Microsoft Gaming CEO Asha Sharma promised "a renewed commitment to Xbox" when she took over from Phil Spencer in February, and in the months since we've seen some movement on that front: A promise to avoid "soulless AI slop," the end of the "this is an Xbox" strategy, and most recently, a reduction of Game Pass pricing—and the removal of Call of Duty games on day one.

Now, thanks to a joint message from Sharma and newly-installed chief content officer Matt Booty that was sent to Xbox employees and subsequently shared publicly, we have perhaps a tiny little bit of a better idea about where it's all going. Beginning with a name change: Microsoft Gaming, which was adopted in 2022 as an overarching title for Xbox Game Studios, Bethesda, Activision. Blizzard, and King, is out, and to replace it, the old Xbox name has been brought back.

That change is actually noted near the end of the post, following a lot of words that don't actually say a lot about Microsoft's plan for its future in gaming: "Games should bring people together through shared experiences," and "Xbox will be where the world plays and creates"—the sort of empty, rally-the-troops mush that can simultaneously make your eyes gloss over and your soul rage. Or maybe that's just me.

Still, there are some rather frank admissions of failure to be found in there. Some of it has been expressed previously in the 'we know we have to do better' way, but it's kind of stark when all rolled together as it is here:

"Players are frustrated. New feature drops on console have been less frequent. Our presence on PC isn't strong enough. Pricing is getting harder for people to keep up with. And core experiences like search, discovery, social, and personalization still feel too fragmented. Developers and publishers are asking for more, too: better tools, better insights, and a platform that helps them grow faster."

There's also a nod to the proverbial "War for Attention" that's become a popular thing to wring hands over in 2026 and "a new generation of players" who "expect more content in familiar places, want to shape the worlds they play in, and want to create and socialize together, not just play together."

The trouble is that while there are certainly challenges a-plenty, most of what it offers in the way of solutions to these challenges are vague promises and platitudes. To wit:

"Xbox will be built to be affordable, personal, and open. We will offer flexible pricing so it's easy to get started and keep playing. The experience will adapt to you, letting you customize how you play, helping you find what you'll love, and connecting you with the right people. And we will be open to all creators, from individuals to the largest studios, giving anyone the tools to reach a global audience and keep their games growing over time.

It's interesting, though, that underneath all that muck is a tacit acknowledgement that PC may be where the future lies. The admission about Xbox not being "strong enough" on PC is obviously telling (and also conjures memories of Microsoft's long history of half-assing its approach to PC gaming), but there's also a deeper segment dedicated to broader PC-driven shifts in the industry.

"Windows now represents more players and more hours and is increasingly where competition is most intense. Players have access to more games than ever, even as the cost and time to build blockbuster titles continues to rise, putting pressure on what gets made and how risk is taken. Some of the biggest recent hits are coming from small teams or even single creators, and places like Roblox are producing experiences that rival major franchises in scale. More players are also choosing subscriptions and services as their primary way to play, with expectations set around instant access, ongoing value, and libraries that evolve continuously."

That part of the memo brings to mind comments made by Google Cloud global director for games Jack Buser, who said in a recent interview that "the only real growth [in games] is coming from Roblox and China." His proposed solution is to go all-in on AI-powered game dev, which—allowing for conversations about how exactly we define "AI"—I think is a terrible idea. But while Sharma promised to steer clear of "AI slop," AI in some sense will, inevitably, play a role in the future of Xbox: The memo says Xbox leadership "will reevaluate our approach to exclusivity, windowing, and AI, and share more as we learn and decide."

And yes, it sounds like games-as-platforms are going to take a more prominent role in Microsoft's plans as well: An oddly isolated line in the memo states that "our new north star will be daily active players."

I can't say the memo gives me any more, or less, hope for a successful future for Xbox than I had yesterday. Microsoft remains the underdog on the console front, and it's never been able to convert its operating system dominance into success in the world of PC gaming. All that can be turned around, sure, but it will take serious time and effort, at a juncture where Microsoft seems to have its mind on other things: The company also launched an early-retirement buyout program for employees today, the first in the company's history, as it continues to focus on AI.

Others are more upbeat about it, though: Industry analyst Mat Piscatella, for one, called the memo "excellent" and said it sets "a strong foundation and a vision for the future."

This is excellent.The read on the current market is spot on. The goals are made clear, and the tactics laid out.It's no guarantee for success, but it does set a strong foundation and a vision for the future.news.xbox.com/en-us/2026/0...

— @matpiscatella.bsky.social (@matpiscatella.bsky.social.bsky.social) 2026-04-23T23:56:41.747Z

Whatever your take on it, at this point it doesn't seem that Sharma—herself a former AI executive at Microsoft—is quite the "palliative care doctor who slides Xbox gently into the night," as original Xbox co-creator Seamus Blackley predicted when she took over from Spencer. At the very least, Microsoft is making a lot more noise than you'd reasonably expect if it was actually looking to pull the plug on Xbox.

The issue for now is that while this memo admits that a lot's gone wrong, it doesn't have much of substance to say about how it will be fixed. And without that, it really doesn't say anything at all.

Masters of Albion has Peter Molyneux's paws all over it. The art style and arch British humor channels Fable, and it's even set in a place called Albion (not the same Albion as Fable, of course). It's all very gentle, very quaint, like I've just opened an Enid Blyton or Jill Barklem picture book.

But then my in-game mentor, who instructs me in all things being-a-god, goes hell for leather: "Fuck off it's a fucking zombie!" she cries. It's a response commensurate with the severity of the situation—midnight zombie attack—but it's the funniest case of tonal whiplash I've experienced in a game for a while. Imagine if Maru in Stardew Valley called you a cockhead every time you proffered a snow yam. That's how it felt.

Masters of Albion is a funny thing and yes, it kinda feels like the explicit, unadulterated version of previous Molyneux games mashed together. It's a god game, because I play as a godly hand in control of everything. It's a management and production line sim in ways that recall both Theme Park and Dungeon Keeper. It also has some light action RPG elements. It definitely feels like Molyneux and his team are making a game designed to be widely played rather than as an experimental commercial gambit, possibly for the first time in 22cans' existence.

God hand

The first order of the day is to revitalise the little hamlet of Oakbridge. In Masters of Albion the protagonist—the god, the hand—is also the mouse pointer. I can construct buildings out of pre-fabbed blocks, all the better to get a little bit of industry going on in this backwater. My farm keeps a nearby mill supplied with wheat, and the mill keeps a nearby factory supplied with flour. At first my town's specialty is making weird sandwiches, which I'm tasked with making manually, by way of learning how the mini in-game economy works. From a selection of ingredients—stale sourdough, moldy white, lettuce, tomato, rat—I click and drag ingredients onto a plate, and the customer delivers their verdict.

It's a wholesome way to live. But the problem with this revitalisation plan is there's a graveyard nearby and it tends to spawn vicious undead by night. Masters of Albion is also, therefore, a tower defense game of sorts. I can fix or mend fences, build catapult sentries, and eventually open a heroes guild where the heroes I employ bed down.

I really love this aspect of the game. My heroes automatically fight zombies by night and I can even help out from the firmament, at first with lightning rods but later with fire. I can also inhabit my heroes and wander the world as them, effectively turning Masters of Albion into a third-person action RPG (you can inhabit workers as well, but no thanks).

Not a very good action RPG though, or at least, not yet: this being early access, my heroes traverse the terrain awkwardly, and the land is rife with weird invisible walls. But inhabiting one of my heroes means I can explore chunks of Albion that my godly hands have yet to win dominion over, and it also falls to them to reconstruct towers that can extend my divine reach, all the better for me to take control over more of Albion. Each of my heroes levels up independently and, eventually, once I've built an armory and weaponsmith, can be kitted out with better gear.

Rat sandwich

The management elements are fine but a little tedious at times. At one point, hundreds of pounds in debt, I needed to build a bunch of new worksites and homes for incoming workers. The only way I knew how to get out of debt was to manufacture 28 "balanced" sandwiches, the problem being that my wheat field was depleted and I needed more than a grand to irrigate purified water to the plot. Thankfully my mill was stocked up with plenty of flour already, but once my supply was depleted I still needed to make 15 pies with only 37 pounds at my disposal.

Manufacturing is slow too: needlessly so, because Masters of Albion doesn't pit you against any clock. Days will run forever until you choose to end them, the idea being that you should be well-equipped to take on the night's zombie incursion. It's possible to speed up production on my farms, mills and factories by double clicking on them—think of it like a foreman's stern tap on the shoulder—but this is still pretty slow as well.

Anyway, I ran out of money and wheat before I could fulfil this task and felt at a loss until I realised I could go adventuring for coins. Sidequesting as one of my heroes is actually a pretty effective way of making money, or at least it is until all sidequests are done. Manufacturing sandwiches is pretty lucrative, but so is bashing hornet nests from trees, and kicking chooks back into their run (it wouldn't be Molyneux without fowl cruelty).

Once I got the money I bought my stupid water purifier, only to struggle endlessly with its placement, which raises the problem of building in general: it's awkward. I want my town to look nice, but it's usually more expedient to just stack required blocks one on top of the other until I have a veritable skyscraper of, say, farmer workshops, all the better to employ workers and boost production. It's also annoying how willing new blocks in hand are to snap to buildings I don't want to be a part of.

The UI, in general, is a bit of a mess. And performance is pretty woeful: my laptop sits pretty much exactly on the recommended specs, but even on minimum settings the game would often lurch, especially when on foot as one of my denizens. 22Cans also writes that the game is Steam Deck ready, and it kinda is, but you're going to need to be comfortable with frame rates that hover around the 22 fps mark on the lowest possible settings.

Yep, it's an early access project alright. But through the mid-development murk I can see huge potential. It's full of fun touches, like the ability to pick my subjects up and drop them into the ocean, though maybe that stands out because there aren't a heap of other ways to assert my divinity. The writing is often superb, and despite the ropey performance it looks really nice on high settings.

I enjoyed my time with Masters of Albion despite its myriad problems, though for anyone less than fanatic about Molyneux' work or the god genre in general, I'd probably advise to wait for further updates. Of course, this being a Molyneux project, you're justified in worrying whether it'll ever be completed. Assuming that it is—22cans predicts a 12 month dev cycle—it's shaping up to be the best Molyneux game since the Fable years. Is that a high bar to cross? Not really, but I'll definitely be coming back.

If you've grown a bit weary of survival games where a stick and a rock can automagically be combined to make a tree-felling axe and you can build a comfortable base in a matter of minutes, how about something a bit more authentic?

Wordless Forest is a FMV (Full Motion Video) survival game that could wind up feeling more realistic than most of what you'll find on Steam—because its creator actually went out into the woods for weeks on end to make it. Solo developer Dogus Cagrici both stars in the game and recorded all the footage himself while hiking alone deep in the real wilderness.

That makes the creation of the game sound like a survival experience in itself, with Cagrici recording footage of himself "in the dense forests and wilderness of four different cities across Turkey for 40 days. Without a production vehicle, I hiked for kilometers carrying heavy camera gear on my back just to reach remote locations," Cagrici says.

"Operating out of a remote forest pension as my basecamp, I hiked deep into territory actively shared with brown bears and wild boars every single day. The danger shifted with the geography; in Muğla, I waded into freezing rivers and worked in soaking wet clothes, while in Antalya, I filmed alone on the edge of treacherous cliffs where a single misstep could have been fatal."

I suspect this is true: "The exhaustion on screen is 100% real," Cagrici says.

The Steam page and trailer for Wordless Forest definitely demonstrates that: we see Cagrici's character drinking from puddles, eating raw fish, and attempting to start a fire and cross icy rivers. Players will be able to make choices to determine what the character does next in life-or-death situations, such as choosing to approach or ignore a mysterious noise in the woods at night. (I wish it showed a third option: run away screaming, because that's how I'd handle it.)

Some of those choices aren't gonna end well: the trailer shows a few of what I assume are game-enders, such as when we see the character falling off a cliff and landing on the rocks below or being dragged lifelessly down the rapids.

If you're wondering why it's called Wordless Forest, it's because the game is completely dialogue free. That's a selling point for me: as a fan of survival films where the dialogue is at a minimum, like Cast Away, All is Lost, and Inside, there's something deeply fascinating about watching someone try to survive while absorbing all the information visually rather than having it plainly spelled out in dialogue or voice over.

I'm also intrigued because Wordless Forest may be the first game of its kind: an FMV survival game. I know there have been a few FMV survival horror games (Phantasmogoria leaps to mind), but I can't think of any straight-up survival games with an FMV format. A search of SteamDB for survival games with the FMV tag only brings up four hits, and none of those look strictly like survival games.

Cagrici says Wordless Forest has been in development for nearly three years. It's due out in August on Steam.

In February this year, Turkiye's Family and Social Services Ministry announced draft legislation to impose various restrictions and overweening oversight on Steam, Epic, and other gaming platforms. The country's parliament has now approved the bill, with some minor changes, which will mean tough times ahead for such platforms, and newly limited access for Turkish players.

The bill is part of a wider legislative push which, the government claims, is about protecting children by restricting access to things like social media. Some of the requirements of the new legislation are pretty boilerplate, such as clearly displaying age ratings and offering "clear [and] easy-to-use" parental tools, but the platform-holders will surely breathe a sigh of relief that the regulator will not have the authority to ban them (though it can throttle traffic after certain violations).

Any gaming platform that has more than 100,000 daily users in Turkiye must now appoint a local representative (though they don't have to open an office). This rep's contact info must be easily accessible on the platform itself, declared to the authorities, and they have to respond to government requests within a strict timeframe.

If the platforms breach the new rules, first up is a warning and one month to correct the violation. If they fail to fix the issue to the authorities' satisfaction within that month, then it's a 10 million Turkish lira fine ($223,000), and after another month of non-compliance the fine trebles to 30 million Turkish lira ($668,000). Should the fines not have the desired effect, this is when the regulator would begin limiting access speeds to the platform in question.

But next to the social media companies, gaming platforms arguably got off lightly. Many more draconian proposals were rejected. The bill no longer grants authorities the power to shut down gaming platforms, and it removes a requirement for mandatory data-sharing with the government, as well as one for maximum playtime limits. A proposal to ban unrated games was rejected, though they will now be automatically assigned an 18+ rating. Individual developers no longer have the duty to verify a user's age: it's down to the platforms.

