Intel launched its long-awaited Core Ultra Series 3 “Panther Lake” mobile chips at CES 2026 this week, promising an alluring blend of long battery life and shockingly great integrated graphics performance thanks to its new Arc Xe3 graphics cores. (It’s true! We benchmarked Panther Lake gaming performance ourselves.) Core Ultra Series 3 is looking pretty damned good, and its strengths could help Intel finally establish a stronger foothold in Steam Deck-style gaming handhelds – a segment long dominated by AMD’s bespoke Ryzen Z1 and Z2 handheld chips. The combo led Nish Neelalojanan, senior director of client product management for Intel, to come out swinging about Panther Lake’s potential advantages in tomorrow’s PC gaming handhelds. “They’re selling ancient silicon, while we’re selling up-to-date processors specifically designed for this market,” Neelalojanan told PCWorld’s Mark Hachman in an exclusive interview. Much like Intel’s Panther logo, Neelalojanan has confident swag about Core Ultra Series 3’s handheld performance.Mark Hachman / Foundry Bold words indeed… but ones that may have a ring of truth to them, as AMD’s lower tiers of Z2 chips lean on both older CPU tech and older GPU tech to help keep costs down. Well, it just so happened that Mark also had a roundtable interview scheduled for the very next day with Rahul Tikoo, senior vice president and general manager of the client business at AMD, to discuss all of the compay’s CES 2026 chip announcements. And naturally, he asked Tikoo about Intel’s comment as part of the far-ranging interview. Here’s the excerpt, lightly edited for clarity. Be sure to check out the rest of the interview for insights into AMD’s new Ryzen AI 400 laptop lineup, the just-released Ryzen 9 9850X3D and rumors of a dual-core X3D chip, and more. I just came back from Intel, where they planned to invest heavily in the handheld space, which you’ve dominated. They claim that you’re selling “ancient silicon.” What’s your strategy going forward in the handheld space? “We’re very committed to the handheld [space]. I mean, we created the space, so it’s a space that we’re very committed to. Here’s the beauty, though, of AMD and why we have a much higher chance of success in that space: because of our console business, or how we develop semi-custom silicon for the console business. You can’t just use mobile silicon and put it in the handheld. You can, but the handheld or the consoles, they care about high graphics. They don’t care about as much compute, and they don’t care about the I/O. So, if you’re putting a notebook chip like Panther Lake in there, and you’re not purpose building it, you have all this baggage that Panther Lake is going to carry around its chiplet architecture. You know, the interconnects of the chiplet architecture, the I/O that they have in there. I mean, it’s a Swiss Army Knife, and it’s good for certain things. We can do that, too. In fact, we do that in the handheld space in some segments. But when you think about the core of the handheld space, they want purpose-designed, purpose-built chips that have great graphics technology, great software like FSR, integration with game developers on Xbox, PlayStation, etc. We can have high battery life, good fidelity of content, high frame rate, and we do that very well.” Intel believes their low-power E-cores give them an advantage, as they extend battery life. Does AMD have a response to that? “We haven’t seen any issues there. I’ll tell you this, Intel does play games sometimes, and it’s very interesting. We had a customer. They said the same thing. They’re like, hey, I can get more battery life with Lunar Lake against the 300 series. So, we’re like, okay, let’s do a quick experiment. And we did this in the lab. And actually, Qualcomm did a video on this too, because we didn’t want to go out and do a video and everything. Qualcomm did a video on this: Lunar Lake has great battery life when measured with MobileMark with the power connected. As soon as you go in DC Mode, battery life climbs while performance drops. The Core i7 performs like a Core i3. So, the E-cores are very good for efficiency, very bad for performance. We balance the two, and we’re already making those choices for our customers and saying, hey, you don’t have to worry about it.” But does that hold true for Panther Lake? We were able to benchmark Core Ultra Series 3 both plugged and unplugged, and the frame rates were surprisingly close in the limited testing available during CES 2026. So there you have it – it appears a full-fledged war (or at least a war of the words) is brewing for the CPUs beating in the heart of the PC gaming handhelds that have taken the world by storm. Will Panther Lake’s potent Arc graphics manage to unseat AMD’s stranglehold on this new class of devices? Time will tell, but it seems clear that both Intel and AMD aren’t shying away from a fight. Read...

