NEW YORK, Jan 26, 2026, 11:21 EST — Regular session.
- CoreWeave surges following Nvidia’s $2 billion investment; Nvidia shares see little movement
- Microsoft gains momentum following the launch of Maia 200 and Triton, tools designed for Nvidia’s CUDA ecosystem
- Traders are focused on mega-cap earnings reports and the Fed’s Jan. 28 decision, searching for signs of AI spending payoffs
AI stocks showed mixed moves on Monday after Nvidia (NVDA.O) announced a $2 billion investment in CoreWeave (CRWV.O), making it the neocloud company’s second-largest shareholder. Nvidia will pay $87.20 per share. CoreWeave, which rents out GPU-heavy computing power, aims to exceed five gigawatts of AI data-center capacity by 2030. The funds will go toward land, power, and infrastructure—not chip purchases. “Nvidia is the leading and most requested computing platform at every phase of AI,” said CoreWeave CEO Michael Intrator. 1
The sector shifts in a jam-packed week. Apple, Microsoft, Meta Platforms, and Tesla are all set to report, with investors eager for proof that hefty data-center investments are boosting profits, not just driving up costs. The Fed kicks off its two-day policy meeting Tuesday, with bets largely on rates staying put. 2
That’s why these chip-and-cloud headlines are under intense scrutiny. It’s not just about who builds the fastest silicon anymore. Investors are digging into who holds the reins on software, the power infrastructure, and the customer budgets.
CoreWeave climbed 9.2% to $101.57 in late morning trading. Nvidia dipped 0.3% to $187.16. Microsoft (MSFT.O) edged up 1.2% to $471.60, while AMD dropped 2.7% and Alphabet added 1.9%. The iShares Semiconductor ETF (SOXX) slipped roughly 0.4%.
Microsoft has rolled out Maia 200, its second-generation AI chip, stepping up as cloud leaders like Google and Amazon Web Services develop their own alternatives to Nvidia’s gear. The company plans to activate Maia 200 this week in a data center in Iowa, with another site in Arizona lined up soon. It’s also bundling the chip with software like Triton — an open-source platform heavily backed by OpenAI — aiming to rival Nvidia’s CUDA developer environment. Built by Taiwan Semiconductor using a 3-nanometer process, Maia 200 features high-bandwidth memory, though not as cutting-edge as Nvidia’s forthcoming chips, and packs more SRAM, which speeds up AI inference, or handling user queries. 3
Microsoft’s chip hasn’t knocked Nvidia off its perch yet. Instead, it offers major clients a new option and shifts the battle to software—a realm where switching costs matter and developers hold sway.
For Nvidia, the CoreWeave investment feels like a hedge on the entire infrastructure build-out. Sure, chips are key, but the real game involves square footage and megawatts. That’s a slower grind.
Still, the setup of these ties raises concerns over “circular” AI funding, while data-center efforts risk delays if power and building progress lag. Any cut to enterprise AI budgets would surface fast in guidance. Given stretched valuations in parts of the AI sector, even a cautious outlook could hit hard without a major miss.
Wednesday marks the next key moment: the Fed will drop its policy statement at 2 p.m. ET, with a press conference set for 2:30 p.m., right after the Jan. 27–28 meeting. 4
