Nvidia CEO Jensen Huang made a bold claim onstage at CES earlier this week when he was introducing the next-generation RTX 50-series GPUs. “The RTX 5070, 4090 performance at $549,” said Huang. It’s a claim that’s been echoed on YouTube, TikTok, and social media networks and has generated a debate over the RTX 50 series and DLSS 4’s Multi Frame Generation.
So, can a $549 RTX 5070 really deliver the same level of performance as a $1,599 RTX 4090? The answer is yes and no, and it all comes down to a “fake frames” argument about DLSS Frame Generation that might not even be a big problem for a lot of PC gamers.
Nvidia’s big RTX 5070 claim is all based on its latest generation of DLSS. “Impossible without artificial intelligence,” admits Huang after promising that the RTX 5070 can deliver RTX 4090 levels of performance. DLSS 4 has a new Multi Frame Generation technique that can generate up to three additional frames per every traditionally rendered frame.
Some PC gamers have long argued that this technique, which Nvidia introduced originally with DLSS 3, is simply “fake frames” and not reflective of the true performance of GPUs that have been using rasterization for decades.
Xbox chief Phil Spencer has been dropping hints about an Xbox handheld for months, but what about Windows handheld gaming PCs? Jason Ronald, Microsoft’s VP of “Next Generation,” tells The Verge that we should expect to see the Windows handheld gaming experience change within this calendar year.
Ronald was a roundtable panelist this evening at an AMD and Lenovo event titled “The Future of Gaming Handhelds,” which was mostly a coming-out party for Lenovo’s new Legion Go S. But he did hint onstage that Microsoft plans to bring the Xbox experience to Windows PCs, rather than the other way around — and expanded on that considerably after we caught up with him later.
“We’ve been really innovating for a long time in the console space, and as we partner across the industry it’s really about how do we bring those innovations that we’ve incubated and developed in the console space and bring them to PC and bring them to the handheld gaming space,” Ronald said.
When we caught up with him after the event, he confirmed that Microsoft is looking at combining Xbox and Windows experiences together — and that we should see changes this very year, rather than needing to wait for an Xbox handheld that might still be years away.
Photo by Sean Hollister / The Verge
Microsoft’s Jason Ronald speaks at the Lenovo / AMD event, flanked by Valve SteamOS designer Pierre-Loup Griffais (left) and AMD chief gaming architect Frank Azor (middle right).
“I would say it’s bringing the best of Xbox and Windows together, because we have spent the last 20 years building a world-class operating system, but it’s really locked to the console,” says Ronald. “What we’re doing is we’re really focused on how do we bring those experiences for both players and developers to the broader Windows ecosystem.”
Right now, Windows sucks on handhelds, to put it rather bluntly, to the point that a community-created fork of Valve’s SteamOS experience can be a far better way to pick up and play games. Ronald is clearly aware of the issues. “We’re focused on really simplifying that and making it much more like a console experience. Our goal is to put the player and their library at the center of the experience and not all the [Windows] work that you have to do today.”
Microsoft has done compact modes for Xbox apps on Windows that are focused on improving the handheld experience, but it’s a lot like putting lipstick on a pig instead of addressing the core experience. “I think we’ll have a lot more to share later this year,” teases Ronald. “I think it’s going to be a journey and I think you’ll see a lot of investments over time that you’re starting to see already, but we’ll have a lot more to share later this year.”
Photo by Tom Warren / The Verge
Microsoft has made several Xbox apps more handheld-friendly over the past year.
How Microsoft goes about this merging of Xbox and Windows will be key, but it doesn’t sound like the company is suddenly going to port a custom Xbox operating system to Windows. It sounds more like Microsoft wants to make Windows better at gaming all up with an Xbox experience on top so the pesky desktop, notifications, and legacy of Windows is hidden away.
“I think, at the end of the day, our goal is to make Windows great for gaming on any device,” says Ronald. “The reality is the Xbox operating system is built on top of Windows. So there’s a lot of infrastructure that we built in the console space that we can bring to the PC space and really deliver that premium gaming experience on any device.”
Specifically, Microsoft has to tackle a lot of the very basics of making Windows more friendly to controllers and getting that Xbox experience to really drive things instead of the taskbar, Start menu, and other elements. “There’s just certain things in Windows that were not designed for if you don’t have a keyboard and mouse, like thumbstick support or joypads and stuff like that,” Ronald admits.
“There’s fundamental interaction models that we’re working on to make sure that regardless of the operating system details it feels very natively like a gaming-centric device and a gaming-centric experience.”
Ronald says the goal is to put an Xbox experience at the center — “not the Windows desktop that you have today.”
Qualcomm teased it was about to enter the mini desktop PC space yesterday, and now Lenovo is announcing two mini PCs that cater to consumers and businesses. Both the ThinkCentre neo 50q and IdeaCentre Mini x will include Qualcomm’s Snapdragon X series chips, enabling Copilot Plus PC features in a desktop form factor.
The IdeaCentre Mini x is a 1-liter mini PC that’s designed for creatives who want a desktop that’s quiet and capable of productivity tasks. It comes with a built-in power supply and an easy-to-open case so you can swap components. It’s not a fan-less mini PC, though, as it has two fans inside to keep things cool.
