Every January, I travel to Las Vegas to attend the Consumer Electronics Show (CES)—the largest technology conference in North America. This is where I (along with 150,000 other attendees) try to spot the latest tech trends that’ll shape the industry for the next 12 months. For the third year in a row, the same two letters buzzed around every brand at the show: AI. And while AI was everywhere at the show, from chatbots to robots to home appliances, I found the new AI features in laptops especially intriguing. If the hype at CES is to be believed, AI may be about to change how your laptop works. As I made my way from booth to booth, two trends stood out again and again: bold promises around battery life that’s managed behind the scenes by AI, and a wave of AI features that work best inside tightly controlled ecosystems.
I wanted to learn more about how AI in laptops may change my day-to-day work, not to mention whether all of this actually proves useful—or if it’s all just gimmick and distraction. To find out, I spoke with CEOs, product managers and directors on the ground in Las Vegas. Here’s what they had to say about how your laptop may be changing for the better—and possibly for the worse.
Key AI Impact Takeaways:
- AI hardware promises longer battery life—but experts don’t yet agree on how that will hold up in real-world use.
- Brands are building tighter AI ecosystems, improving cross-device convenience while also limiting user choice.
Hardware Built For Performance And Battery Life
Modern processors include neural processing units, known as NPUs, and laptop makers claim these NPUs have the potential to enable multi‑day battery life—but the truth is more complicated, since some experts argue this ignores how people actually use their laptops. NPUs run on‑device AI tasks and “are more energy efficient and more specialized than GPUs. They save energy over GPUs,” according to Dr. Josep Torrellas, director of the ACE Center for Evolvable Computing and professor of computer science at the University of Illinois’ Grainger College of Engineering.
Earlier this month, processor maker Qualcomm debuted its new Snapdragon chipsets. The AI-enhanced NPU offerings built into the processors are useful for your productivity, contends Kedar Kondap, senior vice president and general manager, compute and gaming at Qualcomm. They help optimize performance and efficiency, and Kondap says that consumers can expect to “see some incredible performance and multiday battery life.” He adds, “There are some devices that are touting more than 30 hours of battery life. You can go days without having to charge your laptop.”
It Can Be Useful—But It’s Complicated
At least, that’s what should happen under ideal circumstances. But in real life, this correlation between an enhanced NPU and longer runtime isn’t straightforward. “Chipmakers like Qualcomm are directionally right in that multiday laptop battery life is realistic in specific usage patterns,” says Jon Nordmark, cofounder and CEO of AI scaling solutions company Iterate.ai. “But it’s not a blanket guarantee for ‘normal mixed work’ the way most people actually use a laptop.”
Several experts told me the promise of multiday battery life depends heavily on the task. Torrellas argues that AI tasks “use intensive resources,” meaning the more AI you run, the faster your battery drains. In Qualcomm’s case, the brand has specifically built the NPU to handle certain AI tasks more efficiently—tasks like on-device image processing or noise reduction during video calls. And, when you’re completing these tasks, NPUs are “more energy-efficient than GPUs, and hence can help extend battery life over using just GPUs,” says Torrellas.
So, what does that mean? When you’re using processors with AI-based NPUs, the laptop performs certain kinds of AI tasks locally, in small bursts. These commands don’t need to rely on the cloud or using a graphics card to run AI. “Multiday claims fall apart once workloads stop being bursty and become sustained,” he adds. That happens when you put a substantial load on the processor with creative apps, intensive gaming or heavy multitasking. Jonathan Schaeffer, cofounder of Synsira Software Solutions, Inc., agrees that if you’re not using heavy AI applications or burst commands in your daily work, you may not notice a meaningful difference in your laptop’s battery life.
And that’s not all. Long battery life may only be achievable under certain circumstances. It’s not just about your work tasks. It also centers around whether the apps you use—and the operating system itself—is optimized for these AI features. The good news is that you don’t need to worry about any of this—Nordmark says that the onus is on the applications to recognize these capabilities and perform intelligently. According to Nordmark, you’ll see the best results when the AI recognizes what the laptop wants to do and the software distributes work across the CPU, GPU and NPU in a deliberate way. “That entails CPU for control flow and light preprocessing, NPU for steady low-power inference and the GPU only when you truly need high-throughput parallel computing,” he says. In other words, Nordmark says: “It prevents the GPU from becoming the default hammer for every AI nail.”
So will NPUs actually improve battery life? Sometimes—but only when using specific AI tasks, and only when software is optimized. For most people, the gains will be inconsistent, and perhaps not even especially noticeable, depending on how you use your laptop. AI saves energy only when it replaces an existing process—not when it adds additional ones. NPUs alone won’t extend a laptop’s battery life, but it can help under the right circumstances.
Walled Gardens—Harmful Or Helpful?