The large social media firms, on the other hand, got it in the neck. First up, all social media for children under 15 is now banned in Turkiye. In addition, these platforms must now comply with content removal requests that involve national security or public safety within one hour, as well as ensuring any deleted material cannot be re-uploaded. They'll also be required to provide data to the authorities on-demand with a 15-day deadline for compliance, and to fund risk assessments on the social impact of their platforms.

There's also a new framework of fines for social media platforms, and much more onerous consequences if their new obligations are not met. If a platform does not act on an issue within 30 days of being asked, the regulator can now prohibit any new advertising on the platform. Should the authorities still be unhappy, they can then impose bandwidth reductions of 50 percent that can rise to 90 percent.

The problem being, of course, that what's a matter of "public safety" or "illegal content" is very much in the eye of the authoritarian beholder. This bill was introduced under President Recep Tayyip Erdoğan, whose government is routinely accused of authoritarian media censorship and information suppression: it says this is all about ensuring child safety, but being able to squeeze these sites in such a way is a big stick that can be waved at them for all sorts of reasons as the Turkish government sees fit.

Although the law has been passed, it has not yet come into effect. When it does, Steam and everyone else will have six months to comply with the new regulations.

The Assassin's Creed Black Flag remake is just over yon horizon, scheduled to dock this summer, nevermind the leaky hull that Ubisoft always springs when trying to keep a secret. The fourth mainline Assassin's Creed game was an unequivocal pirate game favorite in the years before we had Sea of Thieves. It's still got a place in a lot of hearts thanks to roguish pirate protagonist Edward Kenway and his antics rallying a bunch of pirate legends to a haven in Nassau.

When is the Black Flag Resynced release date?

Assassin's Creed Black Flag Resynced launches on July 9, 2026 on PC (via Steam or Epic), PlayStation 5, and Xbox Series X|S for $60 / £50.

What's new in Black Flag Resynced?

Remake or remaster?

Ubisoft says that Black Flag has been "rebuilt from the ground up" in the latest version of its Ubisoft Anvil Engine, the version it used for Mirage and Shadows. So let's call it a remake, then. Ubisoft says that the upgrade is bringing modern tech like ray-traced lighting, water rendering, and weather simulation.

Who's making it?

Ubisoft Singapore is leading development on Resynced. Yup, that's the studio that did Skull and Bones, for better or worse, "with many of the original game’s developers returning to bring the game to the latest generation of hardware," Ubisoft adds.

Combat and stealth got a rework

Ubisoft was big on highlighting the way that combat has been updated: mostly the ability to do parries and takedowns. On the stealth side, it highlights how Kenway can now freely crouch whenever as a stealth approach.

That plays into all those many tailing missions, which it says have been updated so that being spotted doesn't just instantly fail the mission and instead just means you'll transition to head-on combat to resolve the situation. So stealth is encouraged, but perhaps not required.

What else is new?

• New story scenes, including one with Edward's wife Caroline
• 3 new recruitable crew officers (Lucy Baldwin, "The Padre", and Tobias "Deadman" Smith) with their own story questlines
• The modern day segments connecting to the animus feature "new moments that focus on Edward's internal struggles"
• 10 new sea shanties featuring French artist Woodkid
• New weapon upgrades for your ship
• Cat or monkey pet to live on your ship (Jackd—awwwww, right?)

What's staying the same?

What's NOT in Black Flag Resynced?

When I first opened the warlock skill tree while reviewing Diablo 4: Lord of Hatred last week, I spent a good 20 minutes just reading through all the different skills and upgrades Blizzard stuffed in there. And that was just one of the new classes in the expansion; the other six have been reworked too.

I glanced at the sorceress skill tree and saw all kinds of transformative options in there, like being able to turn your burning meteors into icy comets. However, I should've taken a peek at druids and necromancers—two classes I don't normally play—because they got bigger upgrades than the rest of them.

Druids skills are no longer tied to specific shapeshifting forms. Instead, you select which form you want the skill to transform you into. You can finally decide to live out your fantasy as a purely human druid for storm and earth skills, for example, and all you have to do is select which form you want in the skill tree (for free). This is a big deal for creating builds around items that give you bonuses for sticking to one form, and for anyone, like me, who really isn't a fan of playing as a bear the entire time.

Necromancers might be the biggest winners of the expansion: Blizzard has promoted them to undead generals who can command more minions than ever before. Not only can they finally direct their skeletons to attack targets, they can gather up to what looks like 28 of them, all at once, with the right items.

All their minions now live in the skill tree and work slightly different from before. Skeletal mages are summoned with the necromancer's version of mana, essence, and skeletal warriors show up passively when there are monster corpses nearby. Golems are in the tree, too, if you want one big brute to smash everything apart.

Spoilers for the Lord of Hatred campaign in the video below!

The first thing I thought of when seeing these changes is how much less spamming you have to do to keep your minions summoned during boss fights where they get repeatedly squished. But I'm sure necromancer players will see this and get excited about what the skill tree will let them do to their little bone buddies.

The Book of the Dead, the necromancer's unique, minion-focused class mechanic, will still let you sacrifice your minions for powerful stat bonuses, but won't stop you from still summoning weaker versions of them in combat. This allows you to use them purely as tanks instead of damage-dealers or to activate specific effects on your gear.

All the other classes have lots of new skill variants in their trees as well, but they don't have unique mechanics like druids and necromancers that would need to dramatically change to fit this new era of customization. I'm glad Blizzard put in the extra effort to make sure druids and necromancers can fully take advantage of all the updates to loot, crafting, and skills in Lord of Hatred.

The expansion is launching on April 27 (April 28 for some regions), and will have all of this available from the moment you start the new campaign. Don't listen to the NPCs yelling at you to stay after Mephisto and his army of brainwashed devotees: Saving Sanctuary can wait while you poke around your skill tree and imagine the most creative ways you can take the demon lord down.

Fake messages asking for cryptocurrency to ensure safe passage through the Strait of Hormuz have been sent to multiple ships and shipping companies. The Greek maritime risk management ‌firm MARISKS issued the warning, which was first reported on by Reuters, saying that unknown scammers, claiming to represent the Iranian authorities, had sent various messages demanding transit fees in Bitcoin, Tether and other cryptocurrencies in exchange for "clearance."

Hundreds of ​vessels are currently stranded west of the Strait, which Iran has blockaded in response to the US blockading Iranian ports. In normal times around a fifth of the world's oil and natural gas passes through the shipping channel.

The scam has clearly been inspired by the fact that Tehran has suggested instituting tolls on ships to ensure safe transit (the Iranian authorities have made no comment on the messages). It trialled the plan earlier this month, per the Guardian, requiring tankers to provide their cargo list, destination and ownership before paying a toll of around $1 per barrel of oil: to put that in context, the average oil tanker will be transporting two million barrels. Then the kicker: Iran accepted payment in either Chinese yuan or cryptocurrency.

On April 18 ⁠the strait was temporarily opened subject to Iranian checks, but at least two ships that tried to pass through were fired upon by Iranian warships and reversed course. MARISKS said ​at least one ​of these vessels, ⁠which was hit by gunfire, was a victim of the fraud, adding that "these specific messages are a scam."

MARISKS also gave an example of the messages being received by ships:

"After providing the documents and assessing your eligibility by the Iranian Security Services, we will be able to determine the fee to be ⁠paid in ​cryptocurrency (BTC or USDT). Only then will your ​vessel be able to transit the strait unimpeded at the pre-agreed time."

That ship may not be alone in falling for a crypto scam while seeking safe passage. On April 22 the cargo ship Epaminondas, owned by the Greek company Technomar, was fired upon after apparently receiving a message promising safe passage. Ekathimerini reports the authorities are investigating whether the message in question "may have been fraudulent.".

The Iranians have fired upon ships in other incidents, while since April 12 the US Navy has also been sticking its oar in with a blockade, all while President Donald Trump claims it has "total control" over the Strait, adding the Navy will "shoot and kill" Iranian mine-laying boats.

Crypto-scammers being their scummy selves is the last thing this seemingly intractable situation needs, but here we are. It already feels quaint to look back just a few years to early crypto scams and NFT game rug pulls, now that we've reached full Neal Stephenson novel territory.

Vampire Crawlers may be in a completely different genre than Vampire Survivors, but the progression path still has that signature Survivors DNA. With the right upgrades, you can become an unstoppable force of nature that shreds enemies like paper.

Picking the wrong upgrades can seriously slow you down in the early stages of Vampire Crawlers, though, so you should spend your hard-earned coins wisely. There's thankfully a refund button in the shop, but some of the game's best upgrades have to be unlocked the old-fashioned way. Here's what you should be chasing first as you dive into Vampire Crawlers' dungeons.

The best Crawlers to start with

Crawlers are the cornerstone of your build. Broken runs and crazy stats start with them, so filling the empty seats at the Inn should be your number one priority when starting out. You can have a great time with any of the game's characters, but there are two Crawlers in particular that you should rush toward.

Pasqualina Belpaese is the first truly broken Crawler you can recruit, increasing your hand size when purple cards are played and increasing your Area stat whenever her Crawler card is played.

To get Pasqualina, you need to reach level 20 with Imelda Belpaese on the Inlaid Library stage. You'll unlock Imelda after your first run in the Mad Forest—the very first stage—so you can get Pasqualina in under an hour if you're just starting out.

Pasqualina's whole kit is designed around crowd control, clearing huge waves of enemies with ever-increasing splash damage. Since she boosts your hand size when purple cards are played, you can stuff your deck with a bunch of cheap cards that you can spam to easily clear encounters. Add in a few stronger cards with combo potential to take out bosses and you've got a bulletproof plan of attack!

Next up on the priority list is Gennaro Belpaese, another broken Crawler who overwhelms enemies by increasing your Amount stat instead of Area. He also deals bonus damage whenever red cards are played, making Gennaro a very easy character to win with.

To unlock Gennaro, you have to defeat the Mantichana in the Mad Forest. The Mantichana is the final boss of the Mad Forest stage, so you have to win a full run to add Gennaro to your roster.

Gennaro is all about spamming low-cost cards and filling the screen with a bunch of random attacks. He gets Knife and Spinach cards in his starting deck, so spam Knives and combo with Spinach to boost your Might and make the Knives even stronger. Might also affects Gennaro's bonus damage from red cards, so all these little buffs start to add up fast. Slotting Gems onto those Knife cards will only ramp up the chaos, too.

Splurge on these upgrades at the shop

A returning feature from Vampire Survivors, the shop lets you spend coins to unlock permanent stat upgrades that'll help you survive harder stages. The same upgrades from Vampire Survivors are sold in Vampire Crawlers, but the best game plan is a bit different here.

The first thing you absolutely must buy in Vampire Crawlers is the Recovery upgrade, which restores one HP after every encounter for only 500 coins. The second and third upgrade levels get pretty expensive, so you can hold off on those for now.

The Reroll upgrade is just as important as Recovery—if not more. For only 250 coins, you can unlock the ability to reroll rewards so you can kickstart honing your build. The next upgrade levels are fairly cheap, but you should be fine with just one level at the start of the game.

Next up, you should buy one or two levels of the Duration upgrade. Crawler buffs are so important to the best builds, so you want to maximize the amount of time they're active. It's another cheap upgrade, so you can buy it after just a few runs.

After that, you should buy the first Might and Luck upgrades. They'll increase your stats by 20%, which is a huge boon when you're just starting out. These upgrades will carry you through the early stages of the game, letting you save up coins for the more expensive stuff like armor and Growth.

Add an Arcana to your kit

The Fortune Teller's Arcana tent is a huge upgrade that can drastically improve your builds, but you can easily miss it if you choose the wrong dungeons.

To unlock the Fortune Teller's tent, you need to find the Polentir Relic in the Furious Forest stage. It may seem tempting to go straight to the Inlaid Library after clearing the Mad Forest, but doubling down and trying to clear the harder Furious Forest is worth the effort. After exploring the Furious Forest for a while, the Polentir will eventually drop.

The Fortune Teller's tent lets you equip an Arcana, which provides a big passive buff of your choosing. These are unlocked just like weapons and items, and you can check the required challenges to unlock new Arcanas by hovering over each card.

One of the best Arcanas is Over the Top, which returns a Crawler to the top of your deck once they leave. You can unlock it by playing Crawler cards 100 times, so you'll end up with it sooner or later after a few runs. With Over the Top equipped, you can rack up constant bonus damage with a Crawler like Gennaro or gain a ton of XP with a character like Imelda.

Another strong early-game Arcana is Your Shield My Liege, which lets you keep armor between turns. It unlocks after you've gained 2,000 armor over the course of your adventures. With this Arcana, you can focus entirely on offense and build up armor across encounters over time. There's even a Gem that deals your armor value as damage, so slotting that into a cheap card is an easy path to victory.

Nab these Relics ASAP

Relics are game-changing passive upgrades that unlock new gameplay features in Vampire Crawlers. You'll get some great ones like the Combo Stack and Gem Hammer in the tutorial, but the best ones are hidden in harder dungeons.

As mentioned above, the Polentir should be your first priority once you're out of the Vampire Crawlers tutorial. You'll find it on the Furious Forest stage, accessible after clearing the Mad Forest for the first time. It unlocks the Fortune Teller's tent and the Arcana system.

The Grim Grimoire is another early-game priority. Found in the Library West Wing, it gives you a list of discovered weapon evolutions so you can keep track of the items you need in order to get the strongest cards. You can access the Library West Wing after clearing the Inlaid Library for the first time.

You should also try to clear the Teeny Bridge as soon as possible. It's a tough, combat-centric dungeon that pits you against a long line of enemies, but the reward is the Stardust Anvil. This Relic unlocks the Blacksmith's Shop, which lets you add gem slots to cards in your deck. This is the last major upgrade you'll get for the Village, so don't neglect this dungeon!

After a year of leaks and horsing around that Ubisoft probably didn't find nearly as funny as I did, Assassin's Creed Black Flag Resynced was finally given the official curtain-lifting today. The reveal of "gaming's worst-kept secret" was pretty much what we expected—it's a remake of Black Flag—but hidden in Ubisoft's big announcement roundup is a fun little bit of trivia: Development is being led by Ubisoft Singapore.

