deep-dive February 22, 2026

OpenAI at $730 Billion: The Valuation That Defies Physics

Nvidia is near a $30 billion equity stake in OpenAI. SoftBank, Amazon, and Microsoft are circling a $100 billion round. The implied $730 billion valuation would make ChatGPT’s maker the second-most valuable private company on earth. The math is extraordinary. The risk is real.

Somewhere between September’s grand announcement and this week’s revised terms, the most important financial relationship in artificial intelligence quietly restructured itself. The $100 billion infrastructure partnership between Nvidia and OpenAI - the deal that pushed Nvidia’s market cap above $5 trillion and set off a frenzy of copycat announcements across the industry - was never what it appeared to be. And now, six months later, its replacement tells a more complicated story about who actually has leverage in the AI stack.

Nvidia is reportedly close to finalizing a $30 billion equity investment in OpenAI, at a pre-money valuation of $730 billion. This is not an infrastructure agreement. There is no commensurate commitment for OpenAI to buy Nvidia chips. Nvidia gets stock. OpenAI gets capital. The arrangement is the inverse of what was announced last September - and that inversion matters.

The Deal That Dissolved

The September announcement was a letter of intent. Nvidia would invest in OpenAI over several years as it brought new AI supercomputing facilities online. The first tranche - $10 billion - would deploy when the first gigawatt of capacity was completed. The arrangement was framed as a strategic partnership: Nvidia money funding OpenAI infrastructure, built on Nvidia hardware.

Markets loved it. Analysts called it proof that the AI buildout was demand-driven rather than speculative. For a brief moment, the two most powerful companies in artificial intelligence appeared to have locked arms.

Then cracks appeared. The Wall Street Journal reported in January that the $100 billion deal was “on ice.” Nvidia’s own quarterly filing contained a careful hedge: “no assurance that we will enter into definitive agreements with respect to the OpenAI opportunity.” OpenAI began talking to other chip suppliers. The circular logic of the original deal - Nvidia investing in a customer who would then spend that money buying Nvidia chips - started attracting scrutiny it couldn’t survive.

By February, the structure had changed entirely. Nvidia is investing for equity returns, not supply chain lock-in. The two companies have separated the financial relationship from the commercial one. Whether OpenAI continues to buy Nvidia chips at scale is now a separate question from whether Nvidia profits when OpenAI’s valuation rises.

The $730 Billion Question

OpenAI is seeking to raise more than $100 billion in total from this funding round. Alongside Nvidia, the company is reportedly in discussions with SoftBank, Amazon, and Microsoft. The implied $730 billion pre-money valuation would place OpenAI just behind SpaceX as one of the most valuable private companies on earth - nearly twice the valuation Anthropic achieved in its $30 billion round earlier this month.

The arithmetic of that valuation deserves scrutiny.

OpenAI is burning through cash at an extraordinary rate. Its infrastructure costs - the compute required to train and serve its models - consume capital faster than revenue can replenish it. ChatGPT’s market share has declined from 86.7 percent to 64.5 percent over the past year. The company is now trailing Anthropic in enterprise software adoption, having spent years treating the enterprise market as secondary to its consumer product.

To bridge that gap between spending and revenue, OpenAI has begun testing ads in ChatGPT - a move that prompted Anthropic to launch a pointed advertising campaign attacking the practice. The pivot toward advertising signals that the current subscription revenue model is insufficient at the scale OpenAI requires.

So what exactly is investors paying $730 billion for?

The answer is simultaneously simple and speculative: the bet that OpenAI, despite its current financial profile, is the irreplaceable center of gravity in a market whose total size has not yet been calculated. Investors are not buying OpenAI’s current cash flows. They are buying the option on a world where AI becomes the dominant interface between humans and digital services - and OpenAI retains enough market position to extract value from that transition.

That bet could be right. It could also be the defining example of what happens when strategic importance becomes untethered from financial reality.

Nvidia’s Strategic Pivot

For Nvidia, the restructured deal reflects a more sophisticated approach to the AI stack than the original infrastructure announcement suggested. The September deal was vertical integration by another name: Nvidia buying its way into guaranteed demand. The new arrangement is pure financial exposure - a bet on OpenAI’s equity value rather than its chip consumption.

This shift is significant because it decouples Nvidia’s return from OpenAI’s hardware decisions. If OpenAI diversifies its chip supply - building more on AMD, Google TPUs, or its own custom silicon - Nvidia’s investment still appreciates if OpenAI’s overall valuation increases. The company has eliminated its supply chain dependency as the primary mechanism of return.

Jensen Huang was unambiguous in a recent CNBC interview: there was “no question” Nvidia would invest in OpenAI’s next round. The speed and certainty of that commitment, even after the original deal collapsed, signals that Nvidia views equity exposure to frontier AI companies as a core part of its capital allocation strategy - not merely a marketing exercise.

The $30 billion figure, if finalized, would represent one of the largest single-investor commitments in a private funding round in history. It would also give Nvidia a meaningful ownership stake in the company most likely to shape what people expect AI to do - which, in turn, shapes what hardware they demand to run it.

The Hardware Ambitions

Layered beneath the funding discussions is a separate development that has received less attention: OpenAI is reportedly developing its own consumer hardware, including a smart speaker. The Information broke the story this week. The project would position OpenAI directly against Amazon’s Alexa and Apple’s HomePod - not as a software partner but as a device manufacturer.

This ambition is not new. OpenAI has been exploring hardware strategies since its acquisition discussions with Jony Ive’s AI device startup began last year. What is new is the specificity: a smart speaker suggests OpenAI is targeting ambient, always-on AI interaction rather than screen-based interfaces. The living room, not the laptop.

