Is Nvidia Really Just Artificially Propping Up AI Chip Demand?

© <a href="https://www.flickr.com/photos/crabchick/42326535221/" target="_blank" style="font-size: 100%">House of cards</a> (<a href="https://creativecommons.org/licenses/by/2.0/" target="_blank" style="100%">CC BY 2.0</a>) by <a href="https://www.flickr.com/people/crabchick/" target="_blank" style="100%">crabchick</a>

Nvidia (NASDAQ:NVDA) delivered a blowout earnings report yesterday, easing fears that artificial intelligence (AI) spending might be slowing. The chipmaker reported fiscal third-quarter revenue of $57 billion, up 62% year-over-year and well above estimates of $55 billion. Data center sales hit $51 billion as adjusted earnings came in at $1.30 per share, beating expectations. 

CEO Jensen Huang called Blackwell chip sales “off the charts,” and guidance for Q4 pointed to $65 billion in revenue — again topping Wall Street forecasts. Adding to the momentum, just days earlier Nvidia and Microsoft (NASDAQ:MSFT) announced they will invest up to $10 billion — and $5 billion, respectively — in Anthropic, which in turn committed $30 billion to Azure compute (largely Nvidia-powered) and up to 1 gigawatt of direct Nvidia hardware. 

This mirrors recent massive deals with OpenAI and xAI. With hyperscalers and AI labs scrambling for supply, demand looks insatiable. But is it entirely real?

The Circular Money-Go-Round Raising Eyebrows

Despite the stellar numbers that seemingly crushed the billionaire-led bear thesis, Wall Street remains divided. A persistent worry is that Nvidia is fueling much of its own explosive growth through circular investments — essentially vendor financing on steroids. 

Since 2020, Nvidia has deployed over $53 billion across 170 AI-related deals, with a staggering $23.7 billion in 2025 alone via 59 transactions, according to PitchBook. The pattern is clear: Nvidia pours cash into AI startups and cloud providers, which then turn around and spend heavily on Nvidia GPUs, either directly or through hyperscaler partners.

The fresh Anthropic deal fits perfectly. Nvidia’s up-to-$10 billion infusion helps fund Anthropic’s scaling, while Anthropic’s $30 billion Azure pledge flows to Microsoft, which buys Nvidia chips to deliver it — and Anthropic also commits to massive direct Nvidia compute purchases. 

Bulls and Bears Agree?

Similar dynamics drove earlier blockbuster investments in OpenAI (reportedly up to $100 billion in some arrangements) and xAI (perhaps as much as $15 billion). Analysts like Seaport Research Partners‘ Jay Goldberg — who maintains the lone Sell rating on Nvidia stock — call these ties “very murky” and warn they create artificial demand. Critics argue that if Nvidia’s investment faucet is turned off, a material chunk of chip orders could evaporate overnight.

Even some Nvidia bulls acknowledge the optics issue. Bernstein‘s Stacy Rasgon views the strategy as a savvy use of Nvidia’s enormous cash pile to cement ecosystem dominance in a winner-take-most market. Yet, the opacity of these private deals makes it tough to pinpoint exactly how much revenue is truly “round-tripped.” 

CFO Colette Kress highlighted on the earnings call that partners like Microsoft, Google, and Meta Platforms (NASDAQ:META) are generating real returns from AI deployments, citing their own reports of monetization progress. Still, skeptics point to historical parallels — like vendor financing during the dot-com era — that masked weakening fundamentals until the music stopped.

Broader risks lurk, too. If frontier AI models fail to deliver expected ROI fast enough, or if competitors like Advanced Micro Devices (NASDAQ:AMD), Broadcom (NASDAQ:AVGO), or custom silicon from hyperscalers gain traction, the interconnected web could unravel. Geopolitical hurdles, such as advanced AI chip sales being blocked in China, add another layer of uncertainty despite strong overall demand.

Key Takeaway

Nvidia’s investments aren’t pure smoke and mirrors — they strategically lock in customers, accelerate AI adoption, and position Nvidia as the indispensable backbone of the entire stack. Underlying demand from sovereign AI projects, enterprises, and inference workloads appears robust, backed by hyperscalers’ ongoing multi-hundred-billion capex plans and Nvidia’s half-trillion-dollar order backlog for 2025 and 2026.

That said, the circular component introduces undeniable fragility. A non-trivial slice of Nvidia’s hypergrowth relies on its own capital recycling, amplifying bubble risks if monetization lags or spending rationalizes. 

For investors, this demands vigilance: Nvidia remains the clearest AI infrastructure play, and its valuation still looks attractive despite the hypergrowth it has enjoyed, but despite the chipmaker’s cheerleading, demand could still slow in a recession. Sustained organic buying, of course, would silence the doubters, but any hint of dependency could trigger a sharp rerating.