OpenAI Raised $122B at $852B. Amazon's $35B Is Contingent on AGI.

The largest private financing deal in Silicon Valley history closed on March 31, 2026. OpenAI raised $122 billion at a post-money valuation of $852 billion. To put that number in context: $852 billion is larger than the entire GDP of the Netherlands. It is more than the market capitalisation of JPMorgan Chase. It is the highest valuation ever achieved by a private company in any industry, in any country.

And $14 billion of that company's revenue — produced in a year where it earned $2 billion per month — will still be lost this year.

The gap between those two facts is where the most important questions about OpenAI live. Here is the full breakdown of the $122 billion round, the investors, the clauses, and what the IPO roadmap actually looks like.

Who Put In What: The Investor Breakdown

The round was led by three strategic investors whose combined $110 billion accounts for 90 percent of the total raise:

Amazon: $50 billion — the largest single commitment in the round. Amazon's investment is split: $15 billion is unconditional. The remaining $35 billion is contingent on one of two events occurring by the end of 2028: either OpenAI completes an IPO, or OpenAI achieves what the investment documents define as artificial general intelligence. This is the most unusual clause in the round and the one with the most significant implications.

SoftBank: $30 billion — SoftBank CEO Masayoshi Son has been one of the most vocal AGI believers in the investment world. His $30 billion commitment to OpenAI follows SoftBank's leadership of the $500 billion Stargate initiative announced in January 2026. SoftBank's position across OpenAI and Stargate makes it the largest single private bet on AI infrastructure in history.

Nvidia: $30 billion — Nvidia's investment is strategic in the most direct possible sense. OpenAI is Nvidia's largest single GPU customer. Nvidia's H100 and H200 clusters form the backbone of OpenAI's training and inference infrastructure. An equity stake in OpenAI at $852 billion, if the company reaches a $1 trillion IPO valuation, is also a significant financial return. But the more important dynamic is the signal: Nvidia is deepening its alignment with OpenAI at a moment when competitors including Google (which has its own TPUs), Amazon (Trainium chips), and Microsoft (Maia chips) are all building custom silicon specifically to reduce Nvidia dependence.

Microsoft: undisclosed participation — Microsoft has committed $13 billion to OpenAI in previous rounds since 2019. Its participation in the March 2026 round was confirmed but the amount was not disclosed. Microsoft's position is complex: it is simultaneously OpenAI's largest cloud infrastructure partner, a major revenue source through the Azure OpenAI Service, and a competitor through its own Copilot products.

Other investors: Andreessen Horowitz, D.E. Shaw Ventures, MGX (UAE sovereign wealth), TPG, and T. Rowe Price Associates also participated. For the first time, $3 billion was raised from individual retail investors through bank channels — a structural innovation that creates a public-market-like investor base before an IPO exists.

A $4.7 billion revolving credit facility was also established, supported by JPMorgan Chase, Citi, Goldman Sachs, Morgan Stanley, Wells Fargo, Mizuho, Royal Bank of Canada, SMBC, UBS, HSBC, and Santander.

Amazon's $35B Contingency Clause: What AGI Means in a Legal Contract

The most consequential detail in this funding round is not the total amount or the valuation. It is the clause in Amazon's investment agreement that makes $35 billion contingent on either an IPO or AGI.

What does "AGI" mean in a legal contract?

This is not a philosophical question. Amazon's lawyers and OpenAI's lawyers had to agree on a specific definition that could be tested objectively in a court if disputed. That definition has not been publicly disclosed in full. But the existence of the clause tells you several things:

First, both parties believe AGI is a plausible event horizon. You do not write a legal clause around something you consider impossible. The fact that Amazon's lawyers agreed to a definition both parties could sign means there is a specific technical or capability milestone that both companies believe might be reached before 2028.

Second, the IPO alternative creates a natural hedge. If OpenAI goes public before 2028 — which its confidential SEC filing suggests is the intent — Amazon's $35 billion triggers regardless of AGI. The AGI clause is the more exotic pathway; the IPO is the base case.

Third, the clause makes Amazon a very interested party in OpenAI's AGI timeline communication. Any public statement OpenAI makes about AGI progress that could be interpreted as meeting the threshold definition will trigger legal and financial consequences. This creates a real tension between OpenAI's research transparency interests and its financial structure.