Image: Lenovo
Lenovo’s IdeaCentre Mini x has Qualcomm’s latest Snapdragon X series chips inside.
At the front, there’s a single USB-A 3.2 Gen 2 port and a single USB-C 3.2 Gen 2 port, alongside a headphone jack and a power button. At the rear, Lenovo has equipped its IdeaCentre Mini x PC with two USB-A 3.2 Gen 2 ports, a single USB-A 2.0 port, a USB-C 4 port, a HDMI 2.1 port, a DisplayPort 1.4 connection, and an ethernet port.
The IdeaCentre Mini x can be configured with either a base Snapdragon X chip or the Snapdragon X Plus, up to 32GB of RAM, and up to 1TB of storage. Lenovo is using the latest Wi-Fi 7 connectivity inside the IdeaCentre Mini x, too.
Image: Lenovo
Lenovo’s ThinkCentre neo 50q is designed for small and medium businesses.
Lenovo’s ThinkCentre neo 50q is designed for small and medium businesses, and it has a similar set of specs. The ThinkCentre neo 50q will also be powered by Qualcomm’s Snapdragon X chip or the X Plus, with up to 16GB of RAM instead of the 32GB found on the IdeaCentre Mini x. At the front of the ThinkCentre mini PC, there is a single USB-A 3.2 Gen 2 port, a single USB-C 3.2 Gen 2 port, and a headphone jack and power button. At the rear, there are two USB-A 3.2 Gen 2 ports, two USB-A 2.0 ports, a HDMI 2.1 port, a DisplayPort 1.4 connection, and an ethernet port. Lenovo has also equipped the ThinkCentre neo 50q with Wi-Fi 6E connectivity, instead of the Wi-Fi 7 support available on the IdeaCentre Mini x.
While the IdeaCentre Mini x looks like the better option both in terms of optional specs and design, it will also be priced lower than the ThinkCentre mini PC. The IdeaCentre Mini x will be available starting in April, with an expected starting price of $659.99. Lenovo’s ThinkCentre neo 50q QC will be available slightly earlier in February, with a starting price expected at $849.
Warhammer 40,000: Darktide running on a GeForce RTX 5090 at 4K | Image: Nvidia
Nvidia is revealing a big upgrade to its Deep Learning Super Sampling (DLSS) technology today. DLSS 4 will include new neural rendering capabilities that, on systems with the new RTX 50 Series GPUs, can do Multi Frame Generation, generating “up to three additional frames per traditionally rendered frame, working in unison with the complete suite of DLSS technologies to multiply frame rates by up to 8X over traditional brute-force rendering.”
According to Nvidia, that’s a big enough upgrade to make 4K 240fps, fully ray-traced gaming possible. Also, in an upgrade that will work on all GeForce RTX GPUs, DLSS games with Ray Reconstruction, Super Resolution, and DLAA can be updated to new transformer AI models that use the same tech as AI tools like ChatGPT.
Nvidia says its new frame generation model is 40 percent faster and uses 30 percent less VRAM than the old one.
Here’s a slide showing which features are available on which series of RTX GPUs:
Image: Nvidia
DLSS 4 arrives more than two years after Nvidia introduced DLSS 3 with Frame Generation to boost performance in a variety of games. Nvidia then introduced DLSS 3.5 in August 2023 with an AI-powered Ray Reconstruction technique to improve the quality of ray tracing and introduce path tracing (full ray tracing) in select titles. In August last year, Nvidia revealed it now has more than 600 games and applications with RTX support.
At launch, Nvidia says there will be 75 games and apps that support Multi Frame Generation, with Alan Wake 2, Cyberpunk 2077, Indiana Jones and the Great Circle, and Star Wars Outlaws supporting it when RTX 50 Series GPUs launch.
Nvidia is officially announcing its RTX 50-series GPUs today. After months of leaks and rumors, the next-generation RTX Blackwell GPUs are now official, and there are four of them on the way.
Nvidia CEO Jensen Huang revealed the RTX 50-series GPUs during a CES keynote this evening, announcing a $1,999 RTX 5090, a $999 RTX 5080, a $749 RTX 5070 Ti, and a $549 RTX 5070. Nvidia’s new RTX 5090 and RTX 5080 GPUs will both be available on January 30th, with the RTX 5070 Ti and RTX 5070 to follow in February.
Image: Nvidia
Nvidia’s RTX 5090 GPU is surprisingly small.
The RTX 50-series GPUs include a new design for the Founders Edition, with just two double flow-through fans, a 3D vapor chamber, and GDDR7 memory. All of the RTX 50-series cards are PCIe Gen 5 and include DisplayPort 2.1b connectors to drive displays up to 8K and 165Hz.
Surprisingly, the RTX 5090 Founders Edition will be a two-slot GPU and will be capable of fitting inside small form factor PCs — a big departure from the size of the RTX 4090. The RTX 5090 has 32GB of GDDR7, a memory bandwidth of 1,792GB/sec, and a massive 21,760 CUDA cores.
