Nvidia (NVDA) Defends AI Chip Longevity Against Michael Burry's Criticism as Stock Hits Record High
Key Takeaways
- •Nvidia's stock rose about 3% on Friday to an intraday record near $237.75, exceeding its May all-time closing high and bringing the year-to-date gain to roughly 27%.
- •Michael Burry contends that major technology companies may be inflating current earnings by assuming overly long useful lives for AI hardware, a dispute centered on depreciation schedules for servers and data center equipment.
- •Nvidia has responded to the accounting criticism by pointing out that its largest AI hardware customers, Meta Platforms and Google, have lengthened their AI infrastructure depreciation periods from four years to six.
- •Nvidia's board approved an incremental $150 billion share repurchase, expanding total buyback authorizations to $235 billion through fiscal 2028, the largest program ever announced by a corporation.
- •Morgan Stanley upgraded Nvidia to its top semiconductor recommendation after meetings with CEO Jensen Huang, noting that AI infrastructure bottlenecks are shifting from chip supply toward data center real estate, power availability, and financing.

Nvidia (NVDA) shares advanced approximately 3% during Friday's trading session, reaching a new intraday record and surpassing the all-time closing peak the chipmaker established in May. Shares traded near $237.75, extending momentum that has delivered a year-to-date gain of roughly 27%.
The advance followed a week in which the company paired its market milestone with a major capital-return decision, and the rally was powered by two distinct catalysts unfolding simultaneously: an intensifying debate over the actual useful life of AI chips, and a shareholder-return initiative of unprecedented scale.
The Depreciation Controversy
Michael Burry, the investor famous for forecasting the 2008 financial crisis, has raised concerns that technology giants are being too conservative in estimating how quickly their AI hardware becomes obsolete. Accounting choices of that kind, he argues, could artificially inflate current earnings figures. Depreciation schedules determine how quickly companies expense servers and data center equipment, making the assumptions behind them a sensitive subject for corporations spending billions on AI computing capacity. At its core, the dispute turns on how long a GPU remains economically productive: the shorter the assumed life, the faster the costs hit the income statement and the lower the reported profits during the heavy buildout years.
Nvidia is not sitting idle. The semiconductor manufacturer is pointing to evidence from key clients, noting that Meta Platforms and Google — both among the largest buyers of AI computing hardware — have extended their AI infrastructure depreciation timeframes from four years to six. Nvidia offers this as a direct counter to those questioning the industry's accounting practices: the companies purchasing the most AI hardware are, by their own accounting judgments, concluding the equipment stays useful longer, not shorter.
Why the Buyback Matters
The second catalyst was more transparent. Earlier this week, Nvidia's board approved an incremental $150 billion share repurchase authorization. Combined with existing authorizations, the company now has $235 billion available for buybacks extending through fiscal 2028 — the largest share repurchase program ever announced by any corporation.
The magnitude of the program sends a clear message: Nvidia's leadership is confident the company will continue generating substantial cash flow despite heavy investment in AI infrastructure expansion. Buybacks of this scale also return capital directly to shareholders by reducing the outstanding share count. Because an authorization functions as a ceiling rather than a schedule, execution takes place over time within the fiscal 2028 window, giving the program a multi-year horizon.
Wall Street Responds
The Street responded positively. Morgan Stanley elevated Nvidia to its top semiconductor recommendation on Friday after conducting meetings with CEO Jensen Huang and senior management, citing ongoing robust AI demand and an expanding customer ecosystem.
The investment bank also identified an emerging trend deserving attention. According to the firm, AI infrastructure constraints are transitioning away from semiconductor availability toward data center real estate, electrical power availability, and financing capacity. That marks an important evolution in the narrative: once chips are no longer the limiting factor, advantage shifts to organizations controlling affordable electricity and strategically located facilities, making power procurement and site availability the constraints to monitor next.
The Bigger AI Picture
Nvidia's moves are occurring within a broader AI investment landscape, as the technology sector increasingly focuses on what follows the initial infrastructure buildout phase. Anthropic's anticipated public offering is contributing to the momentum. A stock market debut for the company behind Claude would provide another significant benchmark for investor appetite toward AI companies requiring massive capital to scale operations.
For Nvidia in particular, the dual narrative of historic shareholder returns combined with an active defense against accounting critics provides renewed momentum entering the fourth quarter. Rather than allowing the depreciation discussion to fester quietly, the company has chosen to engage directly.
Nvidia also collaborates extensively with financial institutions to facilitate approximately $500 billion in capital deployment supporting the AI ecosystem — a figure that illustrates the enormous stakes involved in debates over hardware longevity and infrastructure spending patterns. How the story develops from here will be visible in concrete data points: hyperscalers' depreciation disclosures, the pace of buyback execution against the $235 billion authorization, and any firm timing attached to Anthropic's anticipated debut.
Source: Blockonomi