$2 Billion Monthly Revenue, $14 Billion Annual Losses: The Financial Reality

OpenAI is generating $2 billion in revenue per month. Annualised, that is $24 billion. By almost any measure, that is an extraordinary figure for a company that did not have a commercial product until late 2022.

It is also not close to covering costs. OpenAI's estimated 2026 losses are $14 billion against that $24 billion revenue run rate. The company is spending roughly $38 billion per year to produce $24 billion per year.

Where does the money go? Three primary categories:

Compute costs are the largest single expense. Training frontier models requires thousands of H100 or H200 GPUs running continuously for weeks or months. Inference — serving ChatGPT responses to the 200 million weekly active users who pay for or use for free OpenAI products — runs continuously at scale. Compute alone consumes an estimated $3 to $5 billion per year at current usage levels.

Research and development is the second major category. OpenAI employs several thousand researchers, engineers, and safety staff. The compensation to attract and retain frontier AI research talent is at the top of the technology industry pay scale.

Infrastructure and energy — data centre costs, electricity, cooling, networking — round out the primary expense categories.

The $14 billion annual loss figure is not a sign of mismanagement. It is a deliberate bet: OpenAI is spending at a loss now because the company believes that whoever achieves the next threshold of AI capability will capture an outsized share of a market that could be worth trillions annually. The $122 billion raise extends the runway for that bet to play out.

The company's revenue growth has been rapid enough that profitability is not structurally impossible — it is a question of whether compute costs scale slower than revenue as efficiency improves. The trend in inference costs (which have fallen more than 95 percent for GPT-4-class capabilities since 2023) suggests the economics improve over time.

The Nonprofit to For-Profit Conversion: Why It Changes Everything

OpenAI was founded as a nonprofit in 2015. In 2019, it created a "capped-profit" subsidiary to raise venture capital, but the nonprofit retained control and the for-profit returns were capped. This structure is why OpenAI could not do a traditional IPO: you cannot list a nonprofit-controlled company on public markets in standard form.

The conversion to a fully for-profit corporation began in late 2024. The process has faced legal challenges from state attorneys general concerned about the conversion of charitable assets to private ownership, and from Elon Musk through ongoing litigation. As of mid-2026, the conversion is substantially complete but some legal proceedings remain open.

The OpenAI Foundation — the successor entity to the original nonprofit — received equity in the for-profit company as compensation for transferring control. At the $852 billion valuation, the Foundation's stake is worth approximately $180 billion. That number creates an unusual situation: the world's wealthiest charitable foundation (by a large margin — the Gates Foundation has roughly $50 billion in assets) is also an OpenAI shareholder with significant financial interest in the company's commercial success.

The for-profit conversion is not just a legal formality. It changes OpenAI's governance, accountability, and incentive structure in ways that its founders and critics have both noted. The original structure was designed specifically to prevent commercial pressure from overriding safety considerations. Whether the new structure maintains those protections is a live debate inside and outside the company.

The IPO Roadmap: Q4 2026 Target, $1 Trillion Goal

OpenAI has confidentially filed for an IPO with the U.S. Securities and Exchange Commission. The company has stated publicly that it has not decided on timing and that going public may still take a while. Most analysts and investment bankers tracking the process expect a public listing in Q4 2026 or early 2027.

The valuation target for the IPO is widely reported as $1 trillion. That would represent a 17 percent premium to the $852 billion private round valuation. For comparison, Nvidia crossed $1 trillion in market cap in May 2023 and now trades above $3 trillion. Apple reached $1 trillion in August 2018.

An OpenAI IPO at $1 trillion would be:

• The largest IPO in US history, surpassing Saudi Aramco's $25.6 billion raise in 2019
• The highest IPO valuation in history if the offering prices above Alibaba's $231 billion valuation at its 2014 IPO
• The first listing of a company whose primary product is an AI model rather than a platform, marketplace, or device

The IPO also unlocks the $35 billion contingent portion of Amazon's investment. That alone is a significant financial incentive for both parties to move the listing forward on schedule.

What This Means for Developers and the AI Infrastructure Stack

The $122 billion raise has direct implications for the infrastructure and developer tools ecosystem.