If successful, it would give OpenAI a distribution channel outside the app stores and browser tabs controlled by Apple and Google - platforms that could, at any moment, introduce competing AI products or impose unfavorable terms. Hardware is expensive and operationally complex, but it offers a direct relationship with users that software cannot fully replicate.

The timing of the hardware push alongside the funding round is not coincidental. OpenAI needs the capital to pursue both tracks simultaneously: maintain its frontier model development and build the consumer hardware infrastructure to own the interface layer.

The Consolidation Signal

What this week’s investment news actually signals is the acceleration of a consolidation dynamic that has been building since late 2025. The frontier AI layer is narrowing. Anthropic’s $30 billion round. OpenAI’s $100 billion round. Meta’s announced $65 billion capital expenditure for AI infrastructure. Google’s continued dominance in compute with TPUs. The companies that can sustain the capital intensity of frontier model development are fewer than the number of companies currently claiming that position.

See also: NVIDIA Rubin and N1X: Rewriting the Rules of AI Hardware.

For related context, see Inference Economy: How Compute Markets Reshape Power | X01.

The arithmetic of that valuation deserves scrutiny.

OpenAI is burning through cash at an extraordinary rate. Its infrastructure costs - the compute required to train and serve its models - consume capital faster than revenue can replenish it. ChatGPT’s market share has declined from 86.7 percent to 64.5 percent over the past year. The company is now trailing Anthropic in enterprise software adoption, having spent years treating the enterprise market as secondary to its consumer product.

To bridge that gap between spending and revenue, OpenAI has begun testing ads in ChatGPT - a move that prompted Anthropic to launch a pointed advertising campaign attacking the practice. The pivot toward advertising signals that the current subscription revenue model is insufficient at the scale OpenAI requires.

So what exactly is investors paying $730 billion for?

The answer is simultaneously simple and speculative: the bet that OpenAI, despite its current financial profile, is the irreplaceable center of gravity in a market whose total size has not yet been calculated. Investors are not buying OpenAI’s current cash flows. They are buying the option on a world where AI becomes the dominant interface between humans and digital services - and OpenAI retains enough market position to extract value from that transition.

That bet could be right. It could also be the defining example of what happens when strategic importance becomes untethered from financial reality.

Nvidia’s Strategic Pivot

For Nvidia, the restructured deal reflects a more sophisticated approach to the AI stack than the original infrastructure announcement suggested. The September deal was vertical integration by another name: Nvidia buying its way into guaranteed demand. The new arrangement is pure financial exposure - a bet on OpenAI’s equity value rather than its chip consumption.

This shift is significant because it decouples Nvidia’s return from OpenAI’s hardware decisions. If OpenAI diversifies its chip supply - building more on AMD, Google TPUs, or its own custom silicon - Nvidia’s investment still appreciates if OpenAI’s overall valuation increases. The company has eliminated its supply chain dependency as the primary mechanism of return.

Jensen Huang was unambiguous in a recent CNBC interview: there was “no question” Nvidia would invest in OpenAI’s next round. The speed and certainty of that commitment, even after the original deal collapsed, signals that Nvidia views equity exposure to frontier AI companies as a core part of its capital allocation strategy - not merely a marketing exercise.

The $30 billion figure, if finalized, would represent one of the largest single-investor commitments in a private funding round in history. It would also give Nvidia a meaningful ownership stake in the company most likely to shape what people expect AI to do - which, in turn, shapes what hardware they demand to run it.

The Hardware Ambitions

Layered beneath the funding discussions is a separate development that has received less attention: OpenAI is reportedly developing its own consumer hardware, including a smart speaker. The Information broke the story this week. The project would position OpenAI directly against Amazon’s Alexa and Apple’s HomePod - not as a software partner but as a device manufacturer.

This ambition is not new. OpenAI has been exploring hardware strategies since its acquisition discussions with Jony Ive’s AI device startup began last year. What is new is the specificity: a smart speaker suggests OpenAI is targeting ambient, always-on AI interaction rather than screen-based interfaces. The living room, not the laptop.

If successful, it would give OpenAI a distribution channel outside the app stores and browser tabs controlled by Apple and Google - platforms that could, at any moment, introduce competing AI products or impose unfavorable terms. Hardware is expensive and operationally complex, but it offers a direct relationship with users that software cannot fully replicate.

The timing of the hardware push alongside the funding round is not coincidental. OpenAI needs the capital to pursue both tracks simultaneously: maintain its frontier model development and build the consumer hardware infrastructure to own the interface layer.

The Consolidation Signal

What this week’s investment news actually signals is the acceleration of a consolidation dynamic that has been building since late 2025. The frontier AI layer is narrowing. Anthropic’s $30 billion round. OpenAI’s $100 billion round. Meta’s announced $65 billion capital expenditure for AI infrastructure. Google’s continued dominance in compute with TPUs. The companies that can sustain the capital intensity of frontier model development are fewer than the number of companies currently claiming that position.

For Nvidia, getting equity in both Anthropic and OpenAI - the two leading frontier AI labs - is a hedge that ensures it profits from the consolidation regardless of which lab wins. For SoftBank, the OpenAI bet is a do-over of its Vision Fund era ambitions, now focused on a single company rather than a portfolio. For Amazon, Microsoft, and Google, the participation is about ensuring they are not frozen out of the most important AI relationships as the field narrows.

The $730 billion valuation is an artifact of this dynamic. It reflects not what OpenAI earns today but what multiple of the world’s largest technology companies believe they cannot afford to be excluded from. When strategic fear drives capital allocation, numbers stop following conventional financial logic.

Whether that logic ever reasserts itself - and what OpenAI looks like when it does - is the question that will define the next phase of the AI era.