At CES 2026, AMD and Nvidia both signaled that they’re considering bringing back older chip technologies to battle the sudden shortages and extreme price hikes battering the PC ecosystem right know. I warned y’all this could happen. We were chatting about reports of AMD resurrecting older B650 motherboards in a recent episode of our Full Nerd podcast. And, as part of the discussion, I brought up that chip vendors like AMD and Intel are already mixing older silicon into modern lineups in somewhat sneaky ways. (Remember AMD’s laughable decoder ring?) I suggested we could soon witness the return of older nodes to “new” products to help combat soaring PC costs. “Let’s not get crazy,” fellow Nerd Will Smith said after a sigh. Well, Will, I’m not crazy. It’s the world we’re living in that’s crazy. In a series of roundtable interviews with AMD and Nvidia executives, Tom’s Hardware pressed that exact line of questioning. And the answers Tom’s received may simultaneously cast fear into the hearts of long-term PC enthusiasts and spark hope for more affordable PCs in the hearts of normies. A GeForce RTX 3060. Rest in peace, EVGA.Thiago Trevisan/IDG Nvidia CEO Jensen Huang didn’t sound opposed to the idea when Paul Alcorn posited the question of spinning up older GPUs on dated process nodes. “Yeah, possibly, and we could possibly, depending on which generation, we could also bring the latest generation AI technology to the previous generation GPUs, and that will require a fair amount of engineering, but it’s also within the realm of possibility. I’ll go back and take a look at this. It’s a good idea.” It’s not a tacit confirmation that Nvidia would do so, obviously, but it clearly could be on the table. Earlier this month, rumors started swirling about Nvidia potentially reviving the RTX 3060 — a card that launched in 2021 — in this quarter. The Ryzen 7 5800X3D was the fastest gaming processor ever released for AM4 motherboards, landing in 2022. Adam Patrick Murray / IDG AMD’s David McAfee, the VP and GM of Ryzen and Radeon, had obviously thought of the possibility before Tom’s raised it in another roundtable interview. AMD “[is] certainly looking at everything that [it] can do to bring more supply and kind of reintroduce products back into the [AM4] ecosystem to satisfy the demands of gamers that maybe want that significant upgrade in their AM4 platform without having to rebuild their entire system.” McAfee also added that the concept is “definitely something [AMD is] very actively working on.” AMD’s newer, DDR5-based AM5 motherboards launched alongside the Ryzen 7000 series all the way back in 2022. DDR4-based AM4 motherboards have been around since the first generation Ryzen chips launched in 2017. DDR4 is being phased out and isn’t as cheap as it was a year ago, but still costs far less than DDR5 RAM. You can buy 16GB of DDR4 for around $150 at the time of writing, while 16GB of DDR5 will set you back around $250, depending on the kit – and prices are only rising (rapidly) for the newer gen. Is this a good thing or a bad thing? I think it depends on your philosophy. Going backwards is nothing you ever want to see in the technology industry, but with AI datacenters gobbling up all the air (and RAM, and GPUs, and SSDs, and..) in the room currently, PC vendors risk losing mainstream buyers completely if they don’t figure out a way to combat the soaring prices. Either way, it’s definitely a trend to keep an eye on in 2026, where what’s old could apparently be new again. Hopefully you managed to snag a new PC for cheap during Black Friday, because we warned you about that, too. Read...

Nvidia may not have announced any new graphics cards at CES 2026, but it did reveal two stunning technology upgrades: Upgraded G-Sync Pulsar monitors and DLSS 4.5. The problem is it’s damned near impossible to see the advantages each provides in pictures and video — the visual quality differences don’t translate well when captured through a camera lens. The Nvidia promo videos below give good oversight into the new technologies, for example, but you don’t get the full experience secondhand, kind of like VR. Well, friends, I spent two full hours in Nvidia’s booth getting deep-dive, hands-on demonstrations of both G-Sync Pulsar and DLSS 4.5, and let me assure you: They’re amazing. I’m going to try to get my grubby gamer paws on a G-Sync Pulsar monitor after CES to do a deeper analysis, but the upshot? These puppies use technical tricks mixed with complex science to all but solve motion blur. Everything stays crystal clear while you’re panning around without ugly motion blur — the difference is so stark I picked up on it immediately. Gamers who love complex strategy or tactics games (like Civilation or Anno) or esports games like DOTA 2 and LoL should strongly consider picking up one of these 1440P, 360Hz displays over an OLED monitor. Sacrilege, I know, but Pulsar’s IPS display felt that damned good. I also got the chance to poke around with DLSS 4.5, and my demos had a handy toggle that let you switch between DLSS 4 and 4.5 on the fly so you can see the changes instantly. Once again, side by side, the upgrade is clear as day. DLSS 4.5’s new second-gen transformer AI model delivers tangible improvements in some key areas. It damn near solves ghosting, an issue that’s plagued upscaling since infancy, and makes temporal stability — the shimmering effect you might see in small, detailed areas like a ball of yarn or wavering tree leaves — so much better than before. Sharper anti-aliasing also makes imagery more crisp, which was especially noticeable in Outer Worlds 2‘s space scenes and the ornate armor of Black Myth Wukong‘s warrior. Flipping between DLSS 4 and 4.5, the improved visual fidelity was immediately obvious. DLSS 4.5’s new 6x and Dynamic Multi-Frame Gen capabilities make maxing out your monitor’s capabilities seamless, too. Dynamic MFG automatically manages how many AI-generated frames are inserted between traditionally rendered frames, to scale to match your monitor’s refresh rate. I tried it on several systems. In an RTX 5070 desktop attached to a 4K 240Hz display, Dynamic MFG usually ran at 5x or 6x, to be able to match the refresh rate. On a similar system attached to a 4K 165Hz display, it stayed relatively stable at 3x, since you don’t need to create anywhere near as many frames to keep that lower refresh rate fed. Nvidia’s DLSS 4.5 demos had a custom frame rate counter at the top that showed frame rate, latency, as well as the level of Dynamic Multi-Frame Generation. You could also toggle between DLSS 4 and 4.5 with a switch to immediately see and compare differences.Brad Chacos/Foundry Better yet, I felt zero judder or jankiness when Dynamic MFG shifted up or down to the next level. That was a worry of mine when the technology was announced; would you feel it when it suddenly changed from 6x to 4x while transitioning environments? Nope. Bonus: Dynamic MFG is smart enough to shut off AI frames completely when sections of a game don’t need brute force, such as pre-generated cutscenes running at a locked 60fps. I’ll be able to dive much deeper into both G-Sync Pulsar and DLSS 4.5 when I’m back at home with my gaming rig — it’s kind of hard to test gaming fidelity when you’re traveling with a work-issued laptop. But I wanted to get this information out there ASAP since DLSS 4.5’s transformer upgrade and the first G-Sync Pulsar monitors are already being sold online.Don’t be scared. Even if you can’t see the advantages of DLSS 4.5 and Pulsar monitors in a web video, they are there — and they look amazing. (Seriously, Civ and League of Legends fanatics should take a long, deep look at investing in a Pulsar display.) Read...

It’s hard to believe, but Intel’s just-launched Core Ultra Series 3 (Panther Lake) laptop graphics may, in fact, be as good as a laptop from as little as two years ago running a discrete RTX 4000-series discrete GPU. That was the challenge that Intel put before reporters at a special benchmarking session of Panther Lake here at CES 2026. We were invited to take a prototype Lenovo laptop with a Core Ultra X9 388H inside of it, configured with twelve integrated Xe3 GPU cores inside of it, which Intel calls the Intel Arc B390. Reporters were invited to test games and only games, but whatever games and at whatever resolutions we’d like. Unfortunately (maybe?), Panther Lake’s performance was so good that I stopped trying to test multiple games, and instead started running the tests to reflect our own benchmarks recorded on gaming notebooks. Is Panther Lake as good as a RTX 4050 creator-class laptop? The short answer? If not, it’s certainly close. Real proof from real tests; Intel’s Panther Lake.Mark Hachman There’s a wrinkle: Intel supports its latest XeSS technology inside the Arc B390, which supports resolution scaling (render in a low resolution, then upscale it) as well as AI-generated frames — specifically, three AI-generated frames for each rendered frame. Some games let you turn this feature on and off, and I wanted to let Panther Lake render each frame, but also see what would happen when the frame generation technologies were turned on. Core Ultra Series 3: early test results Keep in mind that we were given less than 90 minutes for testing, which limited our available runs. We were also prohibited from testing synthesized benchmarks or those which focused on the CPU — those will be restricted to the formal review process, which should begin sometime soon. We weren’t prohibited from unplugging Lenovo’s laptop, but that wasn’t done for want of time, either. Finally, not all runs were performed thrice, which we do to eliminate any one-time glitches. Lenovo’s prototype Panther Lake notebook.Mark Hachman With that said, Panther Lake’s integrated Arc graphics are the real deal. I compared Intel’s Core Ultra Series 3 chip to the Series 2 chip (Lunar Lake) as well as the Core Ultra Series 1 (Meteor Lake) that Intel has launched over the past few years. I also added Qualcomm’s Snapdragon X1 Elite and AMD’s Ryzen AI 300 chip, Note that I’m eager to benchmark the latest Snapdragon X2 Elite as well as the Ryzen AI 400 laptop processor, but neither company has made those chips available to test. In my Lunar Lake review, I tested the chip against a slew of benchmarks, including two games, Shadow of the Tomb Raider and Cyberpunk:2077. The former is a few years out of date, but with lots of testing to back it up. I tested Cyberpunk at 1080p resolution on Low settings, but with all rendered frames. Mark Hachman / Foundry Compare the 81 fps from the Panther Lake notebook versus the 37 fps from a couple of years ago. That’s incredibly impressive! I then turned to the older Shadow game, where the Core Ultra