Nvidia’s $100 billion OpenAI bet: Strategic masterstroke or high-stakes gamble?

On 22 September, the $4.5 trillion chipmaker announced it would invest up to $100 billion in OpenAI, the maker of ChatGPT, and in return become the company’s primary supplier of AI chips for years to come. The news alone sent Nvidia’s market cap soaring by $180 billion in a single day.

But behind the blockbuster numbers lies a far more complicated story. The deal exposes Nvidia’s reliance on a handful of customers, its vulnerability as a hardware company chasing relentless upgrade cycles, and the risk of alienating Big Tech clients while betting on a combative new rival.

It also raises uncomfortable echoes of the vendor-financing schemes that fuelled the dot-com bubble.

Let's unpack this gripping deal.

The circular money machine

OpenAI plans to build data centres with a combined capacity of 10 gigawatts (GW)—roughly the output of ten nuclear power plants. To get there, it will need 4-5 million of Nvidia’s next-generation AI chips, the “Vera Rubin" platform that hasn’t even launched yet.

Here's where it gets deliciously circular. Nvidia will invest $10 billion in OpenAI for every GW of computing capacity built, starting with the first deployment in late 2026.

Analysts say OpenAI will cover about 71% of these costs in cash, while the remaining 29% will come as equity funding from Nvidia. In effect, Nvidia bankrolls OpenAI’s expansion by buying into the company, money that OpenAI then turns around to spend on Nvidia’s own chips. Bank of America estimates this feedback loop could generate $300-500 billion in revenue for Nvidia over time.

The timing makes the spectacle even more striking. Just months earlier, OpenAI inked a separate $300 billion agreement with Oracle to build more data centres over five years. With annual revenue of only $13 billion, the numbers are starting to look surreal. It’s classic Silicon Valley economics: spend tomorrow’s hypothetical money today, only this time on a scale the Valley has never seen.

The Achilles heel revealed

For all of Nvidia’s dominance, the deal with OpenAI shines a light on a vulnerability that likely keeps chief executive officer Jensen Huang awake at night.

The company’s latest filings show just how concentrated its business is: six customers account for 85% of sales, with a single client contributing 23% of total revenue. And these aren’t ordinary customers, they’re giants like Google, Amazon, Meta, and Microsoft, each pouring resources into reducing their dependence on Nvidia’s chips.

But customer concentration is only part of the problem. Nvidia’s real Achilles heel is that its business model remains tied to hardware sales. Unlike software, which can scale almost infinitely, hardware is constrained by upgrade cycles and eventual market saturation.

That’s why Huang has long pushed for aggressive product release schedules—once every two years, now compressed to a blistering one-year cadence in the AI era. The company needs its customers to constantly refresh their systems to sustain the growth rates baked into its nearly $4.5 trillion valuation.

Warning signs are already emerging. The "second derivative" of Nvidia's growth is slowing, with the most recent quarter marking the lowest growth rate since 2023. While growth is expected to reaccelerate, consensus projections show a steep decline from 58% growth in 2025 to just 17% by 2027.

More concerning, investor scrutiny on AI capital spending is intensifying, with expectations that 2026 will mark a significant slowdown in customer investments.

The worsening frenemies problem

Here’s the twist: OpenAI isn’t just another customer. It’s actively competing with Nvidia’s biggest clients.

Sam Altman, OpenAI's charismatic but combative CEO, has seemingly been throwing digital punches at everyone. He claims he “can't remember the last time he used Google," wants to build a social network to compete directly with Meta, has partnered with Shopify to challenge Amazon's e-commerce dominance, and acquired Johnny Ive's design company to build hardware that could rival Apple. Oh, and he famously has a rocky relationship with Microsoft despite their $13 billion investment in his company.