Ubisoft Singapore is one of the many studios Ubisoft has around the world, but it's notable in this particular context because Singapore is also the developer of Skull and Bones, Ubi's other big pirate fantasy adventure—the game whose well-documented difficulties enabled me to do this:

Skull and Bones was the victim of its own Ubisoft curse, you see. Where Black Flag Resync was destined to leak, Skull and Bones was doomed to delays, at least one of which happened less than two months ahead of its planned release in 2022. That date was itself the result of multiple prior delays: Skull and Bones was announced in 2017 and was originally supposed to be out in the fall of 2018. Missed it by that much.

I'm not suggesting here that Ubisoft is going to mash the brakes on Black Flag Resynced until 2032: As much as I enjoy utterly baseless, wild speculation, I'm pretty confident it will make its July 9 release target.

But recall that work on Skull and Bones began in 2013, right on the heels of Assassin's Creed 4: Black Flag's release, and it was originally envisioned as a multiplayer expansion for Black Flag. After more than 10 years of tossin' on turbulent seas, Singapore finally shoves Skull and Bones out the door, and what do they get told? Yer doin' Black Flag again!

It's a living, as the saying goes, but holy cow. After 13 solid years of working on Edward Kenway and his various legally distinct spinoff adventures, I would imagine the Singapore squad is really looking forward to turning its attention to something different. Hey, maybe Ubisoft can put 'em to work on Beyond Good and Evil 2. Getting that out the door shouldn't eat up more than, oh, another 17 years.

For the record, Ubisoft Singapore isn't working on Black Flag Resynced alone. They're not included on the announcement page but mashed down in the fine print of the actual press release is a list of co-dev studios, and it's basically like someone put Gary Oldman in charge of production: The full roster includes: Ubisoft Barcelona, Ubisoft Belgrade, Ubisoft Blue Byte, Ubisoft Bordeaux, Ubisoft Bucharest, Ubisoft Chengdu, Ubisoft Da Nang, Ubisoft India, Ubisoft Kyiv, Ubisoft Montpellier, Ubisoft Montreal, Ubisoft Philippines, Ubisoft Quebec, Ubisoft Shanghai, and Ubisoft Sofia.

My complaints about the generative AI bubble, and the tech industry in general, can often be reduced to the question 'What are we even doing here?'

Case in point, one of the things Google thinks will save the games industry is the ability to predict when someone is about to stop playing a game so that they can be manipulated, in ways invisible to them, into playing more. To which I have to wonder: What are we even doing here?

The concept was mentioned—not quite in those terms—by Google Cloud gaming exec Jack Buser back at the Game Developers Conference in March. His talk broadly promoted generative AI adoption, hopping across everyday game dev problems AI is being used for: tagging assets, debugging, detecting cheaters. I don't have an especially visceral reaction to those uses for AI, but my stomach churned at the mention of "hyperpersonalization."

AI is great! Here's how we're going to use it against you

AI boosters have been kicking the "hyperpersonalization" term around for a few years now. Microsoft said in a 2024 blog post that the outcome of hyperpersonalized AI marketing should be that "customers' needs are anticipated before they even ask."

When it comes to games, Buser is excited by the idea of generative AI models that invisibly adjust a player's experience so that marketing "feels like it's part of the gameplay experience."

"AI is doing this so well now that it can predict churn before it even happens in extremely robust ways," Buser said, "and it can adjust gameplay as well as offers that you put in front of your players in near real time."

If you're building software that makes Minority Report-style predictions about the likelihood someone will stop playing your videogame so that you can then manipulate them into playing and spending more, I think you should go sit by a body of water, listen to the birds, and ask yourself what the hell you're even doing anymore.

If it's truly impossible to make money selling games without using machine learning to build psychological prisons—like in that one Star Trek: TNG episode where everyone got addicted to willing discs into tubes—then there's no games industry to save. There are just software companies competing to design the best digital nicotine.

We might already describe parts of the industry that way—loot boxes, daily quests, and other tricks all predate AI—but certainly not all of it. Big and small developers are still making games that, whether or not they pay the bills, exist because their creators genuinely want to enrich our lives.

"Many people make games for money, but we make money for games," one triple-A studio founder recently told us. Balatro, despite being a certified time-devourer and one of the best-selling PC games in recent memory, came from a developer who dislikes gambling and says microtransactions make him "want to put [his] computer in the dishwasher." (He hates AI art, too).

But others, especially in Silicon Valley, have clearly decided that making numbers go up is all that matters, and expect us to be thankful that we've been deemed suitable targets for their next number extraction endeavor (or propaganda project). They tell us openly that their goal is to make more money while employing fewer people, and then act surprised when we don't applaud them for it.

Perhaps one day they'll be able to do away with the abstract idea of 'making a good videogame' altogether, designing hits purely by using AI to respond to the dictates of the numbers. They might end up with something like this self-propelled slot machine.

Little Tree Kingdom is about as instantly charming as games get. You want me to help a clan of cute little gnomes build a new town for themselves in a tree? Don't mind if I do! And may I add: "Awwww!"

As if to lull me further into a sense of security, it's also a bit of a deckbuilder—one of my favourite genres. Each turn, your hand of cards dictates what you can add to the colony. Some cards let you modify the tree, extending its roots or branches or adding leaves and flowers to it. Others will build little buildings for the gnomes, like houses to increase their population or mines to gather up underground resources.

The overall goal is to balance the "welfare" of the community across a number of different categories including happiness, security, and essence (whatever that is). Let any of those scores drop below zero, and you'll literally break the gnomes' hearts. Three broken hearts is game over.

At first, that threat seems pretty mild: just plonking down whatever comes into my hand each turn seems to keep my gnomes thriving. I find myself more focused on the shape of my tree than any worries about resources. The branch and root cards allow me to grow things out however I like, and then place objects on the new segments, allowing for some creative beautification of the town.

Every now and then, a festival is held and a trader comes along, allowing me to grab up some new cards. Having not worried too much about gnome capitalism thus far, I'm pretty light on gems to buy with, but scoop up a few basic cards to keep me going. I shouldn't need any more than that, right?

Wrong. You see, the problem is, I've been ignoring a pretty big looming issue: my resources are not infinite. Cards can only be played a certain number of times before they're destroyed, and outside of buying new ones, there's no other way to replenish or expand my options. It's not long before I'm ending up regularly drawing barely useful hands that can only give me slight bonuses rather than actually developing my town in significant ways. My whimsical little shopping trips have not really been enough to keep me stocked.

With weather events and other catastrophes always threatening to chip away at my welfare scores, that means I've dug myself a bit of a hole. I'm still just about scraping by, though, until all the ghosts start appearing.

It turns out this adorable forest is home to more than just cutesy gnomes—for reasons I should probably interrogate the little guys about later, it's also full of vengeful evil spirits, who are intent on tearing down everything I've built. Past a certain point they'll periodically spawn at the edges of the map, and begin their inexorable advance towards the tree.

Keeping them at bay means building up adorable little defences: spiked walls, arrow towers, even growing thorns on the tree. I guess this is also a tower defence game?

But, again, I can't simply build those things whenever I want—I need the cards, and that means I should have planned ahead about 5 days earlier. What makes things even trickier is that defences have a very small area of effectiveness. An arrow tower, for example, will only fire at the area directly around it, making it useless against a ghost on the other side of the tree or underground. Sculpting my tree purely for aesthetics was not wise after all—now I need a more defendable shape.

As the angry dead ravage Gnomeville, I'm afforded plenty of time to ponder my mistakes. Little Tree Kingdom is not, after all, a contemplative little cosy sim. Behind its cutesy smile works a devious mind, and really the way to approach it is as a sort of puzzle roguelike rather than a city builder.

The key to success is planning for disaster and aggressively grabbing at ways to compensate for your tree's current weaknesses. That humble little merchant is a lifeline—in later runs, I focus much more on cash generation so I can reroll his store for exactly the cards I desperately need to survive another nightmarish fairytale winter.

Relaxing it is not, but it is good fun. I've played a lot of roguelikes, but I don't think I've ever tried one quite like Little Tree Kingdom before, and even from a city builder perspective it's almost like the usual formula in reverse.

Normally in the genre you're always expanding and gaining more resources. You constantly unlock more options, and you're always equipped with the tools you need to progress. Instead, this is a game about using dwindling resources effectively, and trying to live with serious limitations on what you can do. That won't be for everyone, but it's a fascinating twist that I certainly didn't expect from the premise.

And you can find out for free whether that's your sort of thing or not—Little Tree Kingdom has a demo on Steam right now.

I haven't 100% Crimson Desert, I haven't even attempted it because, if I can be frank, I'm too sleepy to deal with what ever this, this, or this is. But you know who isn't scared to take on the challenge? Palworld's community manager John "Bucky" Buckley, who recently shared that he completed Crimson Desert's achievements.

"I have 100% Crimson Desert," Bucky says in a social media post. "Simply put, Crimson Desert made me feel a sense of wonder and adventure that no other game has. The only thing that comes close for me is Morrowind. A truly breathtaking experience, and one that I enjoyed ever second of. Easily my game of the year."

It's certainly impressive, but on further inspection of Bucky's screenshot my emotions shifted to a primal sort of terror. Next to the 100% metre which shows he's finished all 34 achievements you can see the play time, which is currently sat at 189.1 hours—bone chilling.

You know who else finds this impressive? The Crimson Desert devs: "Wow, in under 200 hours? That's impressive! We're glad you enjoyed the journey." I've only put hundreds of hours into a few select games, and thousands into the one which shall not be named (Overwatch). But that's over multiple years, Crimson Desert has only been out for just over a month.

But even after so many hours Bucky reveals that he's not even close to being done yet: "Ironically, 100%ing Crimson Desert actually means you completed every challenge. I still have like 20 plus hours worth of side missions still available to do. Such a fantastic game. Genuinely made me feel like a little kid again."

Like I said before, I've not given Crimson Desert a look in yet. But it sounds like an interesting game if all my colleagues are to be believed, alongside the over 5 million sales it managed to swipe up just a couple of weeks after launch. Although one thing's for sure it'll be a very, very long time until I can boast about 100% Crimson Desert.

After more than a year of leaks that got so out of hand they forced Ubisoft's legal department to crack a smile, the big moment is finally upon us: Assassin's Creed Black Flag Resynced has been formally, fully, and finally revealed to the world. And it won't be too long before we get to actually play it, as Ubisoft also unveiled the launch date of July 9.

We've also got the Black Flag Resynced PC system requirements:

• Minimum System Requirements For 1920x1080, 30 FPS, Low Preset, Ray Tracing – Standard, Upscaler – Balanced
  • CPU: Intel Core i7-8700K 3.7 GHz, AMD Ryzen 5 3600 3.6 GHz
  • GPU: NVIDIA GeForce GTX 1660 (6GB), AMD Radeon RX5500XT (8GB) or Intel ARC A580 (8GB)
  • RAM: 16GB (Dual-channel setup)
  • Storage space: 65 GB (SSD required)
  • OS: Windows 11
• Recommended System Requirements For 1920x1080, 60 FPS, Medium Preset, Ray Tracing – Standard, Upscaler – Balanced
  • CPU: Intel Core i5-10600K 4.1 GHz, AMD Ryzen 5 3600 3.6 GHz
  • GPU: NVIDIA GeForce RTX 3060 (12GB), AMD Radeon RX6600XT (8GB), or Intel ARC B580 (12GB)
  • RAM: 16GB (Dual-channel setup)
  • Storage space: 65 GB (SSD required)
  • OS: Windows 11
• Recommended System Requirements For 2560x1440 (2K), 60 FPS, High Preset, Ray Tracing – Standard, Upscaler – Balanced
  • CPU: Intel Core i5-11600K 3.9 GHz, AMD Ryzen 5 5600X 3.7 GHz
  • GPU: NVIDIA GeForce RTX 3080 (10GB) or AMD Radeon RX6800XT (16GB)
  • RAM: 16GB (Dual-channel setup)
  • Storage space: 65 GB (SSD required)
  • OS: Windows 11
• Recommended System Requirements For 3840x2160 (4K), 60 FPS, Ultra Preset, Ray Tracing – Extended, Upscaler – Quality
  • CPU: Intel Core i7-12700K 3.6 GHz, AMD Ryzen 7 5700X3D 3 GHz
  • GPU: NVIDIA GeForce RTX 4090 (24GB) or AMD Radeon RX7900XTX (24GB)
  • RAM: 16GB (Dual-channel setup)
  • Storage space: 65 GB (SSD required)
  • OS: Windows 11

A Core i7-8700K, GTX 1660, and 16GB RAM struck me as a little on the steep side for a minimum spec—although some of that feeling may arise from the fact that Black Flag originally came out in 2013. You shouldn't need that much horsepower to run a 2013 game, right?

It's not really a 2013 game, of course: Black Flag Resynced is "rebuilt from the ground up" in Ubisoft's Anvil engine, and as PC Gamer hardware writer Nick Evanson noted, the Resynced requirements are quite similar to those of Assassin's Creed Shadows, which he said "makes sense, because they're both on Anvil."

Black Flag Resynced promises "updated visuals and enriched gameplay, including parry driven combat, improved stealth and parkour, deeper naval mechanics, and new narrative content." Among that new content is a trio of new officers who can join you on your adventures—Lucy Baldwin, the Padre, and Dead Man Smith—each with their own unique quest that reveals more about their histories. They'll also bring new special abilities to your ship, like double shots from broadside weapons.

"Assassin's Creed Black Flag Resynced matters deeply to us and to you," game director Richard Knight said during today's showcase. "That's why we took the time to rebuild it with the care it deserves. This remains a solo adventure and character-driven experience. It is not an RPG. The focus stays on how you play and how you explore the world."

That might seem like kind of an odd thing to focus on, but there's a point: Creative director Paul Fu said that because Ubisoft wants Resynced to be a "pure story-driven adventure" that's "fully focused on Edward's adventures," it will not offer the multiplayer elements that were in the original game.

Here's a deeper dive into the updated gameplay:

And some new screens:

It's been quite a ride, with spoilers coming from merch makers, voice actors, age rating boards, and various "anonymous sources," who probably aren't used to having to stand in line when they want to spill some beans.

Ubisoft didn't take that leakiness especially well at first but eventually had a change of heart about the whole thing and decided to embrace the absurdity of it all. Which was understandable, really: Have to laugh to keep yourself from crying, and all that.

The standard edition of Assassin's Creed Black Flag Resynced goes for $60, and is up for pre-purchase now on Steam, the Epic Games Store, and the Ubisoft Store. Ubisoft also confirmed during today's showcase that the original Assassin's Creed 4: Black Flag will remain available for purchase.