This all adds up to a GPU that Nvidia says will be two times faster than the RTX 4090, thanks to DLSS 4 and the Blackwell architecture. But it will come at a cost of power consumption, as Nvidia says the RTX 5090 will have a total graphics power of 575 watts and a recommended PSU requirement of 1000 watts. That’s 125 watts more than the RTX 4090, but hopefully the RTX 5090 will be a lot more power efficient so that you’ll rarely be using the full 575 watts.
Nvidia demonstrated Cyberpunk 2077 running on an RTX 5090 with DLSS 4 at 238fps compared to 106fps on an RTX 4090 with DLSS 3.5. Both GPUs are running the game with full ray tracing enabled.
The RTX 5080 is designed to be twice as fast as the RTX 4080 and will include 16GB of GDDR7 memory, a memory bandwidth of 960GB/sec, and 10,752 CUDA cores. The RTX 5080 will have a total graphics power of 360 watts and Nvidia is recommending a 850-watt power supply. Nvidia is promising big performance gains with the RTX 5080 over the previous RTX 4080 model as a result of these specs.
Nvidia is also launching an RTX 5070 Ti and RTX 5070. The RTX 5070 Ti includes 16GB of GDDR7 memory, a memory bandwidth of 896GB/s, and 8,960 CUDA cores. The RTX 5070 has 12GB of GDDR7, a memory bandwidth of 672 GB/sec, and 6,144 CUDA cores. The RTX 5070 Ti will have a total graphics power of 300 watts and require a 750-watt PSU, while the RTX 5070 has a total graphics power of 250 watts and only needs a 650-watt PSU.
Nvidia says the RTX 5070 Ti will be 2x faster than the RTX 4070 Ti, and the RTX 5070 should be twice as fast as the RTX 4070. Huang even claimed on stage at CES that the RTX 5070 will deliver “RTX 4090 performance at $549,” but this will undoubtedly be because of DLSS 4 improvements and not pure rasterization performance.
Nvidia is also bringing its RTX 50-series to laptops, with the RTX 5090 laptop GPU debuting with 24GB of GDDR7 memory. The RTX 5080 laptop GPU will ship with 16GB of GDDR7 memory, the RTX 5070 Ti with 12GB of GDDR7 memory, and the RTX 5070 with just 8GB of GDDR7 memory. RTX 50-series laptops will be available starting in March from a variety of PC makers.
Image: Nvidia
RTX 50-series laptops are coming in March.
Huang demonstrated Nvidia’s RTX Blackwell GPUs with a real-time rendering demo at the beginning of the company’s CES keynote today. The demo included new RTX Neural Materials, RTX Neural Faces, text to animation, and even DLSS 4. “The new generation of DLSS can generate beyond frames, it can predict the future,” says Huang. “We used GeForce to enable AI, and now AI is revolutionizing GeForce.”
Nvidia’s new RTX Neural Shaders can be used to compress textures in games, while RTX Neural Faces aim to improve face quality using generative AI. The next generation of DLSS includes Multi Frame Generation, which generates up to three additional frames per traditional frame and can multiply frame rates by up to 8x over traditional rendering, according to Nvidia.
DLSS 4 also includes a real-time application of transformers to improve image quality, reduce ghosting, and add higher detail in motion. The DLSS 4 upgrade will even work on existing RTX GPUs, as features have been upgraded to the new transformer AI models. You can read more about DLSS 4 right here.
Nvidia’s RTX 50-series announcement comes more than two years after the RTX 4090 and RTX 4080 were announced, based on Nvidia’s Ada Lovelace architecture. Nvidia’s RTX 40-series of GPUs focused on improving ray tracing with Deep Learning Super Sampling (DLSS) version 3, and the RTX 4090 delivered some truly impressive performance gains over the previous RTX 3090 GPU.
AMD is previewing its next generation of GPUs at CES today, based on its latest RDNA 4 architecture that includes AI-powered FidelityFX Super Resolution 4 (FSR 4) upscaling. The Radeon RX 9070 XT and Radeon RX 9070 will both be available in Q1 from a variety of video card manufacturers, but AMD isn’t detailing specifications, pricing, or exact release dates just yet.
The announcement is light on concrete information. AMD says it has built this architecture from the ground up and that the GPUs built on RDNA 4 will include “a significant boost in AI.” AMD has optimized the compute units in RNDA 4, improved its ray-tracing engine and performance, and upgraded its media encoding quality. Built on a 4nm process, the RX 9070 XT and RX 9070 cards will include AMD’s second-generation AI accelerators, third-generation ray-tracing accelerators, and second-generation radiance display engine.
Image: AMD
AMD is only providing small details about its RDNA 4 architecture.
This RDNA 4 architecture will also enable AMD to launch FSR 4 with these 9000-series GPUs. FSR 4 is a machine learning-powered update to AMD’s upscaling and frame-generation technology that’s been developed specifically for RDNA 4 and its dedicated AI accelerator hardware. That means you’ll only be able to get FSR 4 with a Radeon RX 9070-series graphics card right now, and it will be supported in games with FSR 3.1 already integrated.