Series chip shone again when run at 1080p in Low settings. A 112 fps frame rate is more than playable. To my knowledge, these were all rendered frames as well, as the games was released long before XeSS shipped. Mark Hachman / Foundry Those framerates push into gaming laptop territory. So why not use the settings that we use for our gaming laptop tests? The difference isn’t that great: 1080p settings at the Highest setting for Shadow generated a frame rate of about 75 fps, which puts it a bit below the other laptops I quickly assembled, but it still incredibly great for integrated graphics — on par with the RTX 3060 from just a few years ago. Mark Hachman / Foundry Ditto for Cyberpunk, whose Panther Lake scores are just below the others — all of which pack discrete GPUs rather than integrated on-CPU graphics. Basically, these two charts show that Intel still has a little work to do when purely rendered frames are considered, as least compared to laptops with standalone graphics cards. Intel believes (and why wouldn’t they?) that you’ll turn on frame generation when you can, however, simply to smooth out the gameplay. And when you do that, things change dramatically. When image upscaling and frame generation were turned on in Cyberpunk (1080p, Low), the frame rate jumped from 81 fps to a whopping 169 frames per second. Again, that was evidence that I was wasting my time comparing Panther Lake to other integrated graphics; those scores are something that a dedicated gaming laptop will deliver. My tests showed, however, that you can’t get too crazy — not quite yet, anyway. At 1080p Ultra settings, the Lenovo laptop with the Core Ultra Series 3 chip inside produced 47 frames per second in Cyberpunk with frame generation and upscaling turned off. Though, to be fair, that remains incredibly impressive for integrated graphics, and is more than playable for people unfamiliar with hulking gaming machines. Mark Hachman / Foundry With XeSS turned on, the benchmark score leaped to 111 fps — again, fantastic for integrated graphics! But set to the maximum Ray Tracing Overdrive setting, the laptop produced scores of 8 fps with frame-gen off, and 34 fps with it on — barely, barely playable. On-chip graphics aren’t quite ready to challenge discrete GPUs (Nvidia ones specifically) when you’ve got all the bleeding-edge visual bells and whistles turned on. Still, if you’ve wondered if you should take Intel’s claims seriously…yes, you probably should. I’m eager to give Panther Lake a more thorough examination soon. Read...

Lenovo’s Yoga Mini i is Lenovo’s answer to one of the hottest categories around: the mini PC. And yes, it’s round, with a light bar that’s as productive as it is fun. Lenovo claims that this devices takes up a liter of volume. It does not, according to a rather indignant product manager that insisted that the total volume might be closer to 0.85 liters instead. In any event, the Yoga Mini i will ship in June for an estimated starting price of $699. Mini PCs have begun to surge in popularity, partly because they can offer a substantial amount of computing horsepower in very little space. They’ve become the territory of ambitious Taiwanese and smaller Chinese vendors, but Lenovo, traditionally at or near the top of the PC vendor list, is determined to make its mark. How? Though a nifty little light bar that runs underneath this mini PC. Naturally, you can configure the color within a Windows application, and you can tell the Mini i to light the bar or flash it in a variety of scenarios: when it detects your presence, when something happens (like an email), or any number of other configurable situations. The Lenovo Yoga Mini i mini PC can be controlled via the Lenovo AI Turbo Engine app.Mark Hachman / Foundry The Lenovo Yoga Mini I uses Wi-Fi sensing, a technology that Intel debuted in the Core Ultra 200, or “Lunar Lake.” Imagine sitting by a still, foggy pond, whose surface begins to ripple and splash as something moves through it. Wi-Fi sensing can’t tell the Mini i who’s there, but it can wake up the device. The mini PC has an integrated fingerprint sensor to identify and authenticate the user. Lenovo’s demo showcase proclaimed its close alliance with Intel, and no wonder: the Mini i includes the core Ultra X7 358H inside of it, one of the “Panther Lake” chip variants due for a more formal unveiling here at CES 2026. This mini PC can be literally held in the hand, but the Lenovo Yoga Mini i really isn’t portable.Mark Hachman / Foundry The Lenovo Yoga Mini i weighs just 1.32 lb, and Lenovo feels that it’s small enough to be moved from room, or even into a backpack. I don’t agree; disconnecting all those cords and cables will be a pain. It’s 5.12 inches in diameter, and just under 2 inches thick. On the outside, there’s a Thunderbolt 4 port, two 10Gbps USB-C ports (one designed to accept power), an HDMI 2.1 interface, and a 5Gbps USB-A port, too. An Ethernet jack accepts up to 2.5Gbps inputs. Ports, ports, ports adorn the Lenovo Yoga Mini i .Mark Hachman / Foundry Lenovo is still saying that the Mini i can include up to 32GB of LPDDR5x memory and up to a 2TB PCIe SSD, apparently banking on what the company says is a stockpile of memory and storage components to help offset sharp price increases. We also don’t know the minimum specification. But $699 is a pretty diminutive price for this mini PC. Read...

