China’s Lenovo, the world’s largest personal computer maker, is sharpening its push into artificial intelligence by partnering with U.S. chip giant Nvidia to help AI cloud providers deploy data centers at unprecedented speed.

The company announced the tie-up on Tuesday at the Consumer Electronics Show (CES) in Las Vegas, framing it as a direct response to surging global demand for AI infrastructure. Lenovo said the partnership is designed to cut the time it takes to bring AI data centers online from months to just weeks, a critical advantage as competition intensifies among cloud providers racing to support large-scale AI workloads.

Under the programme, Lenovo will combine its liquid-cooled hybrid AI infrastructure with Nvidia’s computing platforms, offering what it described as an end-to-end solution for building so-called AI factories. Liquid cooling has become increasingly important as AI chips consume more power and generate more heat, turning thermal management into a key constraint on scaling.

“Lenovo AI Cloud Gigafactory with NVIDIA sets a new benchmark for scalable AI factory design, enabling the world’s most advanced AI environments to be deployed in record-setting time,” Lenovo chief executive Yang Yuanqing said, speaking alongside Nvidia CEO Jensen Huang.

The language reflects a broader shift in the AI industry. As demand moves from experimentation to industrial deployment, speed of execution is becoming as important as raw computing power. Cloud providers are under pressure from customers building large language models and AI services to add capacity quickly, even as supply chains for chips, power, and cooling remain tight.

For Lenovo, the partnership also signals ambition beyond its traditional PC stronghold. While the company remains best known for laptops and desktops, it also has a sizeable server business and has been positioning itself as a full-stack AI infrastructure provider, spanning devices, data centers, and software.

That strategy was on full display at CES. Alongside the Nvidia announcement, Lenovo showcased an AI platform, a range of concept devices, and its first foldable smartphone under the Motorola brand, highlighting how it plans to weave AI across its entire product portfolio.

Yang also unveiled Qira, a personal AI system designed to operate across Lenovo and Motorola PCs, smartphones, tablets, and wearables. Unlike standalone assistants, Qira is intended to work continuously in the background and integrate third-party services. Lenovo said the system would be able to offer services from companies such as travel firm Expedia, suggesting a move toward an ecosystem model rather than a single-device assistant.

The approach mirrors a broader industry trend, where hardware makers are trying to differentiate themselves by embedding AI deeper into everyday use, rather than treating it as a bolt-on feature. Lenovo is aiming to lock users into its ecosystem while gathering data and usage patterns that can inform future products by controlling both the devices and the AI layer that runs across them.

Lenovo also used the event to showcase concept AI glasses, placing them alongside companies such as Alibaba and Samsung Electronics, which are also exploring AI-powered wearables. In addition, it previewed an AI assistant wearable under “Project Maxwell,” designed to offer users real-time assistance, another signal of how AI is spilling beyond screens into ambient, always-on devices.

The Nvidia partnership sits at the center of this broader push. Nvidia’s chips have become the backbone of AI computing globally, and aligning closely with the company gives Lenovo credibility with cloud providers looking for proven, scalable solutions. The tie-up extends Lenovo’s reach deeper into enterprise and cloud infrastructure by pairing its platforms with Lenovo’s hardware, integration, and global supply chain.

The announcement also comes at a time when geopolitical and supply-chain considerations loom large. A Chinese company working closely with a U.S. AI chip leader highlights how interdependent the global AI ecosystem remains, even as governments talk more openly about technological decoupling.

However, there is the challenge of execution for Lenovo. Competing in AI infrastructure means going head-to-head with established server and systems players, while also keeping pace with rapid advances in chips, cooling, and software. But by promising faster deployment and tighter integration with Nvidia’s platforms, Lenovo is aiming to become a practical partner for cloud providers under pressure to scale.

Lenovo’s message at the CES is that it no longer wants to be seen only as a PC maker adapting to the AI age. Thus, with Nvidia at its side and a growing lineup of AI-driven devices and platforms, it is trying to claim a seat at the table of companies shaping how AI is built, deployed, and used.

Semiconductor stocks have started the year on a strong footing, with gains concentrated not in flashy logic chips but in a quieter, more fundamental corner of the industry: memory.