Matt Booty, president of game content and studios at Microsoft, appeared on stage at AMD’s CES keynote earlier today to confirm FSR 4 will be available in Call of Duty: Black Ops 6 later this year.
AMD hasn’t hinted at how FSR 4 will compare to Nvidia’s DLSS technology or what type of performance uplift or image quality improvement we should expect to see with this next generation of FSR.
It’s also not clear exactly what performance the RX 9070 series of GPUs will deliver against the competition, but in a branding slide for RDNA 4, AMD appears to suggest the 9070 series will offer similar performance to the Nvidia GeForce RTX 4070 Ti and RTX 4070 Super. Nvidia is expected to announce its RTX 50-series GPUs later today.
Image: AMD
FSR 4 will use dedicated AI accelerators on next-gen GPUs to upscale games.
AMD is also adding new AI-powered features to its Adrenalin software. These include the ability to generate images with AI models, summarize local documents, and ask an AMD chatbot questions about graphics settings and more.
If you’re wondering why AMD has jumped from the Radeon 7000 series straight to 9000-series GPUs, the company says it will use the 8000-series branding for RDNA 3.5 mobile GPUs. AMD is only previewing FSR 4 at CES and promises to provide more details on the upscaling technology and its latest RDNA 4 GPUs ahead of the Q1 launch.
AMD is unveiling its latest flagship desktop CPU, the Ryzen 9 9950X3D, today at CES. After the 9800X3D wowed us with its gaming performance, we’ve been waiting to see what AMD’s second generation of 3D V-Cache technology could do with more cores and higher boost frequencies. AMD is now claiming the 9950X3D will be “the world’s best processor for gamers and creators.”
The 9950X3D includes 16 Zen 5 CPU cores (32 threads), a max boost frequency of 5.7GHz, and 144MB of total cache. It has a higher TDP over the 9800X3D at 170 watts instead of 120 watts, but this extra power appears to translate to big improvements in creator benchmarks and gaming.
Image: AMD
AMD’s benchmarks against Intel’s Core Ultra 9 285K.
AMD says the 9950X3D should be around 8 percent faster on average than the previous 7950X3D, based on benchmarks run on 40 games at 1080p. The gaming performance should be similar to the 9800X3D, with AMD claiming it’s within 1 percent. AMD even claims the 9950X3D is 20 percent faster than Intel’s Core Ultra 9 285K in those same games, but Intel is about to deliver a performance update for the 285K that could close that gap.
In content creation tasks, the 9950X3D should be around 6 percent faster than the 7950X3D in Premiere Pro tasks and around 13 percent faster in Photoshop. On average, AMD says the 9950X3D will be 13 percent faster for creator tasks than the 7950X3D, based on 20 apps tested. The big claim from AMD is that the 9950X3D will be 10 percent faster than Intel’s Core Ultra 9 285K, something that could make the 9950X3D an easy pick for anyone wanting great desktop and gaming performance in a single CPU.
AMD is also launching a Ryzen 9 9900X3D processor with 12 cores (24 threads), a 5.5GHz max boost, 140MB of cache, and a 120-watt TDP. Much like the 9800X3D, both of these new X3D chips use the second generation of AMD’s 3D V-Cache technology that sees the cache now sit below the processor cores. This change allows the processor cores to have better access to cooling, and the cache is now less sensitive to high temperatures, allowing the CPU to boost to higher frequencies and deliver better performance in both gaming and creator tasks.
AMD says both the 9950X3D and 9900X3D will be available at some point in March, but it’s not detailing pricing just yet.
AMD is also announcing new “Fire Range” mobile CPUs today, which are the latest X3D-series laptop parts. You can read more about the flagship 9955HX3D laptop chip and other mobile-focused AMD announcements right here.
Qualcomm teased Copilot Plus mini PCs earlier today, but HP is already announcing its own Copilot Plus mini PC and all-in-one desktop PCs, and neither will be powered by Qualcomm. HP has opted for AMD’s new Ryzen AI Max chips for its mini PC and Intel’s latest Core Ultra 200V series processors for the all-in-one.
The OmniStudio X all-in-one will come in both 31.5-inch and 27-inch variants, with the ability to configure the PC with up to Intel Core Ultra 7 258V processors. These chips include a 47 TOPS NPU that’s powerful enough to handle Microsoft’s latest Copilot Plus PC features.
Photo by Antonio G. Di Benedetto / The Verge
HP’s 31.5-inch OmniStudio X all-in-one PC.
The 31.5-inch model comes with a 4K IPS panel that supports HDR 600 and up to 550 nits of brightness, but it’s not a touchscreen display. You’ll have to opt for the 27-inch model if you want a touchscreen, but HP has opted for a 1080p IPS display here that can hit up to 300 nits of brightness. The 27-inch model also comes in a non-touch variant at 1080p and even a 4K IPS option that doesn’t support HDR or touch and can reach up to 350 nits.
Both sizes come with built-in speakers and microphones and a five-megapixel IR camera above the display. The larger model even includes the ability to charge and control your laptop through a USB-C 20Gbps port at the rear that also supports DisplayPort in and out. Both also have 10Gbps USB-A and USB-C ports at the side, with a single USB-C 20Gbps port at the rear, alongside 2 USB-A 10Gbps ports, a single HDMI 1.4 out port, a single HDMI 2.1 in port, an ethernet port, and a headphone jack. Both also come with Wi-Fi 7 support.