If you’ve been waiting — like we have — for truly useful artificial-intelligence applications to land on your laptop, Lenovo has an answer: Qira, a Lenovo-authored AI app that will live on new, select Lenovo PCs and Qira smartphones in the first quarter of 2026. Lenovo describes Qira as an “ambient” intelligence, which might be both good or bad; Windows’ Clippy was famously an assistant which tried to understand what you were doing and offer assistance. Qira sounds like something similar, though with the intent that it “follows” you from Lenovo device to Lenovo device, or on to a Motorola smartphone as well, using a combination of agents and other tasks. Lenovo says that this will be marketed as Lenovo Qira, launching on “select” devices in the first quarter, and as Motorola Qira on smartphones later on. Lenovo says that Qira was designed for privacy, running locally as well as in conjunction with “secure” cloud services. “Every aspect of the Lenovo Qira experience is designed to be secure, ethical, and accountable,” Lenovo says. I didn’t really have a chance to see Qira in action before CES 2026, where Lenovo launched the technology. But the company describes Qira as performing three key functions: presence, actions, and perception. Yes, Lenovo Qira looks a bit like an LLM.Mark Hachman / Foundry Qira can “proactively surface suggestions,” or it can be invoked by saying “Hey, Qira” or by clicking the app’s icon. Lenovo says that you’ll be able to specify documents or “memories” for Qira to access, but that it also “orchestrates actions across apps and devices, coordinates agents, and moves work forward without forcing users to manage every step themselves,” using agents or even offline. The idea is that will develop a “living model of the user’s world,” understanding “context, continuity, and personal patterns over time.” Naturally, a Windows PC like Lenovo’s will already have Microsoft’s Copilot running. It will be interesting to see if the two can interact, or if Lenovo will try to push Copilot to the background instead. That’s a lot of buzzwords that could mean just about anything, depending on the context. Native applications are polarizing enough already: some users like an absolutely “clean” Windows installation, while others appreciate apps like Lenovo’s Vantage software, a centralized command and control center for configuring various aspects of Lenovo laptops, such as function keys or whether a laptop’s charging ports work while the laptop is in a sleep state. Lenovo Qira, presumably transcribing voice input.Mark Hachman / Foundry I personally like Vantage, but there’s a major difference between clicking through a series of actions in a centralized app, and then giving access to personal documents to an unknown AI. I can’t help but suspect that Lenovo will have a kill switch in place for certain customers. What, specifically, can Lenovo’s Qira do? Some of Qira’s abilities sound familiar: “Write for Me,” for example, is something most AI’s can do, penning some text in an appropriate style or voice. Catch Me Up is something apps like Slack offer: the ability to summarize an active chat Here, it “highlights what matters, and helps you re-enter your work.” Similarly, “Pay Attention” provides translations and transcriptions when enabled, as well as AI summaries, similar to Otter.ai or other transcription services. Another use case for Lenovo Qira.Mark Hachman / Foundry Others feel a bit more experimental. A Live Interaction feature “enables real-time, multimodal interaction while you are sharing your screen” — whatever that means. “Next Move” sounds like the weirdest, offering “proactive, contextual suggestions based on what you’re doing in the moment, with continuity across devices evolving over time,” Lenovo said. “It surfaces useful next actions to help you move forward without extra steps.” Qira, naturally, is a big bet for Lenovo. Corporate customers are sure to give Qira a doubtful eye…but many of those same customers are being actively encouraged to use AI to save time and resources. A vocal cadre of consumers actively hate it. We still don’t know which devices Qira will debut on. But as one of the largest PC companies in the world, Lenovo is almost obligated to give AI a try. We’ll have to see if it can pull it off. Read...

Smart appliances that can be controlled with voice commands are nothing new, but IAI Smart is showing a new line of Emerson Smart appliances at CES that respond directly to voice commands. They don’t need a smart speaker in the middle, and they don’t rely on a broadband connection, an app, or anything other infrastructure—everything is processed locally. If you’re leery of the privacy and security vulnerabilities of IoT devices, this could be the answer. Emerson Smart devices—tower fans, space heaters, air fryers, and smart plugs, to start—use IAI Smart’s proprietary SmartVoice technology, which embeds natural-language voice processing directly into the appliance. Each device has an integrated microphone, so you can speak a wake word relevant to the appliance you want to use: “Hey Fan,” “Hey Heater,” or “Hey Air Fryer,” for example. Most also include an onboard speaker to provide audible confirmation of your command without relying on an intermediary device or an internet connection. There’s nothing new about smart plugs, except that this Emerson Smart model can be controlled with voice commands without depending on Wi-Fi.IAI Smart Emerson