Shares of the world’s biggest memory makers have surged, reflecting how artificial intelligence is reshaping demand patterns across the chip sector and tightening supply in critical components.

South Korea’s SK Hynix and Samsung Electronics, the two largest memory chipmakers globally, are up 11.5% and 15.9% respectively so far this year. In the United States, Micron has climbed 16.3%. The rally comes as investors bet that AI-related demand, which drove chip markets through 2025, is not fading but intensifying.

At the heart of the move is memory’s central role in AI computing. Training and running large AI models designed by companies such as Nvidia and AMD requires vast amounts of fast, high-capacity memory to move data efficiently between processors. As cloud providers and tech giants pour billions of dollars into AI data centers, memory has emerged as a bottleneck.

One segment has been especially important: dynamic random-access memory, or DRAM, used extensively in AI servers. Prices for DRAM surged sharply in 2025 as demand outpaced supply, and that pressure has not eased. Counterpoint Research expects memory prices to rise another 40% through the second quarter of 2026, extending what analysts increasingly describe as a full-blown cycle rather than a brief spike.

“The recent rally across the semiconductor space has been driven largely by the memory side of the market rather than logic chips,” Ben Barringer, head of technology research at Quilter Cheviot, said in an email to CNBC. “We’re seeing a combination of very strong demand from AI workloads and relatively constrained supply, particularly in high-bandwidth memory, which is essential for training and running large AI models.”

High-bandwidth memory, or HBM, has become one of the most sought-after components in the AI supply chain. It sits close to processors in advanced packaging configurations, enabling faster data transfer and lower power consumption. SK Hynix is widely seen as a leader in this area, supplying HBM used in some of Nvidia’s most powerful AI accelerators, a position that has strengthened its earnings outlook.

That backdrop explains the optimism heading into earnings season. Samsung is expected to report a 140% jump in fourth-quarter operating profit, according to LSEG estimates, marking a sharp turnaround after a prolonged downturn in its memory business. Micron’s earnings per share are forecast to rise more than 400% year-on-year in the December quarter, reflecting both higher prices and improving utilization rates.

The rally has spilled beyond memory producers themselves. Investors are increasingly positioning for a broader AI-driven expansion across the semiconductor value chain. Intel shares are up 7.6% year-to-date, while Taiwan Semiconductor Manufacturing Co., the world’s largest contract chipmaker, has gained 10%. Both companies manufacture a wide range of chips and are expected to benefit as customers ramp up spending on advanced semiconductors tied to AI workloads.

Equipment suppliers are also riding the wave. ASML, the Dutch firm whose lithography machines are essential for producing the most advanced chips, has seen its shares rise 15.2% this year. Bernstein on Sunday raised its price target on ASML from 800 euros to 1,300 euros, implying about 24% upside from Tuesday’s trading level.

“ASML stands to benefit enormously from the wave of capacity expansion planned for 2026 and 2027,” Bernstein analysts wrote, pointing specifically to memory. They said the company would gain “from the upcoming DRAM super cycle,” as manufacturers invest heavily in new fabs and more advanced production lines.

That link is crucial. As memory makers respond to tight supply and strong pricing by expanding capacity, demand for ASML’s tools rises in tandem. Advanced DRAM and HBM production requires cutting-edge manufacturing equipment, locking ASML deeper into the AI investment cycle.

Recent signals from industry executives have reinforced the bullish narrative. SK Hynix has pointed to the possibility of an extended HBM supercycle, suggesting demand could remain elevated well beyond a single year.

“Recent comments from SK Hynix pointing to a potential HBM supercycle have reinforced the idea that this is not just a short-term bounce, but a more structural shift linked to the ongoing build-out of AI infrastructure,” Barringer said. “That has helped improve sentiment across the sector, especially for companies with direct exposure to AI-driven memory demand.”

The emerging picture is one where memory, long treated as the most cyclical and volatile part of the semiconductor industry, has become central to the AI story. As long as companies continue to scale data centers and push larger, more data-hungry models, memory demand is likely to stay tight.

For investors, this has reframed how the semiconductor rally is being judged. This is not simply about who designs the smartest AI chips, but about who controls the components that make those chips usable at scale. So far in 2026, memory makers are winning that argument.