Image: HP
The rear of HP’s OmniStudio X has a good selection of ports.
If you’re looking for a mini PC instead, HP’s new Z2 Mini G1a looks like it might be one of the most interesting miniature PCs at CES this year. HP is calling it the “world’s most powerful mini workstation,” and it has packed AMD’s latest Ryzen AI Max Plus Pro chips inside with support for up to 16 CPU cores, up to 8TB of SSD storage thanks to dual NVMe modules, up to 128GB of unified memory, and the ability to assign up to 96GB of RAM to the GPU for graphical tasks or AI projects. You’ll also be able to pick between AMD Radeon 8040S, 8050S, and 8060S integrated graphics. It can support up to four 4K monitors.
This mini PC is also a Copilot Plus PC, so it can deliver the AI performance required for features like Recall, Click to Do, and AI-powered image generation and editing in Windows 11.
Image: HP
HP’s Z2 Mini G1a can also be positioned vertically on a desk.
Like many other mini PCs being announced at CES this week, it has a built-in power supply, so it’s ideal for sitting on a desk. It’s also small enough to slot inside a rack, with up to five of these mini PCs able to fill up a 4U rack. HP will even offer a choice of Windows 11 Home or Pro or Ubuntu 24.04 LTS.
HP’s Z2 Mini G1a also has a great port selection, with a single USB-C 10Gbps port at the side, alongside a USB-A 10Gbps port, and a headphone jack. At the rear, there are two USB-A 10Gbps ports, two USB-C Thunderbolt 4 40Gbps ports with USB Power Delivery and DisplayPort 2.1 support, two USB-A 480Mbps ports, two mini DisplayPort 2.1, and an ethernet port. You can also add optional ports at the top of the rear section, with options for additional USB-A and USB-C ports and a variety of ethernet ports.
While HP is announcing its latest all-in-one and mini PCs today, it has not yet provided release dates or pricing.
Qualcomm is introducing another Arm laptop chip to its Snapdragon X series today, lowering the cost of Copilot Plus PCs to around $600. The new Snapdragon X joins the existing Snapdragon X Plus and X Elite and will be available in a variety of devices from Acer, Asus, Dell, HP, and Lenovo in the coming months.
The Snapdragon X is designed for mainstream and budget laptops, so it won’t offer as good of performance as the X Plus or X Elite variants. Qualcomm is still promising better performance per watt compared to Intel’s Core 5 120U processor and better battery life across a variety of tests.
Much like the rest of the Snapdragon X series, the base Snapdragon X chip will include a 45 TOPS NPU that supports Microsoft’s Copilot Plus features. The Qualcomm Oryon CPU has eight cores that boost at up to 3GHz and is built on the same 4nm process node as the rest of the Snapdragon X series.
Image: Qualcomm
The Snapdragon X platform.
It looks like the Snapdragon X will also lay the groundwork for some of the first Qualcomm-powered mini desktop PCs. Qualcomm is promising the “world’s first mini desktop PC powered by Snapdragon X series” tomorrow, so it certainly sounds like more Copilot Plus mini PCs are on the way.
The Snapdragon X will really have an impact on the $600 laptop market. It’s poised to put even more pressure on Intel’s efforts here, and Qualcomm says more than 60 laptop designs from OEMs are currently in production or development based on the Snapdragon X series of chips. More than 100 are on the way by 2026. Windows on Arm momentum isn’t slowing down.
Microsoft is pulling yet another trick to get people to use its Bing search engine. If you use Bing right now without signing into a Microsoft account and search for Google, you’ll get a page that looks an awful lot like... Google.
It’s a clear attempt from Microsoft to make Bing look like Google for this specific search query, and other searches just list the usual Bing search results without this special interface. The Google result includes a search bar, an image that looks a lot like a Google Doodle, and even some small text under the search bar just like Google does. Microsoft even automatically scrolls down the page slightly to mask its own Bing search bar that appears at the top of search results.
Image: Tom Warren / The Verge
The Bing search result for Google right now has a special interface.
While Bing still surfaces search results for Google underneath this spoofed Google UI, a lot of people will see this interface when they configure a new PC and search for Google in the address bar of Microsoft Edge. As 9to5Google points out, it’s a sneaky move from Microsoft to try and keep people using Bing instead of switching to Google.
Microsoft has a habit of this kind of behavior, too. We’ve been cataloging every trick Microsoft has used to convince people to switch to Bing or Edge instead of Google and Chrome over the past few years. Microsoft has modified Chrome download sites, added pop-up ads into Google Chrome on Windows, injected polls into Chrome download pages, and even used malware-like popups to get people to ditch Google.
Google also has its own notifications on its websites to encourage people to download Chrome instead of Microsoft Edge, but they’re nowhere near as aggressive as Microsoft’s use of operating system-level popups and website modifications.