Assassin's Creed Black Flag Resynced will be out on July 9—I already said that, I know, but a reminder never hurts. In the meantime, why not try our new pirate quiz? I can promise* you that "What's a pirate's favorite letter?" is absolutely not one of the questions.

(*not actually a promise, I have no idea)

SpaceX, the private aerospace and AI company founded by Elon Musk, is reportedly planning to make its own GPUs in the not too distant future.

The company plans to go public this summer, with an expected IPO of $1.75 trillion. Part of that process involves filing an S-1 ​registration with the U.S. Securities and Exchange Commission, which details a company's finances and risks prior to going public. Reuters reviewed an excerpt of this document, and spotted that SpaceX lists "manufacturing our own GPUs" under its "substantial capital expenditures."

Now, you and I think of a very distinct, game-ready thing when we hear the term 'GPU', but I suspect SpaceX's plans don't fully fall upon the same page. The odds are these chips will be more specifically geared towards some sort of AI workload, not unlike Google's tensor ​processing units (or TPUs, if you were hankering for yet another hardware initialism).

It's not yet clear exactly how much cash SpaceX might be pouring into this hardware endeavor, but it's hardly a surprising development given Musk's recent team up with Intel. This partnership will see Intel "design, fabricate, and package ultra-high-performance chips at scale" in order to "accelerate Terafab’s aim to produce 1 TW/year" in compute power.

For those that need the refresher, the Terafab project is an advanced AI chip manufacturing complex planned to be built in Austin, Texas. The massive project currently intends to handle chip fabrication, packaging, and testing. It is a megazord effort between SpaceX's xAI unit and Tesla, though it's not yet clear the exact type of chips this fab will produce.

Most recently, Musk said in an earnings call that Terafab will "use Intel's 14A process, which is state-of-the-art and in fact not yet totally complete. But given that by the time Terafab scales up, 14A will be probably fairly mature or ready for prime time, 14A seems like the right move."

To return to SpaceX's GPU plans, it's currently unclear whether a partner such as Intel will fabricate these, or the company will look elsewhere. It's kind of a weird time to announce any fresh hardware venture, especially as GPU giant Nvidia's main manufacturing partner TSMC has its hands very full and many other production lines are similarly fit to busting.

Perhaps unsurprisingly then, SpaceX admits in that aforementioned S-1 registration that it does not "have ​long-term contracts with many of our ​direct chip suppliers." The document continues, "We expect to continue sourcing a significant portion of our compute hardware from third-party suppliers, and there can be no assurance that we will be able to achieve our objectives with respect to ​Terafab within the expected timeframes, or at all."

That's probably not the most attractive prospect for investors, but time can only tell whether SpaceX and its GPU efforts entices some sharks.

I won't claim to be a tactical genius, but when it comes to roguelikes, I've played more than my fair share. These days it doesn't usually take me long to mash enough synergies together to get my first few wins in pretty much any example you chuck my way.

Today though… I have met my nemesis. I am absolutely terrible at Gambonanza. It's embarrassing.

On the face of it, it's a very simple game. Each round, I'm able to deploy a small number of chess pieces on a little board (initially just three), and my goal is just to capture all the enemy pieces before losing all of mine. All the pieces simply move as they do in chess—so a pawn, for example, can only move forward one and capture diagonally forward one, a bishop can move or capture anywhere in a diagonal line, and so on.

In practice, that makes every round a tight puzzle. Bearing in mind every potentially threatened square while manoeuvring your own pieces into the right positions to capture the enemy without then leaving themselves vulnerable to counter-attack—it's enough to wrap my brain in knots.

And I can't afford to be sloppy about it, either. Gambonanza effectively has permadeath: every piece of mine that is captured is gone forever. I can get randomised replacements from the shop between rounds but it's costly, and one bad round is often enough to put me in a hole that I struggle to dig my way out of.

Normally, this is where I'd turn to cheap synergies to save my bacon, but I don't even know where to start in this game. I can try and stack my bench with particular pieces with selective shopping between rounds, but with a concept this abstract, it's hard to know what will be most effective. Is a board full of knights a good strategy? If so, how do I build around it?

The next layer is gambits I can buy to give myself permanent bonuses—this is where the roguelike skullduggery comes in, in theory, with things like earning free pawns whenever your queen is in check, or making bishops also able to move like kings. I can see the rough shape of how some of these things might come together into game-breaking combos, given time, but surviving long enough to get there, that's another question entirely.

Whenever I do manage to muddle through a few rounds, my reward is to face one of Gambonanza's surprisingly eldritch bosses, who in addition to freaking me out with their disturbing appearances each impose their own special power to make things even harder. That includes effects like making squares of the board fall away into the void, or not being able to see their deployment during the planning stage. Even worse, they often have elite pieces, with properties such as being uncapturable until they're the only one left.

It's at this point that my resolve usually starts to crumble entirely (along with a few squares of the board). This is simplicity wielded expertly as a weapon, creating intricate tactical puzzles out of straightforward, 1,500 year old rules, and I have to applaud its cleverness. If only I was anywhere near clever enough to stand up to it.

Now, this is the point where you all go try it and then come back and tell me it's actually a breeze and I'm a fool. The good news is, you can have that pleasure completely free: Gambonanza has a demo on Steam that you can play ahead of its full release on May 1 this year.

World of Warcraft's latest patch has come with… well, let's just say it's not exactly landed well with the playerbase. That's owing to a swarm of nasty bugs that have made it through the net—ranging from genuinely game-breaking to tiny frustrations, but gathering speed into one big mess.

One of the funnier outcomes of Blizzard's difficulty with QA this patch cycle, however, comes from the Outlaw Rogue, and it's either completely unintended, or just an accidental consequence of something that seems cool on paper but in practice, uh, gets people killed.

I've been playing this specialisation since Dragonflight, and it's always held a special place in my heart—I like the fast-paced, proc-focused elements of it, and Midnight tidied up the spec's Roll the Bones feature so I could merrily keep shooting people without a loud WeakAura screaming at me to toss some dice every 30 seconds.

It has, however, been a touch underperformant in Midnight, and needed some tweaks. Blizzard, like a mischievous genie, has granted my wish—but not without dire consequences.

There's a new talent in the Trickster hero tree, and it's called Cloud Cover. Whenever you activate Adrenaline Rush (you will be doing this a lot), you drop a huge smoke bomb. Attacks made from within the cloud have a chance of applying Fazed, which is the Trickster's damage debuff.

It's an objective improvement to the Trickster rogue for both Outlaw and Subtlety specialisations (that share the Trickster tree) and it's flavourful as hell—I think it's genuinely cool that my rotation now includes throwing down a smoke bomb and doing a bullet storm from within it with my dual pistols a-swashbuckling.

There's just one problem: This is a game about not standing in the fire. And you know what makes it hard to stand in the fire? A giant sodding smoke cloud. Look at this thing.

This is a massive problem, because it means that if you want to do content as a Trickster rogue, you've got to basically ask a whole group of very patient strangers to bear with you while you drop a giant muddy puddle in the middle of everything. And while there are options in the menu for adjusting what you see from other players, the point remains that I still gotta deal with it, man. And I'm already bad at not standing in the fire.

Anyway, I hope Blizzard at the very least makes the cloud a little less thick. I do genuinely adore the flavour of it, just not the way it clashes with just about every design decision in every piece of content in the game. At least I don't have to chug a darkness potion to up the contrast, anymore.

The gaming industry is in a bit of a shambles right now, and everyone from developers to gamers would agree. AI's infiltration into the gaming industry is the latest scare for players and devs alike, but it's also a solution according to others. Just yesterday, we reported that Jack Buser, Google Cloud's global director for games, views AI as a salve for the industry's problems.

Well, in an interview with Mobilegamer.biz, that very same Jack Buser states that more game studios are using AI-powered tools in development than are willing to publicly reveal.

"I think what players don’t realise is that their favourite games right now were already built with AI," Buser says. "Those games have shipped. We did a survey around Gamescom last summer with studios all over the world. Roughly nine out of 10 game developers told us, yeah, we’re using it."

He explains that, when we've seen reports like 50% of developers believing generative AI to be bad for the industry—a dramatic increase over the past few years—"that gap is basically the developers' willingness to tell you whether the fact of the matter is it’s being used." Given the very quick, strong reactions towards AI right now, he suggests some studios aren't willing to take the risk in disclosing it.

Right now, Google Cloud's AI tools like Gemini and Nano Banana Pro are, in Buser's words, removing "the drudgery and repetitive, low-value work". For example, he cites Capcom, which has used these tools to "rapidly generate…countless ideas" (such as "pebbles on the side of the road") and "curate" them before the art director can get the team working on making them for real.

It's worth noting that Capcom hasn't been shy about its use of AI in game development, though where it draws the line is unclear. It "will not implement assets generated by AI", but it does use it to "enhance efficiency and boost productivity."

Buser believes that the dislike and distrust of AI tools in games will soon shift once players realise it's "actually helping me get my favourite games faster", and with "more innovation in the industry because there’s more room to take risk".

I agree with Buser's summary of the state of the industry—rightly pointing out burgeoning budgets and more players are sticking to older games—but I'm less convinced that AI is the golden ticket.

I'll be the first to admit that I've been rather critical of our new AI-integrated, "woops-it's-broken" world. That's not to say I'm completely down on the tech, though. I like me a dose of DLSS, and vibe coding has shown some genuine promise in terms of bringing software development to the masses.

I'm also a huge Lego fan. Which means I'm happy to admit that Brickbacklog.com, a vibe-coded web tool that helps you identify and source missing pieces for your old Lego sets, is a very cool thing.

The tool's creator, James Bruce, has written a Linkedin post describing its inception. They're a Lego connoisseur like myself, only with what sounds like a much, much larger collection.

"I love buying old bargain bulk boxes and figuring out what sets were mixed in there," says Bruce. "But finding the right bits to rebuild those sets can be tedious, especially when there are 5 kg of bricks to wade through.

"So one Saturday evening, I decided to make a web-based tool to help me find the pieces I need: something where I could enter a set number, see all the parts, and use a dead-simple tick-box interface to check them off, ideally with a satisfying sound."

No such tool was found to exist. But despite having two decades of experience as a web developer and programmer, Bruce decided to take the easier route and see if ChatGPT Codex could make one instead.

"Within fifteen minutes, it had built a working prototype. Less than an hour later, I’d bought a domain name and uploaded the first version. It was rough, and it didn’t work for larger sets yet, but the proof of concept was there, and it was already useful," says Bruce.

"By Sunday morning, I’d added new features, fixed bugs, and made sure it could scale for massive sets. Then I added a way to save your progress without ever needing to log in, so no sensitive user data is stored. And it all just… worked."

This doesn't quite track with my own (limited) experience with vibe coding, but being a dev to begin with probably helps in terms of prompting the AI in the right direction. Anyway, the resulting tool looks incredibly useful, especially for someone who's planning a move in the near future, like myself.

I have far too many completed Lego sets in my home at the moment. And I know that, no matter how carefully I disassemble them, half my Lego Technic cars will end up without their wing mirrors by the time I unpack them at the other end.

But lo and behold, even my 2022 Ford GT is supported by the app. And yes, half those little doohickies are probably spread underneath various bits of furniture in my house already, and I've simply not noticed they've fallen off. But when I do, I'll be using this little tool to help order replacements—and that's pretty cool if you ask me.

The idea that the whole AI thing is a bubble fit for bursting isn't novel. But when a US senator thinks a bubble is obviously what it is, well, that's particularly worrying.

"I know a bubble when I see one," said veteran senator Elizabeth Warren to the Vanderbilt Policy Accelerator event in Washington (via The Verge). Warren's central critique is that the revenues from AI services aren't keeping up with the huge investments required to provide them.

“If AI companies are unable to increase revenues with lightning speed, they won’t be able to service their massive debt loads,” Warren said. “And because of shady accounting strategies, the first big stumble will have everyone running for the exits, potentially triggering destabilizing losses in the financial sector and another 2008-style financial crisis.”

Warren says AI companies will need to generate roughly $2 trillion in annual revenue by 2030. She claims that in 2025 the industry generated just $20 billion in revenue, 1% of what they would need to earn by 2030 "just to break even".

The similarities with the 2008 financial crisis are particularly worrisome for Warren. "The parallels to the 2008 financial crisis are striking: the reckless behavior of a few billionaires and Big Tech CEOs has turned a promising technology into a structural risk to our financial system," she said, adding, "American families and workers cannot afford another economic catastrophe. They are still picking up the pieces left by the Great Financial Crisis of 2008."

Warren also claims that the AI bubble is currently four times the size of the housing bubble that lead to the 2008 crisis. Warren likewise thinks tech CEOs are acutely culpable in that they are all too aware of the risks they are taking on all our behalf.

"AI companies are aware of these risks—very aware. Instead of reducing their borrowing, slowing their rate of growth, and cleaning up their balance sheets, they are making the classic billionaires’ move: they are quietly lining up for a handout. They have already lobbied the Trump administration for taxpayer funding and guarantees to cover themselves if things go south," she explained.

As for her thoughts on how to mitigate all this, her first step would be to restore the "guardrails" that used to restrict the activities of big Wall Street banks. Here, she's probably talking about the repeal of the Glass-Steagall Act in 1999, legislation which most notably enforced a separation of commercial banks, which hold the deposits of regular consumers, from investment banks involved in more speculative activities. Well, until it didn't, making the 2008 crisis possible, some would argue.

She also wants a new digital regulator to enforce anti-trust and consumer protection laws, and to see big tech paying its "fair share" in taxes. Of course, it will be easy for critics to dismiss Warren's combined warning and call for action as shrill fear mongering from a politician not normally aligned with corporate interests.

But at the very least she's objectively not wrong about the huge sums of money currently at risk in the AI industry. And she's hardly alone in finding that very troubling indeed.

Marathon was the fourth best-selling game (by revenue) in the US when it released in March, and the sixth best so far this year, according to market performance analyst firm Circana. That's quite the surprising statistic given the vitriol surrounding its very existence and the constant doom-watching of its Steam Charts. It's even more impressive when you consider that the stat is based on revenue (not the pure number of copies sold), since Marathon only costs $40—it's not a full-price game, pulling in $60 dollars a pop.

