Anthropic Is Borrowing $60 Billion to Fund Google's Chips at Over 10% Interest. Project North Star Is the Largest AI Chip Financing in History — and the Clearest Window Yet Into What Frontier AI Actually Costs.

Bloomberg reported on September 29, 2026 that OpenAI is in talks with Abu Dhabi's MGX fund and BlackRock to raise at least $30 billion in a new funding round that would value the company at approximately $1.4 trillion. The report, sourced from people familiar with the negotiations, marks the most significant private-market AI valuation event since OpenAI's March 2026 round closed at a $122 billion raise and an $852 billion post-money valuation. At $1.4 trillion, OpenAI would be worth more than every publicly traded company in the world except Apple, Nvidia, Microsoft, and Saudi Aramco. It would also be worth approximately as much as Meta — a company that generates more than $50 billion in annual operating income.

OpenAI's case for that valuation rests on a revenue trajectory that is real, a product position that is genuinely dominant in consumer AI, and a set of platform assumptions about where AI is going over the next decade. This analysis examines those assumptions, the mechanics of the round, the reasons the IPO slipped to 2027, and what the fundraise signals for the enterprise AI buyers who have built critical workflows on OpenAI's models.

The Round: $30 Billion at $1.4 Trillion

The $30 billion round would be the largest private AI funding event since OpenAI's March 2026 raise. At the time of Bloomberg's September 29 report, OpenAI had not confirmed the terms and the final investors had not been locked in. The two most discussed anchor investors were MGX — an Abu Dhabi-based AI fund backed by UAE sovereign wealth infrastructure — and BlackRock, the world's largest asset manager. MGX has discussed contributing up to $10 billion, which would make it the single largest check in the round. BlackRock is in discussions to participate alongside MGX as part of a broader syndicate.

The March 2026 round context matters for understanding the magnitude of this raise. In March, OpenAI closed $122 billion at an $852 billion post-money valuation — itself a record for a private AI company at the time. The new round, if completed at $1.4 trillion pre-money, would represent a 64% increase in valuation in approximately seven months. For context, public market tech companies almost never achieve this kind of valuation acceleration in a seven-month window without a fundamental change in their business — a major product launch, an M&A event, or a category-defining financial result. OpenAI's case is that the combination of GPT-6.1 Sol's launch in October, the Stargate infrastructure buildout, and the ChatGPT consumer franchise has compounded its revenue growth to a level that justifies the step-change.

The round is explicitly framed as a pre-IPO raise. OpenAI CEO Sam Altman told Fortune on September 12 that going public in 2026 would be "ill-advised" and that the company needs to remain focused on AI safety governance. A 2027 IPO is the working assumption across investor discussions. The $30 billion private round, if it closes, eliminates any financial pressure to accelerate the IPO timeline — OpenAI can reach 2027 on private capital.

There is one discrepancy in the reporting worth flagging. At least one outlet cited a "$140 billion pre-money valuation" figure, which conflicts sharply with the $1.4 trillion number from Bloomberg, Fortune, the Financial Times, the Wall Street Journal, and Semafor. The $140 billion figure appears to be a reporting error — OpenAI's post-money valuation from its March round alone was $852 billion, making a $140 billion pre-money impossible for a new raise. The $1.4 trillion is the figure from the five primary outlets that covered the story.

The Revenue Base: What OpenAI Is Actually Making

The most contested fact in the OpenAI fundraise story is the revenue number. Bloomberg's reporting placed annualized revenue above $40 billion as of late September 2026. Other outlets have cited figures closer to $70 billion. The gap is large enough to be material to any valuation analysis, and it likely reflects different measurement methodologies — whether the figure includes committed enterprise contract value versus actual recognized revenue, whether it counts consumer subscriptions at face value or net of refunds, and whether the timing is annualized from a single quarter or based on a trailing twelve-month figure.

What is not in dispute is the trajectory. OpenAI launched GPT-6.1 Sol and the Dots always-on agent at DevDay in October 2026, extending its product lead in both enterprise API and consumer AI. ChatGPT has more than 500 million active users. The enterprise API is used by tens of thousands of businesses for production workloads — customer service automation, document processing, software development, content generation. OpenAI's operator ecosystem generates significant revenue through both direct API billing and revenue-share arrangements with consumer operators.