It’s 2025, and Microsoft is kicking off the year by reminding everyone that support for Windows 10 ends in October. While the company has been trying to entice Windows 10 users to upgrade to Windows 11 with full-screen prompts throughout 2024, it’s now calling 2025 “the year of the Windows 11 PC refresh.”
Last year, Microsoft kicked off 2024 by declaring it was “the year of the AI PC,” before launching a range of Copilot Plus PCs several months later. As Microsoft edges closer to the end of Windows 10, it’s making its presence at CES felt this week by declaring that refreshing a Windows 10 PC will be more important than buying a new TV or phone in 2025.
“As CES 2025 begins, showcasing the latest innovations in technology, we are excited for the advancements our industry will offer to people around the globe,” says Yusuf Mehdi, executive vice president and consumer chief marketing officer at Microsoft. “And we believe that one of the most important pieces of technology people will look to refresh in 2025 isn’t the refrigerator, the television or their mobile phone. It will be their Windows 10 PC, and they will move forward with Windows 11.”
Mehdi believes that “Windows 11 is available at a time when the world needs it most” and that “the forefront of AI innovation will be realized on Windows.” 2025 should be a bigger year for Windows AI features, particularly after Recall was delayed enough times that it didn’t launch fully in 2024. Microsoft also hasn’t delivered its improved AI-powered Windows Search features to Insiders yet after unveiling them in October.
Microsoft isn’t at CES this week in the traditional sense of having a booth on the show floor or even announcing product news, but its influence will be felt in the myriad laptops that get announced this week, and even unusual announcements of its Copilot AI assistant coming to LG and Samsung TVs. I wouldn’t be surprised if we see some Microsoft executives jump onstage during CES press events this week. Jason Ronald, Microsoft’s vice president of “next generation,” is already confirmed to appear at Lenovo’s handheld gaming PC event, and Windows chief Pavan Davuluri has been appearing at partner keynotes in recent months.
With Windows 11 adoption still lagging behind Windows 10, it’s no surprise to see Microsoft dedicating the year to making sure people upgrade to Windows 11 or buy a new PC. Windows 11 is now the most popular OS for PC gaming on Steam, but with Microsoft offering Extended Security Updates to consumers for the first time ever later this year, it will be interesting to see how many opt to pay $30 for an extra year of updates instead of moving to Windows 11.
Intel is announcing the rest of the Arrow Lake family of CPUs at CES today, with options for thin-and-light laptops all the way up to gaming notebooks. These mobile processors will appear in many of the computers being announced at CES this week, with the 200HX series being paired with next-gen GPUs we’re expecting Nvidia to announce later today and the 200H and 200U series of chips destined for thin-and-light and premium laptops.
While Intel will continue to supply Lunar Lake CPUs for its range of Copilot Plus laptops, the Arrow Lake mobile family won’t be ditching memory sticks after Intel confirmed in October that the Lunar Lake chips were a one-off experiment.
Image: Intel
Intel’s Core Ultra 200HX series of processors.
The new Core Ultra 200HX series will be targeted at gaming laptops and should deliver around 5 percent better single-thread performance and 20 percent multithread performance improvements over previous Raptor Lake-H Refresh processors.
Intel says new gaming laptops with the “latest discrete GPUs” (read that as Nvidia’s RTX 50-series) will be coming in late Q1. The flagship Intel Core Ultra 9 285HX has 24 cores (8 performance and 16 efficiency cores), can boost up to 5.5GHz, has 4 GPU cores, and a 13 TOPS NPU. That’s enough to meet Intel’s definition of “AI PC” but not enough for Microsoft’s Copilot Plus features.
Consumer laptops that aren’t gaming-focused will come with Intel’s 200H or 200U series of processors. The H variants have a base power of 28 watts, apart from the flagship Core Ultra 9 285H that pushes the power requirements up to 45 watts. These H-series CPUs have a new Intel Arc GPU inside that delivers around 15 percent better graphics performance over previous Meteor Lake chips.
Image: Intel
The Core Ultra 200HX series lineup.
CPU performance on these H chips should be around 15 percent better, too, for single-thread tasks. There are five chips available for laptop makers here, including the flagship Intel Core Ultra 9 285H with 16 cores (6 performance, 8 efficiency, and 2 low-power efficiency cores), a boost clock of 5.4GHz, and 8 GPU cores.
Intel will also release the U series of its Core Ultra 200 processors for laptops that are much more focused on battery life and thin-and-light designs. These chips have a base power of just 15 watts and only turbo up to 57 watts. The top Intel Core Ultra 7 265U processor includes 12 cores (2 performance, 8 efficiency, and 2 low-power efficiency cores) and can boost up to 5.3GHz.
While the 200HX series of chips won’t appear until late Q1 in gaming laptops, the 200H and 200U chips should start shipping in thin-and-light laptops in the coming weeks.
LG and Samsung have both announced their 2025 smart TVs at CES this weekend, and some of them will include access to Microsoft’s Copilot AI assistant. Both TV manufacturers are chasing the artificial intelligence hype train with dedicated AI sections on their smart TVs that include a shortcut to a Copilot web app.