This means that Marathon, on the surface, beat out the likes of Pokémon Pokopia, Monster Hunter Stories 3: Twisted Reflection, and, perhaps most surprisingly, Crimson Desert. In first place was MLB: The Show 26, followed by Resident Evil: Requiem which is still hitting hard in its second month, and then WWE 2K26 in third place, ahead of Marathon.

March 2026 Top 20 Best-Selling Premium Games - U.S. (Dollar Sales, Physical and Digital from digital data sharing publishers, excludes add-on content)

— @matpiscatella.bsky.social (@matpiscatella.bsky.social.bsky.social) 2026-04-23T14:24:32.071Z

However, it's possible that this ladder is slightly different in reality. Circana explains that Nintendo doesn't share digital sales numbers with it, but I wouldn't be surprised if Pokopia was actually much, much higher—surely in the top three. Similarly, Crimson Desert's 15th place ranking doesn't take into account digital sales for the same reason, but the developer has revealed it's sold over five million copies in its first month alone. Senior analyst Mat Piscatella did add that Crimson Desert, MLB The Show 26 and Pokémon Pokopia led in Circana's initial projections, too.

Nevertheless, in such a crowded month, Marathon clearly made a dent at release, despite the naysayers. But with reports that Marathon cost over $200 million (remember that Sony also spent $3.6 billion acquiring the studio), it's also not the big hit either party would want it to be.

This strong launch, estimated at 1.2 million sales, hasn't continued into a consistently high player base in the weeks since, at least on Steam where we can view this data. A month later, the 88,000 concurrent player release day peak has dropped to roughly 15-20,000 daily concurrents on Steam, where it seems to be settling. Going forward, I'm most interested in how Bungie will support and hopefully grow the game—its latest battle royale-esque experiment is a great example of that effort.

Samsung is never far from the headlines these days, as it's the biggest manufacturer of DRAM chips, a crucial electronics component that's difficult to procure in the face of massive AI server demands. This has led to record financial figures, with the company estimating an eight times profit increase in its most recent earnings guidance documents.

But while those glowing figures certainly put the company in an enviable financial position, tens of thousands of Samsung workers are now protesting outside of a South Korean factory complex, demanding fairer pay (via Reuters).

The second-biggest South Korean memory manufacturer, SK hynix, is said by the Samsung Electronics Labor Union (SELU) to now pay its employees over two-thirds more in bonuses, and as a result, an estimated 40,000 Samsung workers have taken to the streets.

The union is demanding a 7% increase in base salaries, and that 15% of annual operating profit be allocated as bonus pay—alongside more transparency around salary calculations, and the removal of the current 50% base salary bonus cap.

The SK hynix bonus increases are a direct result of previous union demands for fairer pay in September of last year, which appears to have influenced the Samsung protests. SELU members are now estimated to number around 90,000, which would represent over 70% of Samsung's South Korean workforce.

Speaking to Reuters, Samsung logistics employee and now-protestor Song Yong-gi said: "In reality, many employees are leaving for SK hynix. At the end of the day, more than 90% of employees work for pay, and the compensation gap has become so wide that it's driving these moves."

Reuters also reports that the black vest-clad protestors have been observed walking over a large banner depicting the faces of the company's co-CEO, Chairman Jay Y. Lee, and other Samsung executives.

Samsung has reportedly said that it would make efforts to reach a swift agreement with the protestors. An unnamed company official has also been quoted as saying that production halts as a result of "even a single strike" could take years to recover from.

And a full strike appears to be the protestor's next move. A planned 18-day strike is said to be due to begin on May 21, unless demands are met. Here's hoping that a reasonable solution can be found before then.

After draping a white sheet over himself to baffle surveillance cameras didn't exactly go to plan, YouTuber Chris the Producer has taken a different tact: slapping a giant TV on his back.

"This is the only device," he says, "that allows you to legally evade America's surveillance system." For my part, I can neither confirm nor deny that statement.

The setup involves a chest-mounted action camera, a tactical backpack with frame mounting system, and a "massive TV" which is fitted to the backpack in front of an industrial battery pack that sits between the two. The camera streams to the TV at the back, displaying just whatever is in front of you:

"It's essentially a giant backpack with a camera on one side that live streams to the TV mounted on my back, therefore obscuring my body and face, making me entirely invisible."

I'm not sure about that "entirely invisible" part, but if it "works pretty well on dogs" which are "basically the same thing" as surveillance cameras, then who am I to say?

If you want to make one yourself, he does break down the cost:

• GoPro with chest mount and cords: $400
• TV (stolen from family): $250
• North Face backpack: $110
• Super-heavy New Yeti 1500 battery pack: $1,500

According to the appropriately succinct video description, this "kinda set me back a bit but I did evade surveillance cameras for a few hours." The YouTuber does note the ideal setup would be to have it double-sided, however, so you don't have to keep your back to the cameras. That's what bumps up the cost to about $3,000.

What do you think, worth the money? I'll probably stick to just avoiding leaving my apartment instead—the big brain solution. Cameras can't see what isn't there. Just doing my part to combat the ever-encroaching panopticon. Wait, what do you mean they're encroaching digitally, too? To the fields we go, I guess.

Casey Yano, co-creator of Slay the Spire and co-founder of studio Mega Crit Games, has given a new interview to Edge magazine about designing the sequel to a roguelike deckbuilder that, honestly, pretty much got it right first time. With some games it's obvious what a sequel could build on or improve, but when your starting point is so finely tuned that it's scooping PCG's Best Design award… well, where do you even start?

Dark Souls obviously. Yano says that elements of FromSoftware's series were one of the major starting points for where they wanted to take the sequel. The specific inspiration being how, across the Dark Souls games, players can essentially build the same character. But the difference is in what that specific build will be facing. Then the idea ran smack-bang into playtesters.

"We did mess around with, like, what percentage of the card pool should be different for returning characters," Yano tells Edge. "At first, it was less."

"If you say 'I'm going to make a katana build' you can make essentially the same exact character again in Dark Souls 2 and 3—you're just up against new challenges," Yano says. "I like that, personally."

But when Mega Crit started playtesting the early builds, players recoiled at the idea. "They were not jiving with that at all: 'We need new stuff!'" says Yano.

There's definitely something in Dark Souls whereby finding a particular style of weapon is almost like a comfort blanket: I was always delighted to find my first halberd-style weapon in any entry, because I adore the moveset and know how it works in different scenarios. Experimenting is great and all, but how about this one weapon that you know will always get the job done?

Slay the Spire 2 has ended up hewing closely to the formula of the original, though it does add a lot more new stuff than was maybe the case at an earlier stage. It's had a hugely successful launch, though with its most recent updates, many players are grumbling about the direction it's taking. But given that the original game had seven years of post-launch support, this clearly isn't the end of the road.

One thing that will stay, however, is the game having a definitive endpoint. In a post-launch Q&A, Yano addressed the idea of adding an 'infinite mode' of some kind to Slay the Spire 2, and in his reasoning references another classic deckbuilder of recent times.

"Probably not, the deckbuilding gets less exciting the longer a run goes on," says Yano. "I think games like Balatro do a better job due to infinite strategies being possible less often. The scaling of their game's mechanics feels more natural—whereas it's not really the case for us."

Crimson Desert is a massive game with a heap of weird and wonderful systems. Pet dress-up, anyone? What about riding a polar bear into battle? Even with animal affinities aside, that means a lot of moving parts, which in turn means many, many, many squeaky wheels crying out for grease.

As such, developer Pearl Abyss maintains a list of known issues that runs the gamut from quirky NPC behaviour, to the game simply not playing ball on some set-ups. You'll remember that Intel Arc GPU support wasn't available at launch, though now the game is somewhat playable. Although, I say 'somewhat' as even Pearly Abyss notes ongoing crashes and display issues when running Intel's upscaling tech XeSS.

Specifically, Pearl Abyss notes crashes on Intel Arc A770 GPUs when using XeSS, and that "the game screen does not display properly when 'Intel XeSS 3.0' or 'Intel XeSS Frame Generation' is enabled on Intel Arc A-series graphics cards." This is after the latest patch introduced improvements to "the quality of 'Intel XeSS 3.0' upscaling and 'Intel XeSS Frame Generation."

In other words, issues persist but Pearl Abyss appears to be working on it. That's just as well because XeSS isn't the only upscaling tech still coming a cropper in Pywal.

This week's patch made improvements to "the quality of 'AMD FSR Ray Regeneration'." However, according to the list of known issues, using the upscaling tech can still lead to blurry, rainy environments, as well as your screen whiting out if you happen to be gaming on an Nvidia GTX 1060 GPU.

Thankfully, developer Pearl Abyss has been pumping out patches since release, aiming to introduce a suite of improvements long-term. For one thing, this week's patch actually makes the house worth decorating now, and a control scheme refresh now makes it even easier to fly across Pywel.

Bottom line, Pearl Abyss is all-in on Crimson Desert for the foreseeable. Hopefully that means that gamers affected by the known upscaler issues shouldn't have to wait too long for their turn to drink in everything Pywel has to offer.

Just because you can, it doesn't mean you should. Wise words and, dare I say, highly relevant in the context of the pursuit of ever more realistic VR experiences. To wit, would you really want to include accurate odours?

Hold that thought, because we haven't addressed the "can you" bit. It turns out a group of researchers think they can, courtesy of a new approach that involves using ultrasound to stimulate the scent-processing part of the brain. Indeed, they claim the technology can be used not just for smell but also as a highly efficient method for inputting data into the brain. Yes, really.

"We pointed an ultrasound probe at the scent-processing region of the brain to obtain different sensations. Different focal spots corresponded to different smells, which we’ve replicated first-try on two people and validated with a blind trial," the researchers say via this blogpost.

The mere suggestion of such a technology has been enough, inevitably, to have some observers recoil. One commenter on X suggested a possible consequent scenario of, "walking into a VR chat lobby and immediately projectile vomiting as your headset psychically projects the scent of yeast-infected rotting fish from the pregnant fox OC on the other side of the room while it tanks your framerate." Charming.

Specifically, this technology claims to work by stimulating the olfactory bulb. The location of the olfactory bulb makes stimulation with ultrasound particularly tricky, the team says, but not impossible.

"We found that you can place the transducer on the forehead and aim the ultrasound downward towards the olfactory bulb. While this isn’t a perfect solution because the frontal sinuses can weaken the signal, careful device positioning above the sinuses still allows us to reach our general target region."

Fine tuning included using an MRI scan of the subject's skull to optimise the positioning of the ultrasound pad, and then various parameters addressing the actual ultrasound being directed at the olfactory bulb.

That includes the frequency, focal depth and using pulsed output. Thus far, replicated smells include "fresh air, with a lot of oxygen," "the smell of garbage, like few-day-old fruit peels," and "a campfire smell of burning wood."

Anywho, it seems that areas of olfactory bulb are associated with particular smells and are quite densely packed. "The distance between freshness and burning was ~3.5 mm," the researchers say.

Exactly how serious all this is remains hard to gauge. There's clearly an element of whimsy in the diagrams created by the team, for instance. Moreover, I'm not sure what to make of the claim that this simulated smell technology can be used as an interface for injecting data into the brain.

Maybe I should let the research team explain this bit:

"The nose has 400 distinct receptor types, and we can distinguish subtle combinations of their activations, so they could serve as a channel of writing directly into the brain, as a means of non-invasive neuromodulation.

"The olfactory system potentially allows writing up to 400, if not 800 due to two nostrils, dimensions into the brain. That is comparable to the dimensionality of latent spaces of LLMs, which implies you could reasonably encode the meaning of a paragraph into a 400-dimensional vector.

"If you had a device which allows for this kind of writing, you could learn to associate the input patterns with their corresponding meanings. After that, you could directly smell the latent space. A bit of ultrasound, a breath in - and you understood a paragraph.

"People are able to develop synesthesia - being able to hear colors and see smells, and it might be possible to extend that to semantics. However, at this stage it is speculative."

To all this you can also add other questionable pursuits of VR fidelity, including an electronic tongue that can simulate, among other things, fish soup, and another device that zaps your neck and can induce feelings of motion. I'll let you decide whether any of this is a good idea.

Pearl Abyss cannot leave Crimson Desert alone for more than a handful of days. The developer's been pumping out patch after patch, each one somehow bigger than the last. I've been having mixed feelings on how much the game has been changing in such a short period of time, but even I have to admit I can get behind this latest update.

There is a ridiculous amount of big quality-of-life updates to unpack here, but the biggest two? Easily the introduction of difficulty options and an abundance of storage-related changes that actually now provide some purpose to the house near Howling Hill that definitely still isn't empty 100+ hours in on my playthrough.

Okay, it is, but that's about to change! And not just because there's now a handful of new design layouts to choose from.

It's because Pearl Abyss has added a ton of different chest types that can be placed in houses now: the Study Gatherables chest, the Kuku Cooler, the Collectibles chest, and newly-functional wardrobes. Not only does each one serve a different purpose, but they also act as global storage—meaning that items placed in these chests can be automatically pulled for crafting recipes without having to go and grab them yourself. Here's what each one does:

• Sturdy Gatherables chest: 1,000 slots for general materials
• Kuku Cooler: 40-330 slots for food items and ingredients
• Collectibles chest: 1,000 slots for quest items and crafting recipes
• Wardrobe: 100 per wardrobe, up to 1,000 slots for gear

Now I admittedly do kinda wish we could just bung everything in a single chest and call it a day, but having some loose organisational structure in place does go a long way to tackle Crimson Desert's chaotic inventory system. But this update also kinda fixes that, too—inventory is now split into five categories to make finding stuff you need a whole lot easier.

Honestly, this would've been a big enough change to satisfy me for this update. But like I said, Crimson Desert has difficulty options now too. Easy mode makes enemies slower and less aggressive, while widening the parry and dodge windows and reducing the overall damage you take.

Meanwhile, hard mode does the opposite: the parry and dodge window narrows, opponents get faster and more aggressive, and you take more damage. Food items also won't apply their effects until after you finish consuming them entirely, making the moments you choose to replenish health or spirit a little more thoughtful.

Unfortunately, all good patches must come with a sacrifice. This time it's to the damage enemies take when you lob a tree at them. Pearl Abyss says it's been "slightly reduced," so maybe hurling a giant hunk of wood at them is still viable, just less so now. Bummer.

There are hundreds of little changes and fixes—the full list you can check out on the Crimson Desert website—but the last big change that I'm incredibly jazzed about is the addition of a search function and filters to the map.