The revenue mix matters for the valuation. Consumer subscription revenue (ChatGPT Plus at $20/month, ChatGPT Pro at $200/month, the $500/month tier introduced in October 2026) is high-margin and recurring but faces churn risk as competitors improve. Enterprise API revenue is growing rapidly but is under ongoing price competition — GPT-6.1 Sol launched at $2 per million input tokens, a significant reduction from prior frontier model pricing, and Luna at $0.10 per million is essentially priced as a commodity inference service. The Stargate compute infrastructure, while enormous, is a cost center funded through debt and equity rather than a revenue driver.

The revenue trajectory that justifies a $1.4 trillion valuation requires OpenAI to grow from its current annualized run rate to at least $80-100 billion in annual revenue within the next two to three years — and to do so with operating margins that demonstrate a path to profitability. That is not impossible. But it is a demanding assumption that the market will scrutinize intensively when OpenAI files its IPO prospectus.

Why the IPO Slipped to 2027

Sam Altman's September 12 Fortune comment that a 2026 IPO would be "ill-advised" was widely read as a safety-governance comment, and that is what he said. But there are at least three other structural reasons why 2026 was always unlikely and 2027 is the earliest feasible window.

The nonprofit conversion. OpenAI is in the process of converting from its unusual hybrid structure — a capped-profit LLC overseen by a nonprofit board — to a conventional Delaware for-profit corporation. This conversion requires the existing nonprofit entity (OpenAI Inc.) to be compensated for giving up its governance rights and its economic stake in the for-profit subsidiary. The compensation package is complex: it involves a mix of cash, a stake in the new entity, and a commitment to fund OpenAI's charitable mission. Negotiating that compensation, getting state attorneys general approval in California and Delaware, and filing the amended certificates of incorporation takes time. Going public before the structural conversion is complete would create major disclosure complications in the S-1.

Sam Altman's equity situation. Under OpenAI's original structure, Altman received no equity — he joined as CEO without the standard founder's stake because of the nonprofit governance model. This created a peculiar situation where the company's CEO and public face has no direct financial stake in its success. Addressing this through a new equity grant or restricted stock plan is straightforward in principle but requires board approval, shareholder agreement, and potentially IRS review given the tax treatment of grants in a converting entity. The equity situation needs to be resolved cleanly before an S-1 would be digestible to public market investors.

The revenue growth window. A company going public at $1.4 trillion needs to show public market investors a multi-year revenue growth story that justifies the multiple. The optimal moment to go public is when growth is still accelerating — not when it is decelerating. If OpenAI's annualized revenue is $40-70 billion in October 2026 and growing at 30-40% annually, the optimal IPO window is when it can show $80-100 billion in trailing revenue with sustained growth. That window opens in mid-to-late 2027 on even conservative assumptions.

The Investor Economics: MGX, BlackRock, and Sovereign Capital

The choice of MGX and BlackRock as anchor investors in a pre-IPO round tells a specific story about where the capital for frontier AI is coming from and why.

MGX is not a traditional venture capital fund. It is an Abu Dhabi investment platform established in 2024 with backing from Mubadala Investment Company — one of the UAE's largest sovereign wealth funds — and ADQ, another Abu Dhabi sovereign entity. MGX's explicit mandate is to invest in AI and advanced technology, both directly and through partnerships with major US technology companies. It was one of the investors in the Stargate joint venture announced in January 2026, which committed $500 billion over four years to US AI infrastructure built primarily by Microsoft, SoftBank, and Oracle.

MGX's interest in OpenAI is therefore a strategic extension of the broader UAE-US AI partnership rather than purely a financial investment. The UAE has been one of the most aggressive sovereign actors in AI infrastructure buildout — it has committed to building large-scale data centers, attracting AI researchers, and partnering with American AI companies to establish itself as a global AI hub. A $10 billion stake in OpenAI is simultaneously a financial investment and a diplomatic and industrial policy instrument.