LG is adding an entire AI section to its TVs and rebranding its remote to “AI Remote,” in an effort to sell consumers on the promise of large language models. While it’s not clear exactly how Copilot works on LG’s latest TVs, the company describes access to Copilot as a way to allow users to “efficiently find and organize complex information using contextual cues.”
LG hasn’t demonstrated its Copilot integration just yet, but it has shown off its own AI Chatbot that’s part of its TVs. It appears Copilot will be surfaced when LG TV users want to search for more information on a particular subject.
Image: Samsung
Samsung is showing off its AI Vision features at CES this week.
Samsung also has its own Vision AI brand for its AI-powered TV features this year, which include AI upscaling, Auto HDR Remastering, and Adaptive Sound Pro. There’s also a new AI button on the remote to access AI features like recognizing food on a screen or AI home security features that analyze video feeds from smart cameras.
Microsoft’s Copilot will be part of this Vision AI section. “In collaboration with Microsoft, Samsung announced the new Smart TVs and Smart Monitors featuring Microsoft Copilot,” says Samsung in a press release. “This partnership will enable users to explore a wide range of Copilot services, including personalized content recommendations.”
I asked Samsung for more information or images of Copilot in action, but the company doesn’t have anything more to share right now. I’ve also asked LG and Microsoft for more information about Copilot on TVs and neither company has responded in time for publication. Without any indication of exactly how Copilot works on these TVs, I’m going to chalk this one up as a gimmicky feature that LG, Samsung, and Microsoft clearly aren’t ready to demo yet.
Nvidia’s RTX 5090 has leaked today in the form of a marketing image of the unannounced next-gen GPU. VideoCardz has obtained a box shot of the RTX 5090, which suggests that the rumors of 32GB of GDDR7 memory are true.
While the packaging of the unannounced Inno3D RTX 5090 iChill X3 doesn’t reveal more specs about Nvidia’s flagship next-gen GPU, it does suggest that this particular model will ship with a 3.5-slot cooler.
The RTX 5090 is expected to have double the VRAM of the RTX 5080, which is rumored to include 16GB of GDDR7 memory. It’s also rumored to include 21,760 CUDA cores, nearly 1.8TB/s of memory bandwidth, and a TDP of 575 watts — 125 watts more than the RTX 4090.
We shouldn’t have long to wait until details about the RTX 50-series of GPUs are official. Nvidia is hosting a CES keynote tomorrow night, where the GPU maker is widely expected to announce its RTX 5090, RTX 5080, RTX 5070 Ti, RTX 5070, and even an RTX 5090D model for China. Rumors suggest the RTX 5080 could debut first on January 21st, followed by Nvidia’s other RTX 50-series cards.
Microsoft has been working on dynamic animated wallpapers for Windows 11 for a few years, and now a former designer has revealed exactly what they look like. Sergey Kisselev, a former motion designer and 3D artist at Microsoft, has detailed his work on the dynamic animations for Windows 11 that were originally supposed to ship in 2023 but have likely been canceled now.
Kisselev worked on Windows design elements and Microsoft’s Fluent design system for more than eight years before departing to Amazon in 2022. In his post on Behance, Kisselev describes the dynamic wallpapers as “part of the Windows Creative Direction Team’s efforts to celebrate a new centered signature composition for Windows 11, highlighting its centered Start Menu and taskbar.” The dynamic wallpapers were part of an effort that was “explored for Microsoft’s low-cost devices, primarily targeting educational users,” according to Kisselev.
Windows Central reports that these dynamic wallpapers were originally supposed to ship as part of the 23H2 update for Windows 11, but that never happened. Windows watcher Albacore says the dynamic wallpapers feature has been scrapped, and that unfinished parts of it shipped in both Windows 11 version 22H2 and 23H2, but were removed in the latest 24H2 update.
These dynamic wallpapers look very similar to what Microsoft does with its Xbox dashboard, and the only way to currently get animated wallpapers in Windows is to install a third-party app like Wallpaper Engine. Microsoft used to support videos as wallpapers with its Windows DreamScene feature in Windows Vista, so it’s surprising to see Microsoft do all of this design work to bring a similar dynamic wallpaper feature back and then never ship it.
Asus’ upcoming mini PC has a Copilot button. | Image: Asus
Ever since Microsoft first introduced its Arm-based Copilot Plus laptops in June, I’ve been wondering when we might see Copilot Plus features appear on desktop PCs. Six months on, it’s clear we’re about to see mini PCs that deliver the AI performance required for features like Recall, Click To Do, and AI-powered image generation and editing in Windows 11. These mini PCs might even help Microsoft compete with Apple’s latest Mac Mini.
Asus became the first PC manufacturer to announce a mini PC that’s Copilot Plus capable in September. It then revealed the full specs of its upcoming NUC 14 Pro AI last month, ahead of the Consumer Electronics Show (CES) that kicks off next week. Asus’ mini PC even has a Copilot button on the front and is almost identical to the size of Apple’s latest Mac Mini.
The timing of Asus’ spec drop came on the same day that Taiwanese company Geekom revealed three new mini PCs that it will showcase at CES. Geekom is releasing a mini PC with AMD’s Strix Point CPUs inside and one with Qualcomm’s Snapdragon X Elite processor, meaning both will be Copilot Plus compatible. The third model is powered by Intel’s unannounced Arrow Lake-H laptop processors, which are...