Once you actually remember to make it out of Hernand and play the rest of the game, Crimson Desert's map gets unruly incredibly fast. It's gigantic, and trying to find specific points of interest becomes a much more tedious task than it has any right to. Being able to search and filter stuff out on the map is gonna be a huge help, one which is encouraging me to dive back in and finally wrap up any loose ends.

Okay, those are the big changes. But my favourite little change? Easily the fact that you can have birds as pets now. I am going to end up looking like an avian Pied Piper by the time I am done gallivanting my way through Pywel.

If birds aren't quite your speed, you might be pleased to know that there are also five new types of pet cat. Plus, a new item that lets them stay on your shoulders longer, a sort of half-apology for fixing a bug where they could end up stuck up around your neck indefinitely. I'm just hoping my new feline friends don't eat my new birds.

Earlier this month, Intel teamed up with Elon Musk to "help refactor silicon fab technology" for Musk's ambitious Terafab project. Now, in a Tesla earnings call, the tech mogul has said he will use the company's upcoming 14A node to make chips for this project.

Terafab is Musk's (one might say) bold venture into making his own chips for xAI, Tesla, and SpaceX to power cars, robots, and data centres in space. It's apparently necessary, given the world seemingly doesn't produce enough chips for these companies' needs. Which isn't as absurd as it might sound, to be honest, as CPUs are reportedly already feeling a shortage squeeze.

There's been hope that one solution to this shortage might be Intel 18A, the process that the previous Intel CEO Pat Gelsinger said he'd "bet the whole company on." We have 18A chips already on the market in the form of Panther Lake mobile processors, and later this year, we'll be seeing them again in Nova Lake for desktop PCs.

14A, on the other hand, is still a couple of years out. Previously, Intel had said that it's not in any hurry to get the process node into full swing "until we know that we have the customers that will accept that demand." Well, how's Tesla as a first customer for you? That should be enough to get the ball rolling, if Musk's Terafab goals are anything to go by.

When he reported on Intel's partnering up with Musk, our Nick hinted that Intel's Lip-Bu Tan would most likely have had foundry and packaging wrapped up in part of that agreement, and this has turned out to be true. That's because Intel didn't exactly have the largest list of external customers for its foundries.

Recently, however, Intel has opened the door to having external customers for 18A, and now with Musk on board for Terafab, we have a first customer penned in for 14A.

Musk expresses some optimism towards the timing of the process:

"We plan to use Intel's 14A process, which is state-of-the-art and in fact not yet totally complete. But given that by the time Terafab scales up, 14A will be probably fairly mature or ready for prime time. 14A seems like the right move. And we have a great relationship with Intel. A lot of respect for the CEO, the CTO, and the new team there. So we think it's going to be a great partnership."

All good news for Intel, then, and presumably for its shareholders too. The company's market cap is now higher than it's been in over 25 years, thanks largely to the Terafab deal plus a multi-year deal with Google.

It all makes a refreshing change from the previous few years of stagnation and seeming issue after issue for Intel. And speaking of refreshes, the latest Arrow Lake Core Ultra 200-S 'Plus' refresh is pretty great, too. Hopefully, things continue in that direction; lord knows the PC hardware market needs as many wins as it can get right now.

The best graphics card is, arguably, the one you can afford right now. Now, that might seem trite, but the reality is that, while prices are frustratingly high thanks to a whole host of reasons (*cough* AI), there are cards worth looking at in every segment. And though you might be swayed by the VRAM debate, our own testing shows that in the real world gaming with an 8 GB GPU is absolutely fine in most cases. It's also interesting to note that prices have essentially stabilised at their current level. While stock might be tight, there is stock and at prices which have remained consistent for the past month, with some actually dropping.

I've tested and gamed on pretty much every new graphics card that's been released in this generation, from the Intel Battlemage cards that launched late 2024, through the initial Nvidia RTX Blackwell launch at the start of 2025, and on to AMD's RDNA 4 cards later last year. I believe that makes me well placed to talk you through what I think today's best graphics card is, and for me it's the AMD Radeon RX 9070.

For most PC gamers this is the graphics card which offers the highest performance for the money, delivering decent 1440p performance, and much improved ray tracing over previous AMD generations of GPU. It's also rocking 16 GB of VRAM to be able to deliver on high graphics settings at that resolution now and into the future, too. It was priced appallingly at launch, but now it's settling down and delivering performance close to the RX 9070 XT and therefore the RTX 5070 Ti, too. It's also been remarkably resilient so far in the face of spiking GPU prices thanks to the ongoing RAMpocalypse.

The absolute top, from a pure performance standpoint, however, still is the Nvidia RTX 5090. That's the best high-end graphics card with no other contenders getting near its mix of style, features, and raw computational grunt. It's also a card where no-one gets close in terms of price, either, especially with prices the way they are right now. At the budget end, we can no longer recommend Intel's B570, it doesn't have the value argument anymore and the reliability and relative performance of the RTX 5050 makes that now the best budget graphics card.

GPU price tracking

🕹️ RTX 5080: $1,290 @ Best Buy
🕹️ RTX 5070 Ti: $990 @ Amazon
🕹️ RTX 5070: $599 @ Walmart
🕹️ RTX 5060 Ti 16 GB: $500 @ Walmart
🕹️ RTX 5060 Ti 8 GB: $380 @ Newegg
🕹️ RTX 5060: $330 @ Amazon
🕹️ RTX 5050: $290 @ Amazon

🕹️ RX 9070 XT: $710 @ Newegg
🕹️ RX 9070: $620 @ Amazon
🕹️ RX 9060 XT 16 GB: $440 @ Newegg
🕹️ RX 9060 XT 8 GB: $350 @ Amazon

🕹️ Arc B580: $290 @ Newegg
🕹️ Arc B570: $260 @ Amazon

Lowest price per GPU

Lowest pricing (Price ($))

Time (Week)

Nvidia RTX 5080 Nvidia RTX 5070 Ti Nvidia RTX 5070 Nvidia RTX 5060 Ti 16 GB Nvidia RTX 5060 Ti 8 GB Nvidia RTX 5060 Nvidia RTX 5050 AMD RX 9070 XT AMD RX 9070 AMD RX 9060 XT 16 GB AMD RX 9060 XT 8 GB Intel Arc B580 Intel Arc B570

Lowest pricing Data

Timestamp (Week)	Nvidia RTX 5080 (Price ($))	Nvidia RTX 5070 Ti (Price ($))	Nvidia RTX 5070 (Price ($))	Nvidia RTX 5060 Ti 16 GB (Price ($))	Nvidia RTX 5060 Ti 8 GB (Price ($))	Nvidia RTX 5060 (Price ($))	Nvidia RTX 5050 (Price ($))	AMD RX 9070 XT (Price ($))	AMD RX 9070 (Price ($))	AMD RX 9060 XT 16 GB (Price ($))	AMD RX 9060 XT 8 GB (Price ($))	Intel Arc B580 (Price ($))	Intel Arc B570 (Price ($))
1	1100	780	550	430	368	300	250	700	600	370	280	260	230
2	1099	750	550	430	340	300	250	700	600	370	270	260	230
3	1000	750	524	430	370	291	250	700	550	370	280	250	230
4	1000	750	546	430	345	296	250	695	550	370	280	250	230
5	1000	749	481	430	340	300	250	670	600	350	270	250	230
6	929	749	499	379	340	289	240	670	600	350	270	250	230
7	999	750	520	430	350	299	240	650	550	350	270	250	230
8	980	749	500	430	342	299	250	650	550	350	270	250	200
9	999	750	500	430	323	299	250	650	560	350	270	250	220
10	989	730	480	430	340	299	250	600	550	360	270	250	210
11	990	699	489	400	320	280	230	600	550	340	250	250	200
12	989	730	500	400	320	279	220	600	530	350	275	250	200
13	1000	750	527	420	342	279	220	600	530	350	280	250	200
14	1000	750	540	420	313	279	220	600	530	370	280	250	202
15	1000	749	499	430	350	279	200	600	530	380	288	240	202
16	1100	750	499	430	350	270	200	620	530	380	288	240	202
17	1230	750	549	430	330	260	235	660	580	390	310	250	200
18	1268	829	540	450	330	290	240	700	580	390	310	250	200
19	1380	899	549	450	330	299	240	720	580	400	310	260	200
20	1350	1000	620	520	330	299	240	720	600	440	310	290	200
21	1365	950	620	520	351	299	260	730	590	440	350	290	250
22	1365	950	620	520	360	299	260	730	630	400	346	300	250
23	1290	950	630	520	360	340	260	730	630	440	346	300	250
24	1290	950	625	520	360	340	260	730	620	440	346	300	250
25	1250	950	630	550	360	349	260	730	610	440	346	300	250
26	1260	950	604	550	360	349	290	730	610	440	340	300	250
27	1290	950	640	570	360	349	290	730	630	440	340	300	250
28	1250	900	630	515	390	349	290	720	620	440	346	290	250
29	1290	990	636	500	390	349	290	710	620	449	350	290	250
30	1290	990	630	500	390	330	290	710	620	440	350	290	260
31	1290	990	599	500	380	330	290	670	620	440	350	290	260

This data represents the lowest price we can find for each GPU from a range of US retailers from August 2025

Quick list

1. Best overall graphics card: AMD Radeon RX 9070

Previous

1 / 5

Next

1440p gaming performance

Black Myth Wukong (1440p High)Cyberpunk 2077 (1440p RT Ultra)F1 24 (1440p Ultra High)Homeworld 3 (1440p Epic)Metro Exodus Enhanced Edition (1440p Ultra)The Talos Principle 2 (1440p Ultra)Total War: Warhammer 3 (1440p Ultra)

Avg FPS

1% Low FPS

RX 9070

73

64

RTX 5070

68

60

RX 9070 XT

82

73

RTX 5070 Ti

84

74

0255075100

Black Myth Wukong (1440p High) Data

Product	Value
RX 9070	73 Avg FPS, 64 1% Low FPS
RTX 5070	68 Avg FPS, 60 1% Low FPS
RX 9070 XT	82 Avg FPS, 73 1% Low FPS
RTX 5070 Ti	84 Avg FPS, 74 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

43

36

RTX 5070

45

36

RX 9070 XT

52

43

RTX 5070 Ti

61

48

020406080

Cyberpunk 2077 (1440p RT Ultra) Data

Product	Value
RX 9070	43 Avg FPS, 36 1% Low FPS
RTX 5070	45 Avg FPS, 36 1% Low FPS
RX 9070 XT	52 Avg FPS, 43 1% Low FPS
RTX 5070 Ti	61 Avg FPS, 48 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

90

67

RTX 5070

78

60

RX 9070 XT

105

66

RTX 5070 Ti

108

84

037.575112.5150

F1 24 (1440p Ultra High) Data

Product	Value
RX 9070	90 Avg FPS, 67 1% Low FPS
RTX 5070	78 Avg FPS, 60 1% Low FPS
RX 9070 XT	105 Avg FPS, 66 1% Low FPS
RTX 5070 Ti	108 Avg FPS, 84 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

104

62

RTX 5070

112

64

RX 9070 XT

111

62

RTX 5070 Ti

119

59

037.575112.5150

Homeworld 3 (1440p Epic) Data

Product	Value
RX 9070	104 Avg FPS, 62 1% Low FPS
RTX 5070	112 Avg FPS, 64 1% Low FPS
RX 9070 XT	111 Avg FPS, 62 1% Low FPS
RTX 5070 Ti	119 Avg FPS, 59 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

107

74

RTX 5070

100

68

RX 9070 XT

120

54

RTX 5070 Ti

129

89

037.575112.5150

Metro Exodus Enhanced Edition (1440p Ultra) Data

Product	Value
RX 9070	107 Avg FPS, 74 1% Low FPS
RTX 5070	100 Avg FPS, 68 1% Low FPS
RX 9070 XT	120 Avg FPS, 54 1% Low FPS
RTX 5070 Ti	129 Avg FPS, 89 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

74

53

RTX 5070

64

42

RX 9070 XT

84

44

RTX 5070 Ti

81

35

0255075100

The Talos Principle 2 (1440p Ultra) Data

Product	Value
RX 9070	74 Avg FPS, 53 1% Low FPS
RTX 5070	64 Avg FPS, 42 1% Low FPS
RX 9070 XT	84 Avg FPS, 44 1% Low FPS
RTX 5070 Ti	81 Avg FPS, 35 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

96

43

RTX 5070

81

53

RX 9070 XT

102

49

RTX 5070 Ti

90

44

037.575112.5150

Total War: Warhammer 3 (1440p Ultra) Data

Product	Value
RX 9070	96 Avg FPS, 43 1% Low FPS
RTX 5070	81 Avg FPS, 53 1% Low FPS
RX 9070 XT	102 Avg FPS, 49 1% Low FPS
RTX 5070 Ti	90 Avg FPS, 44 1% Low FPS

Our detailed benchmarking shows just how close the RX 9070 gets to the far more expensive RTX 5070 Ti and RX 9070 XT, why it's our pick as the best current GPU to buy in 2026

Upscaling performance

Black Myth Wukong (1440p High + Upscaling quality)Cyberpunk 2077 (1440p RT Ultra + upscaling quality)F1 24 (1440p Ultra high +upscaling quality, 2x FG)Homeworld 3 (1440p Epic + upscaling quality)The Talos Principle 2 (1440p Ultra +upscaling quality, 2x FG)

Avg FPS

1% Low FPS

RX 9070

111

96

RTX 5070

107

81

RX 9070 XT

121

104

RTX 5070 Ti

124

107

037.575112.5150

Black Myth Wukong (1440p High + Upscaling quality) Data

Product	Value
RX 9070	111 Avg FPS, 96 1% Low FPS
RTX 5070	107 Avg FPS, 81 1% Low FPS
RX 9070 XT	121 Avg FPS, 104 1% Low FPS
RTX 5070 Ti	124 Avg FPS, 107 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

149

102

RTX 5070

122

97

RX 9070 XT

174

106

RTX 5070 Ti

167

118

050100150200

Cyberpunk 2077 (1440p RT Ultra + upscaling quality) Data

Product	Value
RX 9070	149 Avg FPS, 102 1% Low FPS
RTX 5070	122 Avg FPS, 97 1% Low FPS
RX 9070 XT	174 Avg FPS, 106 1% Low FPS
RTX 5070 Ti	167 Avg FPS, 118 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