In the tech world, “walled gardens” are everywhere. If you embrace Apple products, for example, you’ll find that your devices only work best when you buy all the products with the same badge–your MacBook, iPhone, Apple Watch—even AirPods and AirTags—are designed to work together seamlessly, but not work well with devices from Samsung or Sony. Increasingly, this is also true when it comes to AI. A laptop brand like Apple doesn’t lock you entirely into Apple Intelligence, its AI ecosystem. But Apple doesn’t go out of its way to make it easy to use ChatGPT, Perplexity or any other AI model.
Microsoft users like myself fare no better, as they are pushed to adopt Microsoft Copilot, including an on-device key on the latest laptop keyboards. AI is raising the walls on these gardens—and whether that’s good or bad depends on what you value. Once locked in, chances are you won’t adopt another model, Shaeffer told me. He also contends that this software strategy restricts user choices, which in turn can slow innovation because developers are limited to what they can connect to and customize. “At the same time,” he says, “tight integration can deliver better performance and better security because the platform can control the full pipeline end-to-end.”
That tight integration means these AI models mostly work well inside their own ecosystems. Try to use it across devices, and it falls apart. According to Ryan McCurdy, senior vice president and president of Lenovo North America, “If you think of AI today, a lot of them are apps that you have to open, and you have to live inside of that siloed app,” he says. “They’re all providing value, but there are certain apps that are better for different tasks.” In other words, today’s AI assistants work independently and aren’t designed to talk to one another.
When I met with McCurdy, Lenovo (in conjunction with its smartphone brand Motorola), had just announced a new AI assistant called Qira which works on both Lenovo and Motorola devices. Similar to Google’s Gemini, it’s built to work across your devices, and its AI feature, called Next Move, is what intrigued me the most. Essentially, if you look up something on your Motorola phone during your commute, Qira recognizes what’s on your screen; when you later open your compatible Lenovo laptop, it automatically surfaces your “Next Move,” offering the documents and tools you need to pick up right where you left off—no manual file transfers required. Lenovo claims Qira breaks down these walls—but does it? Its promise is compelling—true cross‑device working continuity without cloud or manual file uploads.
In speaking with coworkers and industry colleagues, the consensus was clear: They use Google Docs because of its portability across devices, which is essentially what Qira aims to do. But answers splintered when it came to which AI agent they used daily. ChatGPT and Google Gemini are popular ways to run queries, for example, and allow you to upload documents as well. AI tools like these aren’t always optimized for letting you seamlessly work across devices—like moving from a laptop to your phone.
Lenovo’s new AI software promises to solve that problem, reducing friction when moving between devices, no matter what kind of documents with which you’re working. “Cross-device AI is improving because continuity is valuable and you want an experience that follows you,” says Schaeffer. “But the real competitive advantage for platforms is deep integration: by moving your data to their cloud-based site, you get access to hardware acceleration and a rich set of AI tools.”
But despite the reduced friction working across devices, there is a lurking problem. In creating yet more AI software for its devices, Lenovo has erected yet another walled garden. It may solve one problem—continuity—while reinforcing the need to commit entirely to a single brand’s hardware. And that is a problem, because these walled gardens don’t talk to each other, and it makes it harder to mix devices to optimize how you work best.
Whether Qira succeeds remains to be seen, and I’m excited to test it in the future because that seamlessness, if it holds up to testing, may be the biggest game changer if you’re working across laptops and smartphones. If you’re leaning toward an AI tool that can work across devices, the Lenovo and Motorola devices might be worth your investment. To access it, you must use Lenovo and Motorola devices—and new 2026 products. The bigger takeaway here: AI will make laptops feel more personalized and connected, but that convenience increasingly depends on which ecosystem you choose. If you can’t wait for them to ship later this year and prefer a Lenovo-branded laptop, our tech team named the Lenovo ThinkPad X1 Carbon as the best business laptop for its blend of raw power and lightweight build.
AI’s Future: ‘It Knows You Personally’
These two AI components in laptops raise an obvious question. If the current AI upgrades provide both genuine improvements and trade-offs, what AI innovations will shape laptops next year?
CES represents the state-of-the-art, and the consensus among every CEO and tech expert I spoke with is that the biggest AI innovations are still in development. But there’s a caveat.
Don’t expect a single winner, McCurdy says; instead, he predicts that AI agents will likely become more collaborative, and multiple solutions will simultaneously help with tasks and projects. “The message from Lenovo is we’re building a collection of the best-in-breed AIs. It’s when you want it and where you need it. It knows you personally,” he explains.
We’ve seen a glimpse of the partnerships that will shape AI. Google already partners with Samsung to bring Google Gemini to its Samsung Galaxy phones. Last month, in a joint announcement, the brands shared that the next Apple Foundation AI models will incorporate Google Gemini technology and help power Apple Intelligence in the future.
But even as the future of AI becomes more flexible, tensions are still present. Smarter, more personalized AI depends on your preferred hardware and ecosystem. Multiday battery life may manifest for some users, but don’t expect it to be universal. There’s also your cross-device AI, which may feel seamless—but only if you have the correct ecosystem to support it.