Meanwhile, Nvidia’s traditional customers are far from passive. Google leans heavily on its own TPUs, Amazon is making Anthropic’s massive clusters run on Trainium chips, and rivals from AMD to specialty startups are being courted as alternatives. Nvidia itself has been hedging its bets, investing in “neoclouds" like CoreWeave and boosting GPU allocations to Oracle—moves that effectively pit it against its own largest clients.

It’s like running the only burger joint in town, where six regulars account for nearly 85% of your sales—until you overhear them plotting to open rival restaurants. Your solution? Hand out free meals to a flashy newcomer who also plans to compete with your regulars. It may buy time, but it almost guarantees those regulars will race to find another supplier.

The infrastructure reality check

The deal also highlights some uncomfortable practical realities.

OpenAI's planned 10GW of additional power capacity represents almost half of all utility-scale electricity generation added in America during the first half of this year. Even with relaxed infrastructure permitting, bringing this online could take years.

Altman himself acknowledged three major challenges: pushing AI research frontiers, building compelling user products, and overcoming "unprecedented infrastructure challenges" around chips and power supply.

The response to OpenAI's latest model, GPT-5, has been notably underwhelming despite ChatGPT's 700 million weekly active users. This raises questions about whether the massive infrastructure investments will generate proportional returns, or if the industry is building digital cathedrals for a congregation that may not materialize as expected.

Financial theatre or strategic masterstroke?

Some analysts are calling this arrangement “financial theatre"—and the parallels are hard to ignore. It echoes the vendor-financing schemes of the early 2000s telecom bubble, when Nortel and Lucent fuelled their own sales by lending money to customers.

The key difference is that Nvidia is offering equity instead of debt, eliminating repayment obligations but creating new risks around circular dependencies.

Critics argue this reveals weakness rather than strength. If Nvidia truly had unassailable technological superiority and unlimited customer demand, why would it need to fund its own sales? The company has already had to publicly clarify that it won't give OpenAI preferential treatment amid rising antitrust concerns, suggesting regulatory scrutiny may be inevitable.

Then there’s the valuation backdrop. At 41 times forward earnings, Nvidia is priced at levels unseen since the dawn of the AI boom. With consensus forecasts pointing to growth slowing over the next few years, the timing of this deal appears designed to extend the current investment cycle and justify valuations that may already reflect several years of future growth.

The verdict: Brilliant desperation

Nvidia’s $100 billion investment in OpenAI is the ultimate Silicon Valley paradox—a move that is at once brilliant and desperate. Brilliant because it creates artificial demand for Nvidia’s products while positioning the company as kingmaker in the AI ecosystem. Desperate because it exposes the fragility of relying on a handful of customers actively working to replace you.

The deal essentially turns Nvidia into a venture capitalist using its own products as currency, betting that the AI revolution will continue indefinitely and that infrastructure spending will justify today's valuations. It's like a high-stakes game of musical chairs where everyone knows the music will eventually stop, but nobody wants to be the first to sit down.

Whether this proves to be strategic genius or costly folly hinges on two questions: will artificial general intelligence arrive as promised, and will the astronomical infrastructure investments generate proportional economic value?

One thing is certain: Silicon Valley has never seen anything quite like this circular financing arrangement, where the world’s most valuable company essentially funds its own growth by investing in its customers’ dreams of digital dominance.

What's certain is that Silicon Valley has never seen anything quite like this circular financing arrangement, where the world's most valuable company essentially funds its own growth by investing in its customers' dreams of digital dominance.

Disclaimer: This article is for general educational purposes only and does not constitute an offer, recommendation, or solicitation to buy or sell any securities. It may contain forward-looking statements, and actual outcomes can vary. Past performance is not a guarantee of future results.

Neither the information herein nor any opinion expressed should be construed as investment advice. The information and opinions were considered valid by VF Securities, Inc. at the time of publication. Anyone relying on this content does so at their own risk.

Securities markets may experience rapid and unexpected price movements. Investors must conduct independent analysis with their own legal, tax, and financial advisors before making any investment decisions.