261

114

RTX 5070

176

113

RX 9070 XT

292

119

RTX 5070 Ti

226

142

075150225300

F1 24 (1440p Ultra high +upscaling quality, 2x FG) Data

Product	Value
RX 9070	261 Avg FPS, 114 1% Low FPS
RTX 5070	176 Avg FPS, 113 1% Low FPS
RX 9070 XT	292 Avg FPS, 119 1% Low FPS
RTX 5070 Ti	226 Avg FPS, 142 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

123

64

RTX 5070

127

63

RX 9070 XT

136

52

RTX 5070 Ti

128

58

037.575112.5150

Homeworld 3 (1440p Epic + upscaling quality) Data

Product	Value
RX 9070	123 Avg FPS, 64 1% Low FPS
RTX 5070	127 Avg FPS, 63 1% Low FPS
RX 9070 XT	136 Avg FPS, 52 1% Low FPS
RTX 5070 Ti	128 Avg FPS, 58 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

197

90

RTX 5070

144

65

RX 9070 XT

221

107

RTX 5070 Ti

177

126

075150225300

The Talos Principle 2 (1440p Ultra +upscaling quality, 2x FG) Data

Product	Value
RX 9070	197 Avg FPS, 90 1% Low FPS
RTX 5070	144 Avg FPS, 65 1% Low FPS
RX 9070 XT	221 Avg FPS, 107 1% Low FPS
RTX 5070 Ti	177 Avg FPS, 126 1% Low FPS

Our detailed benchmarking shows just how close the RX 9070 gets to the far more expensive RTX 5070 Ti and RX 9070 XT, why it's our pick as the best current GPU to buy in 2026

4K gaming performance

Black Myth Wukong (4K High)Cyberpunk 2077 (4K RT Ultra)F1 24 (4K Ultra High)Homeworld 3 (4K Epic)Metro Exodus Enhanced Edition (4K Ultra)The Talos Principle 2 (4K Ultra)Total War: Warhammer 3 (4K Ultra)

Avg FPS

1% Low FPS

RX 9070

40

30

RTX 5070

37

30

RX 9070 XT

45

39

RTX 5070 Ti

47

33

015304560

Black Myth Wukong (4K High) Data

Product	Value
RX 9070	40 Avg FPS, 30 1% Low FPS
RTX 5070	37 Avg FPS, 30 1% Low FPS
RX 9070 XT	45 Avg FPS, 39 1% Low FPS
RTX 5070 Ti	47 Avg FPS, 33 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

20

17

RTX 5070

21

17

RX 9070 XT

25

21

RTX 5070 Ti

30

24

07.51522.530

Cyberpunk 2077 (4K RT Ultra) Data

Product	Value
RX 9070	20 Avg FPS, 17 1% Low FPS
RTX 5070	21 Avg FPS, 17 1% Low FPS
RX 9070 XT	25 Avg FPS, 21 1% Low FPS
RTX 5070 Ti	30 Avg FPS, 24 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

44

32

RTX 5070

39

30

RX 9070 XT

54

39

RTX 5070 Ti

54

41

015304560

F1 24 (4K Ultra High) Data

Product	Value
RX 9070	44 Avg FPS, 32 1% Low FPS
RTX 5070	39 Avg FPS, 30 1% Low FPS
RX 9070 XT	54 Avg FPS, 39 1% Low FPS
RTX 5070 Ti	54 Avg FPS, 41 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

72

56

RTX 5070

79

55

RX 9070 XT

77

60

RTX 5070 Ti

89

53

0255075100

Homeworld 3 (4K Epic) Data

Product	Value
RX 9070	72 Avg FPS, 56 1% Low FPS
RTX 5070	79 Avg FPS, 55 1% Low FPS
RX 9070 XT	77 Avg FPS, 60 1% Low FPS
RTX 5070 Ti	89 Avg FPS, 53 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

62

47

RTX 5070

57

44

RX 9070 XT

73

50

RTX 5070 Ti

76

36

020406080

Metro Exodus Enhanced Edition (4K Ultra) Data

Product	Value
RX 9070	62 Avg FPS, 47 1% Low FPS
RTX 5070	57 Avg FPS, 44 1% Low FPS
RX 9070 XT	73 Avg FPS, 50 1% Low FPS
RTX 5070 Ti	76 Avg FPS, 36 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

39

24

RTX 5070

35

24

RX 9070 XT

44

32

RTX 5070 Ti

46

25

015304560

The Talos Principle 2 (4K Ultra) Data

Product	Value
RX 9070	39 Avg FPS, 24 1% Low FPS
RTX 5070	35 Avg FPS, 24 1% Low FPS
RX 9070 XT	44 Avg FPS, 32 1% Low FPS
RTX 5070 Ti	46 Avg FPS, 25 1% Low FPS

Avg FPS

1% Low FPS

RX 9070

52

36

RTX 5070

44

30

RX 9070 XT

60

41

RTX 5070 Ti

56

41

015304560

Total War: Warhammer 3 (4K Ultra) Data

Product	Value
RX 9070	52 Avg FPS, 36 1% Low FPS
RTX 5070	44 Avg FPS, 30 1% Low FPS
RX 9070 XT	60 Avg FPS, 41 1% Low FPS
RTX 5070 Ti	56 Avg FPS, 41 1% Low FPS

Our detailed benchmarking shows just how close the RX 9070 gets to the far more expensive RTX 5070 Ti and RX 9070 XT, why it's our pick as the best current GPU to buy in 2026

Synthetic performance

3DMark Port Royal3DMark Time Spy ExtremeBlender 4.3.0 (Junkyard)Procyon generative image creation

RX 9070

15596

RTX 5070

14076

RX 9070 XT

18071

RTX 5070 Ti

19212

05,00010,00015,00020,000

3DMark Port Royal Data

Product	Value
RX 9070	15596
RTX 5070	14076
RX 9070 XT	18071
RTX 5070 Ti	19212

GPU index score

Overall index score

RX 9070

12647

11432

RTX 5070

10670

9967

RX 9070 XT

14315

12555

RTX 5070 Ti

13534

12049

03,7507,50011,25015,000

3DMark Time Spy Extreme Data

Product	Value
RX 9070	12647 GPU index score, 11432 Overall index score
RTX 5070	10670 GPU index score, 9967 Overall index score
RX 9070 XT	14315 GPU index score, 12555 Overall index score
RTX 5070 Ti	13534 GPU index score, 12049 Overall index score

RX 9070

1031.81

RTX 5070

1665.98

RX 9070 XT

1135.6

RTX 5070 Ti

1954.78

05001,0001,5002,000

Blender 4.3.0 (Junkyard) Data

Product	Value
RX 9070	1031.81
RTX 5070	1665.98
RX 9070 XT	1135.6
RTX 5070 Ti	1954.78

Index score

Duration (sec)

RX 9070

2272

43.997

RTX 5070

2889

34.604

RX 9070 XT

2593

38.554

RTX 5070 Ti

3815

26.211

01,0002,0003,0004,000

Procyon generative image creation Data

Product	Value
RX 9070	2272 Index score, 43.997 Duration (sec)
RTX 5070	2889 Index score, 34.604 Duration (sec)
RX 9070 XT	2593 Index score, 38.554 Duration (sec)
RTX 5070 Ti	3815 Index score, 26.211 Duration (sec)

Our detailed benchmarking shows just how close the RX 9070 gets to the far more expensive RTX 5070 Ti and RX 9070 XT, why it's our pick as the best current GPU to buy in 2026

System performance

GPU frequencyPerformance per wattPowerThermals

RX 9070

2349

RTX 5070

2693

RX 9070 XT

2657

RTX 5070 Ti

2692

07501,5002,2503,000

GPU frequency Data

Product	Value
RX 9070	2349
RTX 5070	2693
RX 9070 XT	2657
RTX 5070 Ti	2692

4K frames per J

1080p frame per J

RX 9070

0.26

0.6

RTX 5070

0.24

0.61

RX 9070 XT

0.23

0.51

RTX 5070 Ti

0.26

0.65

00.20.40.60.8

Performance per watt Data

Product	Value
RX 9070	0.26 4K frames per J, 0.6 1080p frame per J
RTX 5070	0.24 4K frames per J, 0.61 1080p frame per J
RX 9070 XT	0.23 4K frames per J, 0.51 1080p frame per J
RTX 5070 Ti	0.26 4K frames per J, 0.65 1080p frame per J

Peak power (watts)

Average power (watts)

RX 9070

270

248

RTX 5070

284

251

RX 9070 XT

366

326

RTX 5070 Ti

342

303

0100200300400

Power Data

Product	Value
RX 9070	270 Peak power (watts), 248 Average power (watts)
RTX 5070	284 Peak power (watts), 251 Average power (watts)
RX 9070 XT	366 Peak power (watts), 326 Average power (watts)
RTX 5070 Ti	342 Peak power (watts), 303 Average power (watts)

Average temperature (°C)

RX 9070

57

55

RTX 5070

79

76

RX 9070 XT

59

56

RTX 5070 Ti

69

65

020406080

Thermals Data

Product	Value
RX 9070	57 Peak temperature (°C), 55 Average temperature (°C)
RTX 5070	79 Peak temperature (°C), 76 Average temperature (°C)
RX 9070 XT	59 Peak temperature (°C), 56 Average temperature (°C)
RTX 5070 Ti	69 Peak temperature (°C), 65 Average temperature (°C)

Our detailed benchmarking shows just how close the RX 9070 gets to the far more expensive RTX 5070 Ti and RX 9070 XT, why it's our pick as the best current GPU to buy in 2026

The RX 9070 is a rather agonising pick as the best graphics card around right now. There are inevitable caveats around my decision, because quite obviously this second-tier RDNA 4 GPU is not the most powerful GPU you can buy. The AMD Radeon RX 9070, however, is the graphics card I think makes the most sense to the most PC gamers.

And, realistically, it's going to be the one that I would suggest to my PC gaming friends if they come looking for a recommendation. Luckily, I don't actually have friends.

The RX 9070 offers close to the performance of AMD's most powerful RDNA 4 GPU, still packs in 16 GB of GDDR6 video memory, can be easily overclocked without demanding much more cooling or power, and handily outperforms Nvidia's similarly priced, lower-spec RTX 5070.

I was relatively cool on the card in my initial review of the GPU, mainly because our first taste of it was a card that XFX hiked the price of out the gate, which made it more than a reference RX 9070 XT. As things have evened out, there are reliably cheaper RX 9070 cards around and when you look at the graphics card market as a whole, the Radeon stands out as a great pick.

Yet I still haven't come to that conclusion lightly. My first thought was going to be recommending the $350 RX 9060 XT, because it is great value, performs well, and pretty much offers performance parity with the more expensive Nvidia option. But it doesn't feel like a great recommendation for the best graphics card, given that it is a third-tier Radeon that isn't particularly exciting.

The Radeon RX 9070 XT, though, is a fantastic card and were it not for the ludicrous over-pricing of AMD's finest it would absolutely be my top pick. With an MSRP of $599 it would be an easy choice for its impressive performance out of the box, stunning under-volting frame rate bumps, and the excellent filip of the new FSR4 upscaler. But the cheapest you can get it for these days is still over $700, with the RX 9070 generally under $600 with very similar gaming performance, especially with some judicious undervolting.

In the end, the extra raster performance of the Radeon RX 9070 over the RTX 5070—and how close it can get to the more expensive RTX 5070 Ti and RX 9070 XT GPUs—makes it the best graphics card to buy right now for most PC gamers.

Read our full AMD Radeon RX 9070 review.

2. Best value graphics card: AMD Radeon RX 9060 XT 16 GB

Previous

1 / 5

Next

Native 1080p gaming performance

Black Myth Wukong (1080p High)Cyberpunk 2077 (1080p RT Ultra)F1 24 (1080p Ultra high)Homeworld 3 (1080p Epic)Metro Exodus Enhanced Edition (1080p Ultra)The Talos Principle 2 (1080p Ultra)Total War: Warhammer 3 (1080p Ultra)

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

73

63

RX 9060 XT 8 GB

71

58

RTX 5060 Ti 16 GB

75

65

RTX 5070 12 GB

95

82

0255075100

Black Myth Wukong (1080p High) Data

Product	Value
RX 9060 XT 16 GB	73 Avg FPS, 63 1% Low FPS
RX 9060 XT 8 GB	71 Avg FPS, 58 1% Low FPS
RTX 5060 Ti 16 GB	75 Avg FPS, 65 1% Low FPS
RTX 5070 12 GB	95 Avg FPS, 82 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

49

39

RX 9060 XT 8 GB

41

26

RTX 5060 Ti 16 GB

51

41

RTX 5070 12 GB

72

56

020406080

Cyberpunk 2077 (1080p RT Ultra) Data

Product	Value
RX 9060 XT 16 GB	49 Avg FPS, 39 1% Low FPS
RX 9060 XT 8 GB	41 Avg FPS, 26 1% Low FPS
RTX 5060 Ti 16 GB	51 Avg FPS, 41 1% Low FPS
RTX 5070 12 GB	72 Avg FPS, 56 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

95

61

RX 9060 XT 8 GB

82

66

RTX 5060 Ti 16 GB

90

71

RTX 5070 12 GB

123

96

037.575112.5150

F1 24 (1080p Ultra high) Data

Product	Value
RX 9060 XT 16 GB	95 Avg FPS, 61 1% Low FPS
RX 9060 XT 8 GB	82 Avg FPS, 66 1% Low FPS
RTX 5060 Ti 16 GB	90 Avg FPS, 71 1% Low FPS
RTX 5070 12 GB	123 Avg FPS, 96 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

106

65

RX 9060 XT 8 GB

111

63

RTX 5060 Ti 16 GB

117

65

RTX 5070 12 GB

125

64

037.575112.5150

Homeworld 3 (1080p Epic) Data

Product	Value
RX 9060 XT 16 GB	106 Avg FPS, 65 1% Low FPS
RX 9060 XT 8 GB	111 Avg FPS, 63 1% Low FPS
RTX 5060 Ti 16 GB	117 Avg FPS, 65 1% Low FPS
RTX 5070 12 GB	125 Avg FPS, 64 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