Smart is not marketing its technology as a replacement for Alexa- or Google-powered smart homes, but SmartVoice’s disconnected nature will be a compelling feature to many. Since all processing occurs on the device itself, recordings of your voice—and your usage data—will never leave your home. And if you have slow or limited broadband service—or an onerous data upload cap—they eliminate the need for persistent connectivity to the internet. The Emerson Smart SmartVoice Air Fryers (one is pictured up top) are the most ambitious products in the new lineup. Available in 5.3-quart ($129.99) and 10-quart ($169.99) sizes, the cookers support more than 1,000 voice commands and have more than 100 cooking presets. Users can issue commands such as “Cook salmon,” “Reheat pizza,” or “Increase temperature,” allowing basic meal prep without ever touching the controls. The company is also showing three SmartVoice tower fans: 29-inch ($89.99), 40-inch $99.99), and 42-inch models ($119.99). The fans have 15-hour sleep timers, wide-angle oscillation, and LED touch controls. The 42-inch model also features an integrated aroma diffuser. SmartVoice Fan-Heaters will be available in two sizes: 25-inch ($129.99) and 32-inch ($169.99). Both provide up to 1,500 watts of heating power, with oscillation options and multiple heat modes. Safety features include tip-over protection and automatic shutoff timers. There will also be Emerson Smart tower fans and fan-powered space heaters with local voice processing.IAI Smart To control lamps or dumb appliances, there will be SmartVoice Electrical Plugs in two configurations: A single-outlet ($24.99) model and a dual-outlet ($29.99) SKU that includes USB charging ports ($34.99). Using the wake phrase “Hey Emerson,” users will be able to issue more than 30 preset voice commands to turn devices on or off, set timers, schedule routines, or group multiple plugs–all without a Wi-Fi connection. Emerson Smart has started with the basics (aside from the air fryer), possibly to find out if there’s a market for its offline approach. If there’s a sizable contingent of buyers who want all the features with none of the connectivity, can refrigerators, washers and dryers be far behind? As novel as these appliances sound, they aren’t the first household products we’ve seen that have local voice-command processing. Simple Human introduced a pricey garbage can that responds to simple voice commands (“open can,” “close can,” “stay open”) way back in 2020. If you’re attending CES in person, Emerson Smart appliances are on display at the Venetian Expo Center, booth #52808. Read...

Love or hate it, upscaling technology like Nvidia’s DLSS have expanded the definition around gaming performance. And while hardware enthusiasts still want to know what to expect for raster performance, free of any software tricks, we’ll have to wait a while longer for a definitive answer. When PCWorld’s own Adam Patrick Murray asked about the RTX 5090 and the future of AI gaming GPUs at a CES 2026 Q&A session for media and analysts, Nvidia CEO Jensen Huang dove briefly into his current view on GPUs, AI, and gaming—one where upcoming video games will house layers and layers of AI: (Transcript lightly edited for clarity.) PCWorld: Adam, with PCWorld, I’d like to talk about gaming for a second— Huang: Yes, it’s awesome. Me too! PCWorld: So Nvidia continues to push DLSS to be better and faster with what’s been introduced— Huang: Pretty amazing, right? Just quickly: Before GeForce brought CUDA to the world, which brought AI to the world, and then after that, we used AI to bring RTX to gamers and DLSS to gamers. And so, you know, without GeForce, there would be no AI today. Without AI, there would be no DLSS today. That’s great. PCWorld: It’s harmonious, yeah. My question, one of my questions, is—is the RTX 5090 the fastest GPU that gamers will ever see in traditional rasterization? And what does an AI gaming GPU look like in the future? Huang: I think that the answer is hard to predict. Maybe another way of saying it is that the future is neural rendering. It is basically DLSS. That’s the way graphics ought to be. And so, I think you’re going to see more and more advances of DLSS. We’re working on things in the lab that are just utterly shocking and incredible. And so I would expect that the ability for us to generate imagery of almost any style from photo realism, extreme photo realism, basically a photograph interacting with you at 500 frames a second, all the way to cartoon shading, if you like. All that entire range is going to be quite sensible to expect. You should also expect that future video games are essentially AI characters within them, and so it’s almost as if every character will have their own AI, and every character will be animated robotically using AI. The realism of these games is going to really, really climb in the next several years, and it’s going to be quite extraordinary, you know. And so I think this is a great time to be in video games, frankly. With the surprise announcement of DLSS 4.5, which will offer RTX gamers further resolution boosts and frame rate bumps, Nvidia does appear to be focused less on where gaming performance improvements come from. Its launch of G-Sync Pulsar display tech (available in select monitors starting this Wednesday, January 7) also reflects this stance, improving motion clarity with a combination of high-end panel specs and clever software control. Read...

