Nvidia stock jumps after Meta locks in millions of AI chips — what traders watch next

February 18, 2026
Nvidia stock jumps after Meta locks in millions of AI chips — what traders watch next

New York, February 18, 2026, 12:30 EST — Regular session

  • Nvidia up about 2% after Meta expanded a long-term AI infrastructure partnership
  • Cadence jumps about 9% on results; Analog Devices rises after an upbeat forecast
  • Traders weigh Fed minutes later today and Nvidia’s Feb. 25 earnings as the next tests

Shares of Nvidia rose about 2% on Wednesday, helping pull AI stocks higher in midday trade. The Nasdaq 100 tracker was up about 1.2%, while a semiconductor ETF gained roughly 1.9%.

The move matters because investors have been asking for clearer proof that the AI spending wave is turning into orders and, eventually, profit. A public multiyear commitment from a top customer can calm nerves, even if it doesn’t settle the bigger debate on returns.

There’s also a practical angle. If the big AI checks keep coming, it lifts more than one stock — it supports the suppliers around the build-out, from chip design tools to the parts that help run power and signals inside data centers.

Nvidia said Tuesday the Meta agreement covers millions of its current Blackwell GPUs — chips used to train and run AI models — and its forthcoming Rubin AI chips, as well as Grace and Vera central processors that compete with Intel and AMD. The announcement signaled Nvidia wants those CPUs to take on more “AI agent” work — software that can plan and act with limited human input — and routine data-center tasks such as databases. Meta is building its own AI chips and has talked with Google about using its Tensor Processing Units, or TPUs, and Nvidia executive Ian Buck said Meta has already tested Vera: “Meta has already had a chance to get on Vera and run some of those workloads,” he said. (Reuters)

In a joint announcement, the companies said Meta will build hyperscale data centers optimized for AI training and “inference” — running a trained model — using Nvidia CPUs, networking and millions of Blackwell and Rubin GPUs. “No one deploys AI at Meta’s scale,” Nvidia CEO Jensen Huang said. “We’re excited to expand our partnership with NVIDIA,” Meta CEO Mark Zuckerberg said. (NVIDIA Newsroom)

Meta shares were up about 0.2% midday. AMD slipped about 0.5%, while Broadcom edged higher.

Arista Networks fell nearly 2.5% as traders revisited whether Nvidia’s Spectrum-X networking push could pressure incumbent data-center switch vendors. Wolfe Research analyst George Notter called the Meta-Nvidia news “a non-event for Arista,” while Morgan Stanley analyst Meta Marshall said the announcement carried “the same information” Nvidia disclosed previously. (Investing)

Cadence Design Systems climbed about 9% after it beat quarterly revenue and profit estimates, helped by demand for chip design tools tied to AI processors. CFO John Wall said strong fourth-quarter contract bookings left Cadence with a record $7.8 billion of work under contract. (Reuters)

Analog Devices rose about 2% after forecasting second-quarter revenue of about $3.5 billion, plus or minus $100 million, above Wall Street estimates, as data-center demand helped offset a tough backdrop. “Our revenue outlook… reflects a new high-watermark for ADI,” CFO Richard Puccio said. (Reuters)

“Figuring out the winners and losers in AI is likely to be a center theme in 2026,” said Paul Stanley, chief investment officer at Granite Bay Wealth Management, as U.S. stocks worked back from an AI-led selloff earlier this month. Traders were also waiting for minutes from the Federal Reserve’s latest policy meeting, due at 2 p.m. ET. (Reuters)

But the upside case still has fragile spots. Any sign that the biggest customers are shifting more workloads to in-house chips, trimming capital spending, or pressing suppliers on prices could hit the group fast, especially into earnings.

Next up for the AI trade is Nvidia’s quarterly report on Feb. 25, which investors are treating as a read on demand for Blackwell systems and the pace of the Rubin rollout. (Nvidia)