27-inch 4K OLED 240Hz monitors seem to be like buses: you wait ages for one and then three turn up at once. Asus, Samsung, and MSI are all announcing the industry’s next-generation QD-OLED gaming monitors that offer the benefits of 4K OLED 240Hz panels at the smaller 27-inch size instead of 32 inches.
All three appear to be using the same fourth generation QD-OLED panel from Samsung Display, which Asus says offers “a longer lifespan over previous-gen OLEDs.” Both the Asus ROG Swift OLED PG27UCDM and the MSI MPG 272URX QD-OLED (who names these things?) include DisplayPort 2.1a (UHBR20), which offers 80Gbps of bandwidth to support 4K at 240Hz without the need for Display Stream Compression (DSC). Samsung’s press release about its Odyssey OLED G8 (G81SF) doesn’t mention DisplayPort 2.1a compatibility, but it’s reasonable to assume it’s part of the spec list.
Image: Asus
Asus’ latest OLED monitor has a 26.5-inch viewable display.
Image: Samsung
Samsung’s Odyssey G8 now comes in a 27-inch OLED 4K 240Hz variant.
MSI and Asus’ models both support DisplayHDR True Black 400, and Asus also supports Dolby Vision HDR. Both MSI and Asus are offering a three-year warranty that includes burn-in protection, but Samsung hasn’t confirmed its warranty situation for its latest G8 model. Samsung also hasn’t fully detailed the specs of its latest G8 OLED model, but it’s reasonable to assume it will support DisplayHDR True Black 400 at the minimum.
Interestingly, Asus’ model only has a 26.5-inch viewable display, but both MSI and Samsung are marketing their monitors as 27-inch ones. With all three offering the 0.03ms response times of OLED, 240Hz refresh rates, and above 160PPI for improved text clarity, the choice will really come down to design, features, and pricing. Unfortunately, Samsung, Asus, and MSI haven’t announced release dates or pricing yet.
Valve revealed earlier this year that it will support the ROG Ally handheld with its Steam Deck operating system, and now it looks like Lenovo will soon have its own SteamOS handheld, too. Evan Blass has posted images of an unannounced Lenovo Legion Go S on X today, revealing black and white variants of a handheld gaming PC. The interesting part? The black one has a Steam button.
While the low resolution images don’t immediately scream SteamOS, if you look closely you’ll notice a Steam logo is visible on a button to the left of the display. Interestingly, the white model doesn’t have the same Steam button — which could indicate Lenovo is preparing Windows- and SteamOS-powered models of its Legion Go S.
Image: Evan Blass (X)
You can clearly see a Steam logo on the button to the left of the display.
PC makers like Lenovo, Asus, and MSI have all opted for Windows-based handhelds so far, but as consumer patience with Windows on tiny screens wears thin and Microsoft’s progress in improving the experience is slow, it seems like OEMs are looking for a better alternative.
Lenovo’s potential entry into a SteamOS handheld comes just days after Valve quietly updated a document with new branding guidelines that include “Powered by SteamOS” hardware. It’s the surest sign yet that Valve is working with more third-party hardware manufacturers, beyond Asus, to certify devices with SteamOS. With CES 2025 just a few weeks away, we could be about to witness a lot more SteamOS-powered hardware.
Valve has been here before, though. At CES in 2013, it tried to tempt PC manufacturers to sign up to its Steam Machines initiative, but it didn’t get far enough with developers to convince them to port games to Linux to make these machines successful. There’s reason to believe this time things will be very different, thanks to Proton. The Steam Deck uses the Proton software compatibility layer and has already proved that many Windows games can even run better on Linux as a result.
Microsoft has quietly ended the sale of new credits and phone number features for Skype this week. Skype Credit and Skype Number are both being discontinued in favor of Microsoft pushing subscriptions instead.
Skype Credit was a pay-as-you-go plan for making calls both locally and internationally, and Skype Number allowed you to have a phone number that people could call and get through to your Skype account. Both have existed in Skype for years, making it a popular choice for calling landlines and mobile phones over the internet.
“New sales of Skype numbers and Skype credit have stopped, but customers can still use existing Skype numbers and credits,” says Amit Fulay, vice president of Microsoft Teams and Skype, in a statement to The Verge. “Users can also purchase new outbound PSTN calling services through monthly Skype subscriptions or use Skype’s free VoIP services for voice or video calls.”
While existing Skype phone numbers will continue to work for now, it seems inevitable that Microsoft will eventually force existing users into a subscription, too. Existing Skype Credit can also be used, but you can’t buy any additional top ups. Even with Microsoft’s Skype subscriptions, there’s still no direct replacement for having a phone number that people can ring to get to your Skype.
Microsoft made the surprise decision to remove ads from Skype earlier this year, alongside new features like AI image creation. The communications app, acquired by Microsoft in 2011, has struggled to remain relevant in recent years up against WhatsApp, Zoom, FaceTime, and other video calling apps. Microsoft has also turned its attention more to Teams in recent years, especially after launching a personal version of Teams in 2020.