A day after AMD announced the Ryzen AI 400 (Gorgon Point) processor for laptops, PCWorld and a handful of other reporters sat down with Rahul Tikoo, senior vice president and general manager of the client business at AMD, to ask about AMD’s client processors: its mobile Ryzen processors, the Ryzen AI Max, desktop processors, and more. Below are excerpts of the interview, edited for space and clarity. Client played a very minimal role in Lisa Su’s keynote last night. What does that mean? Tikoo: It was supposed to be about a 75-minute keynote, and client was about 15 of the 75 minutes, right? So if that gives you a clue, it’s roughly 25% to 30% of the time, and client business is roughly 30% of our revenue right now, right? I mean, so it’s an important part of our revenue profile, and it’s very, very important to us. This is just my characterization, but it appears that the Ryzen AI 400 is a modest upgrade to the Ryzen AI 300, which was a very good chip. How do you see it? I mean Qualcomm, kudos to them for continuing to fight the good fight. But you know, Arm is a big challenge in this marketplace, just because of the application compatibility, I feel really good about our [Gorgon Point] portfolio. Of course, we haven’t had a chance to get our hands on the competitive products yet, but everything that we heard yesterday did not surprise us, because, you know, we have our own market intelligence and what’s happening and what the competitive landscape look like, and so we didn’t see any surprises there. Based on that, what I would say is we have a pretty good head [of steam]. You had a certain number of design wins heading into the Ryzen AI 300, and a number of wins with the Ryzen AI 400. If you can’t give us specifics, which are larger? It’s about the same. What we’re going to see is about between the Ryzen AI 300 and 400 product, and the Ryzen Halo product, we have roughly a little over 250 designs that will be in the market. That’s all three chips. All three chips, yeah, roughly a little over 250 designs, give or take, that will be in the marketplace by the middle of this year, right? Because we just have notebooks that are coming out this month. Desktops will come out in early Q2. [The additional] Strix Halo is also coming out this month. Pro is March. So let’s just call it, the first three-to-four months of the year are going to be busy for us launching the portfolio. You mentioned the AI 400 desktop. It’s going to be a socketed AM5 part? Yeah. It’s a socketed AM5 part. I think the interesting thing about the desktop Gorgon part is that it’s going to be the first Copilot+ part, so the first part with a 60 TOPS NPU. We’ve been working with Microsoft and our partners on optimizing for desktop, because you can imagine desktop has a different set of challenges, right? I think we have a lot of opportunity in that space, and we weren’t there two years ago. We weren’t playing as heavily. We didn’t have enough of a portfolio last year, we had a really reasonable portfolio. This year, we’re going to have even better portfolio. What we’re seeing is a lot of interest in mobile on desktop, even small desktops and even in large desktops, they’re actually putting mobile on because the socket infrastructure is cheaper on mobile. Even traditional desktops? Okay, yeah, even traditional desktops, we’re seeing mobile on desktop now. It’s more relevant in the smaller form factors, like, you know, you have the one liter boxes, the eight liter boxes, the small form factor. So that’s where it’s more relevant, right? But we’ve seen all kinds of desktops use mobile parts. There was a time a few years back where mobile shifted into two categories, high performance and thin-and-light, right? And it’s sort of the same inflection that you see in desktops. Let’s talk about what the prices of RAM and storage are doing, and the effects they’ll have. What are your customers telling you about how they’re going to configure their systems? Are they going to continue on pushing upwards to 2TB SSDs or 16GB of RAM? It depends on the market segment. If you think about creators, they want all the capabilities they can get. Let’s talk about a car company. They’re designing a car. They’re running wind tunnel simulations on a car. Are they going to sweat a 20% or 30% increase in price and say, well, you know, my seven-year research on the car is going to have to be slower? No, they’re going to invest. Now, consumers, on the other hand, you and I, you know, when we sit at home and we’re using the laptop for basic internet, web browsing, or email, we’re going to have to make a choice, right? Do we really need the highest end components in the laptop, or not? Now, we do know there’s a floor. A floor has been set where people like 1TB SSDs are the norm. Nobody buys anything smaller, you know? I mean, even phones, nobody tends to buy anything smaller than a certain capacity, right? So, I think consumers will have to make a choice based on that. But I do expect gamers will continue to invest. Creators will continue to invest. There’s a rumor that AMD was going to launch a Ryzen X3DX2, which didn’t materialize. What’s going on there? X3D dual-cache, right? Stay tuned. Stay tuned. I just came back from Intel, where they planned to invest heavily into the handheld space, which you’ve dominated. They claim that you’re selling “ancient silicon.” What’s your strategy going forward in the handheld space? We’re very committed to the handheld [space]. I mean, we created the space, so it’s a space that we’re very committed to. Here’s the beauty, though, of AMD and why we have a much higher chance of success in that space: because of our console business, or how we develop semi-custom silicon for the console business. You can’t just use mobile silicon and put it in the handheld. You can, but the handheld or the consoles, they care about high graphics. They don’t care about as much compute, and they don’t care about the I/O. So, if you’re putting a notebook chip like Panther Lake in there, and you’re not purpose building it, you have all this baggage that Panther Lake is going to carry around their chiplet architecture. You know, the interconnects of the chiplet architecture, the I/O that they have in there. I mean, it’s a Swiss army knife, and it’s good for certain things. We can do that, too. In fact, we do that in the