Compute demand locks in Nvidia's position. OpenAI's cash allows it to commit to GPU clusters at a scale that only hyperscalers previously reached. For Nvidia, this is a guaranteed demand floor for H200 and next-generation Blackwell and Rubin architecture chips through at least 2028.

The Azure OpenAI Service accelerates. Microsoft's continued participation and Azure's role as OpenAI's primary cloud infrastructure mean Azure OpenAI API capacity grows with OpenAI's compute investment. Developers building on GPT-4o and future models through Azure can expect continued capacity expansion.

OpenAI's API pricing competition intensifies. With $122 billion in the bank, OpenAI can afford to price its API aggressively to defend market share against Anthropic (which raised $3.5 billion in 2024) and Google's Gemini. The era of frontier model pricing dropping 80 to 95 percent per year is likely to continue.

The AGI clause creates a new kind of deadline. Amazon's $35 billion contingency creates a financial incentive — not just a research one — for OpenAI to reach defined AGI milestones before 2028. That pressure flows through to every aspect of OpenAI's research and product roadmap. Developers building on OpenAI's platform should expect an unusually aggressive product release cadence through 2027.

Our Analysis: The Valuation Is a Bet on a Winner-Take-Most Market

$852 billion for a company losing $14 billion per year is not a valuation that makes sense against traditional discounted cash flow models. It makes sense only if you believe the AI market is winner-take-most and that OpenAI is the company that wins it.

That is the thesis every investor in this round is buying. Amazon's $50 billion, Nvidia's $30 billion, and SoftBank's $30 billion are not bets that OpenAI will eventually become profitable in a competitive market. They are bets that OpenAI will be the dominant platform for AI in the way that AWS is the dominant platform for cloud, and that dominant platform position is worth buying at any valuation that reflects a realistic path to getting there.

Whether that thesis plays out depends on factors that no investor in the round controls: whether OpenAI continues to produce the most capable models, whether regulation constrains frontier AI development in ways that favour incumbents or entrants, and whether the AGI capability threshold — whenever it arrives — concentrates value in one company or distributes it across an ecosystem.

The $122 billion raise buys time, compute, and talent to run that race. Whether OpenAI wins it is the open question the next two years will begin to answer.

Key Takeaways

• $122 billion raised at $852 billion valuation on March 31, 2026 — the largest private financing deal in Silicon Valley history; investors include Amazon ($50B), SoftBank ($30B), Nvidia ($30B), Microsoft (undisclosed), and for the first time retail investors ($3B through bank channels)
• Amazon's $35B contingency clause — $15B of Amazon's commitment is unconditional; the remaining $35B triggers only on OpenAI completing an IPO or achieving AGI by end of 2028; the specific AGI definition agreed in the legal documents has not been publicly disclosed
• $2B monthly revenue, $14B annual losses — OpenAI is growing revenue rapidly but spending $38B per year to produce $24B; the loss is deliberate: a bet that capability dominance now translates to market dominance later
• Nonprofit-to-for-profit conversion substantially complete — the OpenAI Foundation's equity stake is worth ~$180B at the $852B valuation, making it the world's wealthiest charitable foundation by a substantial margin; legal challenges from state AGs and Elon Musk remain open
• IPO confidentially filed with SEC — target timing Q4 2026 or early 2027; valuation target reported at $1 trillion; would be the largest IPO in US history by amount raised and highest IPO valuation ever
• Developer implication: OpenAI's $122B runway accelerates compute investment, API pricing competition, and product release cadence through 2027 — the AGI clause creates a financial deadline that makes OpenAI's roadmap more aggressive than any research-only incentive would

Sources

• OpenAI Valued at $852 Billion After Completing $122 Billion Round — Bloomberg
• OpenAI Raises $122 Billion to Accelerate the Next Phase of AI — OpenAI Official
• OpenAI Not Yet Public, Raises $3B from Retail Investors in Monster $122B Fund Raise — TechCrunch
• OpenAI Closes Funding Round at $852 Billion Valuation — CNBC
• OpenAI IPO 2026: $852B Valuation, Risks and Bull Case — Tech Market Briefs
• OpenAI Revenue, Losses, and IPO Valuation: Forecasts Through Late 2027 — Future Search
• Our coverage: GitHub Copilot Drops Flat Pricing — What AI Credits Mean for Dev Teams