• Nvidia (NVDA) shares rose about 2.5% to a record high, their first since May, driven by a new $150 billion buyback authorization and robust AI infrastructure demand.
  • The buyback adds to $85 billion previously authorized, bringing total repurchase capacity to roughly $235 billion through fiscal 2028.
  • Investors are also responding to long-duration AI computing commitments from cloud providers and frontier-model developers, though the complexity of AI infrastructure financing warrants scrutiny.

Nvidia's Buyback Bonanza

Nvidia Corp. shares climbed roughly 2.5% on Thursday to notch their first record high since May, as investors cheered a massive expansion of the company's share-repurchase program and fresh evidence of insatiable demand for AI computing.

The Santa Clara-based chipmaker authorized an additional $150 billion for buybacks, on top of $85 billion previously approved. The total authorization now stands at approximately $235 billion through fiscal 2028. The move signals management's confidence in Nvidia's cash-generation engine, which has been fueled by the AI data-center boom. Buybacks reduce the share count and can lift earnings per share, a metric Wall Street watches closely.

"The scale of this authorization is unprecedented for a semiconductor company," said one portfolio manager at a large asset manager, who asked not to be named because he is not authorized to speak publicly. "It tells you Nvidia sees a multiyear runway of free cash flow that it can't fully reinvest in the business."

The rally also reflects renewed optimism about AI-infrastructure spending. Market reporting points to expanding long-duration commitments from hyperscale cloud providers and frontier-model developers. Those contracts, while not always publicly detailed, underpin expectations that demand for Nvidia's GPUs and systems will remain elevated even as the company's revenue base grows ever larger.

The AI Trade Roars Back

The record high comes after a months-long stretch in which Nvidia traded below its May peak, even as the broader "AI trade" maintained momentum. The stock's breakout suggests investors are again willing to pay a premium for the company's central role in supplying AI data centers.

Nvidia designs graphics processing units, AI accelerators, networking gear, and the Grace Blackwell platform, along with the CUDA software ecosystem that locks in developers. Its customers include hyperscalers, enterprises, research institutions, sovereign AI projects, and gaming enthusiasts. Recent quarterly results showed revenue of about $96.2 billion, up roughly 106% year over year, with adjusted earnings per share of $2.22 beating the $2.09 consensus. Data-center revenue alone was near $89 billion, up about 117% from a year earlier.

Gross margin has hovered around 74.7%, an unusually high level for hardware, though future margins may face pressure from new-product ramps, memory costs, and competitive pricing. CEO Jensen Huang remains at the helm, and no major restructuring is evident.

Financing Complexity and Competition

The bullish case rests on continued hyperscaler capital expenditure, successful product transitions, and the buyback tailwind. But risks are mounting. Companies are increasingly using leases, special financing vehicles, and vendor-supported arrangements to fund AI data centers. These structures can extend the spending boom, but they also raise questions about end-customer economics, collateral value as GPUs age, and whether today's orders reflect durable usage demand.

Competition is also intensifying. Advanced Micro Devices (AMD), Broadcom (AVGO), custom chips from cloud providers, and China's Huawei are all vying for a slice of the AI silicon pie. Broadcom's work on customer-specific AI chips and financing arrangements illustrates that the contest is expanding beyond conventional GPU sales.

Meanwhile, U.S. export controls on advanced AI chips to China remain a wild card. Nvidia has said it complies with all rules, but the restrictions curtail its addressable market in China and encourage domestic alternatives. Reports of chip diversion through third countries have intensified enforcement concerns.

What to Watch

Near term, the stock's trajectory will depend on whether hyperscaler spending plans hold up and whether Nvidia can execute on its product roadmap without margin erosion. Morgan Stanley (MS) recently set a $300 price target, according to market reports, though such targets are opinions, not guarantees.

Long term, the central question is whether AI becomes a broadly monetized computing platform or remains a capital-intensive experiment concentrated among a few deep-pocketed buyers. For now, the record high reflects confidence that Nvidia remains the indispensable supplier to the AI infrastructure boom.

A spokesperson for Nvidia did not respond to a request for comment on the share move.

Correction: An earlier version of this article misstated the total buyback authorization. It is approximately $235 billion through fiscal 2028, not $250 billion.