handheld space in some segments. But when you think about the core of the handheld space, they want purpose-designed, purpose-built chips that have great graphics technology, great software like FSR, integration with game developers on Xbox, PlayStation, etc. We can have high battery life, good fidelity of content, high frame rate, and we do that very well. Intel believes their low-power E-cores give them an advantage, as they extend battery life. Does AMD have a response to that? We haven’t seen any issues there. I’ll tell you this, Intel does play games sometimes, and it’s very interesting. We had a customer. They said the same thing. They’re like, hey, I can get more battery life with Lunar Lake against the 300 series. So, we’re like, okay, let’s do a quick experiment. And we did this in the lab. And actually, Qualcomm did a video on this too, because we didn’t want to go out and do a video and everything. Qualcomm did a video on this: Lunar Lake has great battery life when measured with MobileMark with the power connected. As soon as you go in DC Mode, battery life climbs and performance drops. The Core i7 performs like a Core i3. So, the E-cores are very good for efficiency, very bad for performance. We balance the two, and we’re already making those choices for our customers and saying, hey, you don’t have to worry about it. Can you talk about the desktop X3D processor and the direction that it’s going? It’s a very critical part of our portfolio. I mean, the channel market overall. If you look at IDC, the DIY market is about 30, 35 million units. And give or take, we’re close to 60 points of share in that market, right? We’re pretty high. And then as you look at X3D, which is the top of that market, we have over 80 points of share in that market, and it’s driven by the fact that there’s really nothing else that comes even close in terms of performance. And then with the new X3D part that we just announced, the new part to the stack that, with that boost clock you see on it, it now separates us even more, right? We used to be about 20% better now, or 27% better, when you look at average game performance, and so we’re very committed to that space. That customer base is very demanding, as you can imagine, right? And they’re very vocal. Do you have anything to say about AMD’s ability to supply chips to its customers? We’re using the biggest and the best supplier in the world, TSMC. And our Gorgon portfolio is based on four nanometer technology and is a fully ramped, highly yielding, very proven technology. So, we don’t have the same challenges our competition has where they’re bringing up a new technology. We feel very good about it. No challenges. Threadripper, X3D, and the Ryzen AI Max: these are all innovative though niche products. Does AMD remain committed to all three? We are very committed to those spaces. We’re very, very committed to those spaces. How do you see the Ryzen AI Max going forward? First of all, we will continue to invest in that space. That’s an important space for us. Stay tuned. There will be more announcements in that space over the course of this year. Our focus has been in ramping developers and gamers around that product. You know, thin-and-light gaming is a space where that product has done well. Creative users is another space that product has done well, and now AI. Read...

Smart home device manufacturer Aqara is showing its new Smart Lock U400 at CES. It’s among the first devices of its kind to combine Apple Home Key technology with an Ultra-wideband (UWB) radio to deliver hands-free unlocking when an authorized user approaches the door. Provided you’re wearing your Apple Watch or your iPhone is in your pocket, purse, or backpack, you won’t need to touch the deadbolt to unlock it (compatibility with other smart devices is promised down the road). The trick is made possible by an UWB radio inside the lock, along with support for Matter-over-Thread, although it will also require you to have a Thread-enabled Matter controller inside your home. For Apple Home Key users, that means an Apple TV or HomePod. Using its UWB radio, the Aqara U400 measures time-of-flight (the precise time it takes for a radio signal to travel between the lock and your smart device) and angle-of-arrival (the direction of that radio signal) to determine your location with accuracy down to the centimeter. This ensures that the lock will only open when an authorized user is outside the home approaching the door and not simply walking past it or while they’re inside the house. Family members who don’t use Apple Wallet can unlock the Smart Lock U400 with Aqara’s app for iOS or Android, with a PIN code, fingerprint, or physical key.Aqara Aqara says this approach offers better security since UWB technology is more resistant to relay attacks than Bluetooth-only proximity systems. A built-in gyroscope enables automatic locking once the door is closed, helping prevent homeowners from accidentally leaving their door unlocked. Power comes from a 7.3V rechargeable lithium-ion battery rated for up to six months of use per charge under typical conditions. If the battery runs out, the lock can be temporarily powered via USB-C from a phone or power bank for emergency unlocking. If you’re not an Apple user, Aqara says it will add support for devices that use Samsung Wallet later in the first quarter of 2026. The Aqara Smart Lock U400 also supports tap-to-unlock via Near-field Communication (NFC, via a smartphone or an Aqara key card), as well as with the lock’s fingerprint scanner; personal codes entered on its PIN pad (including one-time and scheduled access); mobile apps; voice control through Alexa, Google Home, Samsung SmartThings, Siri; or a conventional mechanical key. The Aqara Smart Lock U400 It uses a standard 60/70mm deadbolt and supports door thicknesses from 35mm to 55mm, making it compatible with most U.S. residential doors. The exterior keypad panel carries an IP65 rating for dust and water resistance, meaning that the lock is dustproof and can withstand being sprayed with water jets (short of a pressure washer). The Aqara Smart Lock U400 is available now for $269.99 at Amazon or Aqara’s e-commerce site in silver or black (satin nickel and “shadow” black finishes are also promised). This news story is part of TechHive’s in-depth coverage of the best smart locks. Read...