BlackRock's participation follows from its Global Infrastructure Partners acquisition and its subsequent launch of an AI infrastructure investment fund with Microsoft, announced in 2024. BlackRock manages $10 trillion in assets and has been building exposure to AI infrastructure through direct data center investments, private credit to AI companies, and now equity stakes in pre-IPO AI companies. Anthropic's Project North Star $60 billion chip financing deal — with Blackstone, not BlackRock — is a parallel example of large institutional capital flowing into frontier AI infrastructure at scale. The pattern across both deals is the same: institutional capital that cannot wait for public market access is writing very large checks to secure positions in the AI companies they believe will dominate.

The geopolitical dimension is real and worth naming. The United States has used export controls to limit Chinese access to advanced semiconductors. The UAE has positioned itself carefully between the US and China, maintaining relationships with both, and its investment in OpenAI is partly about ensuring access to frontier AI technology regardless of how the US-China AI competition evolves. For OpenAI, accepting sovereign capital from UAE funds is a calculated trade: the capital is essential, and the strategic relationship with a major Gulf sovereign is useful for international expansion. The risk is that dependence on sovereign capital creates governance complications that public market investors will scrutinize.

The Valuation Math: What $1.4 Trillion Requires

The cleanest way to think about whether $1.4 trillion is defensible is to work backward from what OpenAI would need to be worth at IPO to give investors a reasonable return on the pre-IPO price.

If MGX and BlackRock invest at $1.4 trillion and target a 2x return (a relatively modest target for a pre-IPO position), they need OpenAI to IPO at $2.8 trillion or higher, or to return capital through dividends or buybacks equivalent to the target. A $2.8 trillion market capitalization at IPO would make OpenAI the most valuable company in the world at the time of this writing. That is achievable if revenue grows to $100+ billion with strong operating margins, but it is a high bar.

The table illustrates the compounding challenge. At the Bloomberg low-end revenue estimate, investors paying 35x revenue at pre-IPO are betting on sustained revenue growth that takes OpenAI to $80-100 billion before the IPO multiple compresses — and further to sustain a 35x multiple at IPO, which would be unprecedented for any company at that revenue scale.

The more defensible bull case runs through the $60-70 billion revenue estimate and assumes that OpenAI's revenue grows to $120+ billion over the next two to three years at 40-50% annual growth. At that point, a 15-20x revenue multiple at IPO — still aggressive by traditional SaaS standards but plausible for a platform company with consumer and enterprise dominance — would produce the $2-2.5 trillion market cap that makes the pre-IPO price look reasonable in hindsight. The bear case is that revenue growth decelerates before the IPO window opens, public market investors price it at 10-12x revenue on the cautious side, and the IPO price implies a flat-to-down return for pre-IPO investors at $1.4 trillion.

The variable that neither bull nor bear case fully captures is the composition of OpenAI's revenue. Consumer subscription revenue (relatively high margin once the infrastructure is built) is more valuable per dollar than commodity API revenue (very low margin, highly competitive). If OpenAI can grow its higher-margin consumer and enterprise tiers faster than its commodity API business, the margin profile improves and the justifiable valuation multiple expands. The GPT-6 Sol and Luna pricing architecture — with Sol at $2 per million tokens capturing enterprise use cases and Luna at $0.10 per million capturing high-volume commodity workloads — reflects a deliberate strategy to segment the revenue base and defend the high-margin tier.

The Competitive Context: How OpenAI's Valuation Compares to the AI Market

OpenAI's $1.4 trillion target sits in a competitive landscape where the other frontier AI companies are valued dramatically differently. Anthropic, whose most recent equity round implied a valuation in the range of $60-70 billion, has taken a radically different capital structure path — raising enormous amounts of debt against chip lease payments rather than equity at high valuations. The Project North Star $60 billion debt financing, on top of the $35 billion Apollo deal from June, means Anthropic has secured nearly $95 billion in chip-lease-backed debt while keeping its equity valuation modest by comparison.