96

42

RX 9060 XT 8 GB

93

64

RTX 5060 Ti 16 GB

98

42

RTX 5070 12 GB

139

88

037.575112.5150

Metro Exodus Enhanced Edition (1080p Ultra) Data

Product	Value
RX 9060 XT 16 GB	96 Avg FPS, 42 1% Low FPS
RX 9060 XT 8 GB	93 Avg FPS, 64 1% Low FPS
RTX 5060 Ti 16 GB	98 Avg FPS, 42 1% Low FPS
RTX 5070 12 GB	139 Avg FPS, 88 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

73

37

RX 9060 XT 8 GB

66

35

RTX 5060 Ti 16 GB

75

41

RTX 5070 12 GB

92

57

0255075100

The Talos Principle 2 (1080p Ultra) Data

Product	Value
RX 9060 XT 16 GB	73 Avg FPS, 37 1% Low FPS
RX 9060 XT 8 GB	66 Avg FPS, 35 1% Low FPS
RTX 5060 Ti 16 GB	75 Avg FPS, 41 1% Low FPS
RTX 5070 12 GB	92 Avg FPS, 57 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

94

72

RX 9060 XT 8 GB

90

42

RTX 5060 Ti 16 GB

83

54

RTX 5070 12 GB

104

53

037.575112.5150

Total War: Warhammer 3 (1080p Ultra) Data

Product	Value
RX 9060 XT 16 GB	94 Avg FPS, 72 1% Low FPS
RX 9060 XT 8 GB	90 Avg FPS, 42 1% Low FPS
RTX 5060 Ti 16 GB	83 Avg FPS, 54 1% Low FPS
RTX 5070 12 GB	104 Avg FPS, 53 1% Low FPS

Testing the 16 GB version of the AMD RX 9060 XT graphics card against the competition to show why it's our pick as the best value GPU of 2026

Native 1440p gaming performance

Black Myth Wukong (1440p High)Cyberpunk 2077 (1440p RT Ultra)F1 24 (1440p Ultra High)Homeworld 3 (1440p Epic)Metro Exodus Enhanced Edition (1440p Ultra)The Talos Principle 2 (1440p Ultra)Total War: Warhammer 3 (1440p Ultra)

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

48

44

RX 9060 XT 8 GB

48

43

RTX 5060 Ti 16 GB

51

45

RTX 5070 12 GB

68

60

020406080

Black Myth Wukong (1440p High) Data

Product	Value
RX 9060 XT 16 GB	48 Avg FPS, 44 1% Low FPS
RX 9060 XT 8 GB	48 Avg FPS, 43 1% Low FPS
RTX 5060 Ti 16 GB	51 Avg FPS, 45 1% Low FPS
RTX 5070 12 GB	68 Avg FPS, 60 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

30

24

RX 9060 XT 8 GB

26

17

RTX 5060 Ti 16 GB

32

26

RTX 5070 12 GB

45

36

015304560

Cyberpunk 2077 (1440p RT Ultra) Data

Product	Value
RX 9060 XT 16 GB	30 Avg FPS, 24 1% Low FPS
RX 9060 XT 8 GB	26 Avg FPS, 17 1% Low FPS
RTX 5060 Ti 16 GB	32 Avg FPS, 26 1% Low FPS
RTX 5070 12 GB	45 Avg FPS, 36 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

60

46

RX 9060 XT 8 GB

51

41

RTX 5060 Ti 16 GB

56

45

RTX 5070 12 GB

78

60

020406080

F1 24 (1440p Ultra High) Data

Product	Value
RX 9060 XT 16 GB	60 Avg FPS, 46 1% Low FPS
RX 9060 XT 8 GB	51 Avg FPS, 41 1% Low FPS
RTX 5060 Ti 16 GB	56 Avg FPS, 45 1% Low FPS
RTX 5070 12 GB	78 Avg FPS, 60 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

86

65

RX 9060 XT 8 GB

75

58

RTX 5060 Ti 16 GB

97

64

RTX 5070 12 GB

112

64

037.575112.5150

Homeworld 3 (1440p Epic) Data

Product	Value
RX 9060 XT 16 GB	86 Avg FPS, 65 1% Low FPS
RX 9060 XT 8 GB	75 Avg FPS, 58 1% Low FPS
RTX 5060 Ti 16 GB	97 Avg FPS, 64 1% Low FPS
RTX 5070 12 GB	112 Avg FPS, 64 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

71

34

RX 9060 XT 8 GB

68

49

RTX 5060 Ti 16 GB

72

34

RTX 5070 12 GB

100

68

0255075100

Metro Exodus Enhanced Edition (1440p Ultra) Data

Product	Value
RX 9060 XT 16 GB	71 Avg FPS, 34 1% Low FPS
RX 9060 XT 8 GB	68 Avg FPS, 49 1% Low FPS
RTX 5060 Ti 16 GB	72 Avg FPS, 34 1% Low FPS
RTX 5070 12 GB	100 Avg FPS, 68 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

49

35

RX 9060 XT 8 GB

47

27

RTX 5060 Ti 16 GB

51

32

RTX 5070 12 GB

64

42

020406080

The Talos Principle 2 (1440p Ultra) Data

Product	Value
RX 9060 XT 16 GB	49 Avg FPS, 35 1% Low FPS
RX 9060 XT 8 GB	47 Avg FPS, 27 1% Low FPS
RTX 5060 Ti 16 GB	51 Avg FPS, 32 1% Low FPS
RTX 5070 12 GB	64 Avg FPS, 42 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

62

45

RX 9060 XT 8 GB

60

44

RTX 5060 Ti 16 GB

60

44

RTX 5070 12 GB

81

53

0255075100

Total War: Warhammer 3 (1440p Ultra) Data

Product	Value
RX 9060 XT 16 GB	62 Avg FPS, 45 1% Low FPS
RX 9060 XT 8 GB	60 Avg FPS, 44 1% Low FPS
RTX 5060 Ti 16 GB	60 Avg FPS, 44 1% Low FPS
RTX 5070 12 GB	81 Avg FPS, 53 1% Low FPS

Testing the 16 GB version of the AMD RX 9060 XT graphics card against the competition to show why it's our pick as the best value GPU of 2026

Upscaling performance

Black Myth Wukong (1440p High + Upscaling (quality))Cyberpunk 2077 (1440p RT Ultra + upscaling (quality))F1 24 (1440p Ultra high +upscaling (Quality), 2x FG)Homeworld 3 (1440p Epic + upscaling (quality))The Talos Principle 2 (1440p Ultra +upscaling (quality), 2x FG)

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

78

69

RX 9060 XT 8 GB

76

68

RTX 5060 Ti 16 GB

86

72

RTX 5070 12 GB

107

81

037.575112.5150

Black Myth Wukong (1440p High + Upscaling (quality)) Data

Product	Value
RX 9060 XT 16 GB	78 Avg FPS, 69 1% Low FPS
RX 9060 XT 8 GB	76 Avg FPS, 68 1% Low FPS
RTX 5060 Ti 16 GB	86 Avg FPS, 72 1% Low FPS
RTX 5070 12 GB	107 Avg FPS, 81 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

102

73

RX 9060 XT 8 GB

90

60

RTX 5060 Ti 16 GB

96

79

RTX 5070 12 GB

122

97

037.575112.5150

Cyberpunk 2077 (1440p RT Ultra + upscaling (quality)) Data

Product	Value
RX 9060 XT 16 GB	102 Avg FPS, 73 1% Low FPS
RX 9060 XT 8 GB	90 Avg FPS, 60 1% Low FPS
RTX 5060 Ti 16 GB	96 Avg FPS, 79 1% Low FPS
RTX 5070 12 GB	122 Avg FPS, 97 1% Low FPS

1% Low FPS

RX 9060 XT 16 GB

190

113

RX 9060 XT 8 GB

146

122

RTX 5060 Ti 16 GB

136

89

RTX 5070 12 GB

176

113

050100150200

F1 24 (1440p Ultra high +upscaling (Quality), 2x FG) Data

Product	Value
RX 9060 XT 16 GB	190 Avg FPS, 113 1% Low FPS
RX 9060 XT 8 GB	146 Avg FPS, 122 1% Low FPS
RTX 5060 Ti 16 GB	136 Avg FPS, 89 1% Low FPS
RTX 5070 12 GB	176 Avg FPS, 113 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

110

62

RX 9060 XT 8 GB

102

51

RTX 5060 Ti 16 GB

118

63

RTX 5070 12 GB

127

63

037.575112.5150

Homeworld 3 (1440p Epic + upscaling (quality)) Data

Product	Value
RX 9060 XT 16 GB	110 Avg FPS, 62 1% Low FPS
RX 9060 XT 8 GB	102 Avg FPS, 51 1% Low FPS
RTX 5060 Ti 16 GB	118 Avg FPS, 63 1% Low FPS
RTX 5070 12 GB	127 Avg FPS, 63 1% Low FPS

Avg FPS

1% Low FPS

RX 9060 XT 16 GB

135

59

RX 9060 XT 8 GB

127

78

RTX 5060 Ti 16 GB

115

41

RTX 5070 12 GB

144

65

037.575112.5150

The Talos Principle 2 (1440p Ultra +upscaling (quality), 2x FG) Data

Product	Value
RX 9060 XT 16 GB	135 Avg FPS, 59 1% Low FPS
RX 9060 XT 8 GB	127 Avg FPS, 78 1% Low FPS
RTX 5060 Ti 16 GB	115 Avg FPS, 41 1% Low FPS
RTX 5070 12 GB	144 Avg FPS, 65 1% Low FPS

Testing the 16 GB version of the AMD RX 9060 XT graphics card against the competition to show why it's our pick as the best value GPU of 2026

Synthetic performance

3DMark Time Spy Extreme3DMark Port RoyalBlender 4.3.0 (Junkshop)Procyon generative image creation

GPU index score

Overall index score

RX 9060 XT 16 GB

7664

7631

RX 9060 XT 8 GB

7553

7499

RTX 5060 Ti 16 GB

7261

7287

RTX 5070 12 GB

10670

9967

03,7507,50011,25015,000

3DMark Time Spy Extreme Data

Product	Value
RX 9060 XT 16 GB	7664 GPU index score, 7631 Overall index score
RX 9060 XT 8 GB	7553 GPU index score, 7499 Overall index score
RTX 5060 Ti 16 GB	7261 GPU index score, 7287 Overall index score
RTX 5070 12 GB	10670 GPU index score, 9967 Overall index score

RX 9060 XT 16 GB

9779

RX 9060 XT 8 GB

9660

RTX 5060 Ti 16 GB

10173

RTX 5070 12 GB

14076

03,7507,50011,25015,000

Overall index score

3DMark Port Royal Data

Product	Value
RX 9060 XT 16 GB	9779
RX 9060 XT 8 GB	9660
RTX 5060 Ti 16 GB	10173
RTX 5070 12 GB	14076

RX 9060 XT 16 GB

423.07

RTX 5060 Ti 16 GB

1124.94

RTX 5070 12 GB

1665.98

05001,0001,5002,000

Samples per min

Blender 4.3.0 (Junkshop) Data

Product	Value
RX 9060 XT 16 GB	423.07
RX 9060 XT 8 GB	N/A
RTX 5060 Ti 16 GB	1124.94
RTX 5070 12 GB	1665.98

Index score

Duration (sec)

Sec per image

RX 9060 XT 16 GB

1472

4.245

4.245

RX 9060 XT 8 GB

1382

72.338

4.521

RTX 5060 Ti 16 GB

2018

2018

2018

RTX 5070 12 GB

2889

34.604

2.163

07501,5002,2503,000

Procyon generative image creation Data

Product	Value
RX 9060 XT 16 GB	1472 Index score, 4.245 Duration (sec), 4.245 Sec per image
RX 9060 XT 8 GB	1382 Index score, 72.338 Duration (sec), 4.521 Sec per image
RTX 5060 Ti 16 GB	2018 Index score, 2018 Duration (sec), 2018 Sec per image
RTX 5070 12 GB	2889 Index score, 34.604 Duration (sec), 2.163 Sec per image

Testing the 16 GB version of the AMD RX 9060 XT graphics card against the competition to show why it's our pick as the best value GPU of 2026

System performance

GPU frequencyPerformance per wattPowerThermals

RX 9060 XT 16 GB

2849

RX 9060 XT 8 GB

2945

RTX 5060 Ti 16 GB

2687

RTX 5070 12 GB

2693

07501,5002,2503,000

Average clock speed (MHz)

GPU frequency Data

Product	Value
RX 9060 XT 16 GB	2849
RX 9060 XT 8 GB	2945
RTX 5060 Ti 16 GB	2687
RTX 5070 12 GB	2693

4K frames per J

1080p frame per J

RX 9060 XT 8 GB

0.22

0.56

RTX 5060 Ti 16 GB

0.23

0.62

RTX 5070 12 GB

0.24

0.61

00.20.40.60.8

Performance per watt Data

Product	Value
RX 9060 XT 16 GB	N/A
RX 9060 XT 8 GB	0.22 4K frames per J, 0.56 1080p frame per J
RTX 5060 Ti 16 GB	0.23 4K frames per J, 0.62 1080p frame per J
RTX 5070 12 GB	0.24 4K frames per J, 0.61 1080p frame per J

Peak power (watts)

Average power (watts)

RX 9060 XT 16 GB

182

146

RX 9060 XT 8 GB

175

164

RTX 5060 Ti 16 GB

207

182

RTX 5070 12 GB

284

251

075150225300

Power Data

Product	Value
RX 9060 XT 16 GB	182 Peak power (watts), 146 Average power (watts)
RX 9060 XT 8 GB	175 Peak power (watts), 164 Average power (watts)
RTX 5060 Ti 16 GB	207 Peak power (watts), 182 Average power (watts)
RTX 5070 12 GB	284 Peak power (watts), 251 Average power (watts)

Peak temperature (°C)

Average temperature (°C)

RX 9060 XT 16 GB

57

48

RX 9060 XT 8 GB

51

48

RTX 5060 Ti 16 GB

67

65

RTX 5070 12 GB

79

76

020406080

Thermals Data

Product	Value
RX 9060 XT 16 GB	57 Peak temperature (°C), 48 Average temperature (°C)
RX 9060 XT 8 GB	51 Peak temperature (°C), 48 Average temperature (°C)
RTX 5060 Ti 16 GB	67 Peak temperature (°C), 65 Average temperature (°C)
RTX 5070 12 GB	79 Peak temperature (°C), 76 Average temperature (°C)

Testing the 16 GB version of the AMD RX 9060 XT graphics card against the competition to show why it's our pick as the best value GPU of 2026