The contrast between OpenAI and Anthropic's capital structures reflects fundamentally different theories of how to maximize ownership while securing the capital needed to stay at the frontier. OpenAI is raising equity at very high valuations, which is dilutive to existing shareholders but preserves the nonprofit structure until the conversion and gives the company cash with no repayment obligation. Anthropic is preserving equity by raising debt secured by chip lease payments, which is non-dilutive but creates annual interest obligations that must be covered by revenue.

Meta is the most instructive public market comparison. Meta trades at approximately $1.5 trillion in market capitalization as of October 2026, but it generates $50+ billion in annual operating income on $150+ billion in annual revenue, runs at roughly 33% operating margins, and has a long track record of generating free cash flow. OpenAI is asking for a comparable valuation without any of those profitability metrics yet established. The premium over Meta's operating-income multiples that OpenAI commands reflects the belief that OpenAI is a platform company whose network effects and model improvements will eventually produce Meta-like margins — or better.

Microsoft's position deserves a separate analysis. Microsoft invested approximately $13 billion in OpenAI and holds a 49% economic stake in the for-profit subsidiary's profits (subject to the capped-profit structure). At $1.4 trillion, Microsoft's stake would be worth approximately $690 billion on paper — more than Microsoft paid for LinkedIn ($26 billion) and Activision ($69 billion) combined, and a multiple of what it actually invested. Microsoft's incentive to see OpenAI succeed is not neutral; it is profound.

Enterprise Buyer Implications: What the Fundraise Signals

For enterprise AI teams that have built production workflows on OpenAI's models, the $30 billion round at $1.4 trillion carries four practical signals worth acting on.

1. Capital stability through 2027. A company with $30 billion in fresh equity capital and significant existing cash reserves is not at risk of operational disruption for the foreseeable future. The scenario where enterprise teams need to migrate workflows because OpenAI runs out of money is effectively off the table for the duration of the IPO window. This is meaningful for risk assessments on critical applications.

2. Competitive pricing pressure will continue. A well-capitalized OpenAI has every incentive to continue cutting API prices to prevent enterprise workloads from migrating to Anthropic Claude, Google Gemini, or open-weight models. The $1.4 trillion valuation requires revenue growth, and the fastest path to revenue growth is capturing more of the enterprise API market — which means aggressive pricing. Enterprise teams should expect continued price reductions or capability improvements at existing price points.

3. The IPO narrative will favor enterprise metrics. When OpenAI approaches its IPO, the metrics it will need to show public market investors are net revenue retention, enterprise contract sizes, and API consumption growth. This creates incentives to invest heavily in enterprise features, reliability, and support quality over the next 12-18 months. The pre-IPO period is historically when enterprise software companies are most motivated to win and retain enterprise accounts.

4. Build model flexibility anyway. Capital stability does not mean pricing stability forever. Once OpenAI is public, quarterly earnings pressure may push it to raise prices on high-margin enterprise segments that have demonstrated inelastic demand. The Project North Star analysis of Anthropic's capital structure identified this same dynamic — companies with large fixed capital obligations eventually must pass those costs on to customers. Enterprise teams should treat the current competitive pricing window as an opportunity to lock in multi-year contracts and simultaneously maintain active relationships with at least two alternative model providers.

Takeaway

Takeaway: OpenAI's $30 billion round at a $1.4 trillion valuation is simultaneously a record and a bet. It is a record because no private company has ever sought a valuation at this scale with this revenue base, and it reflects genuine investor conviction in OpenAI's platform trajectory. It is a bet because the valuation requires OpenAI to grow from $40-70 billion in annualized revenue to $100+ billion before the IPO window closes — to demonstrate operating leverage, to complete a complex structural conversion, and to defend its consumer and enterprise positions against well-funded competitors. The investors who are writing the $30 billion check — sovereign funds, asset managers, infrastructure investors — are not making a pure financial calculation; they are making a strategic decision that exposure to OpenAI is essential regardless of the multiple. For enterprise buyers, the signal is constructive: a well-capitalized OpenAI with an IPO timeline to manage is a highly motivated enterprise partner for the next 12-18 months, pricing will remain competitive, and the product investment will be substantial. Build on that window deliberately — but build with the flexibility to adapt when the IPO changes the incentive structure.