OpenAI Files Confidential IPO With Goldman Sachs, Targets September 2026 at $850B — Same Week as Anthropic

OpenAI has filed a confidential IPO application with Goldman Sachs and Morgan Stanley as lead underwriters, targeting a September 2026 public listing. Private market buyers have priced the company at $730 billion to $850 billion. Analysts covering the deal expect the listing to push past $1 trillion in market capitalization on opening day, which would make it one of the five largest IPOs in US history.

The filing came 13 days after Anthropic submitted its own confidential S-1, also targeting a 2026 listing. The two filings together set up Q3 and Q4 2026 as the first genuine public market test of frontier AI valuations — the moment when VC-priced AI companies meet institutional fund managers who require revenue multiples that make sense.

What OpenAI's IPO Filing Actually Contains

OpenAI's S-1 is confidential as of June 14, 2026, which means the financial details are not public. What is known comes from filings, CFO statements, and leak reporting:

Revenue: CFO Sarah Friar confirmed in a January 2026 blog post that the company's annualized revenue rate exited 2025 above $20 billion. Full-year 2025 actual revenue was approximately $13.1 billion. The annualized rate of $20B+ implies Q4 2025 and Q1 2026 acceleration — consistent with ChatGPT Enterprise adoption growing and the Codex agent product reaching commercial traction.

Last funding round: OpenAI closed a $122 billion round in late 2025 / early 2026 at an $852 billion post-money valuation. Goldman Sachs and Morgan Stanley were already involved in structuring that round. The IPO continues that banking relationship.

Structure: OpenAI completed its conversion from a hybrid nonprofit/for-profit structure to a Delaware public benefit corporation in early 2026. That conversion was a prerequisite for a public listing — you can't take a nonprofit public. The restructuring was contentious (Elon Musk's lawsuit was one of the complications) but completed before the IPO filing.

Microsoft's position: Microsoft owns approximately 49% of the for-profit OpenAI LLC entity that is being taken public, in exchange for its multi-year $13 billion investment commitment. How Microsoft's stake converts to public shares — and at what dilution — is one of the key unknowns in the S-1.

The $850 Billion Valuation: How It Gets to $1 Trillion

Private market pricing: $730B-$850B. Public market analyst consensus: $1T+ at listing. The gap is explained by two factors.

First, public markets apply a liquidity premium. When shares become freely tradeable on Nasdaq, the pool of potential buyers expands from a few hundred institutional VCs to millions of retail and institutional investors. Supply and demand: same company, more buyers, higher price.

Second, AI is the dominant investment theme of 2025-2026. OpenAI's listing will attract investors who cannot access private AI deals — pension funds, retail investors, ETF managers forced to buy by index inclusion. That demand pressure has historically pushed AI and tech IPOs above private market pricing in the first weeks of trading. Nvidia's 2023-2025 run is the model everyone is using.

The revenue multiple at $1T: $1 trillion / $20B annualized revenue = 50x revenue. That is high by traditional standards but within the range of what Salesforce, Snowflake, and early Microsoft Azure commanded. Whether it holds depends on how fast $20B grows and whether operating margins improve. OpenAI has been cash-flow negative — the S-1 will reveal by how much.

Revenue and Unit Economics: What the S-1 Will Show

OpenAI's revenue comes from three streams, each with different margin profiles:

ChatGPT subscriptions: $20/month consumer, $25/month Plus, $30/month Team, $60/month Enterprise Pro. Enterprise contracts are bespoke. This is the highest-margin revenue stream — software subscription with minimal marginal cost per user once model inference is covered.

API revenue: Charged per token, priced by model tier. GPT-5.5 is the premium model; GPT-5.5 Instant and the Turbo variants are the volume tier. API revenue scales with developer usage and is margin-sensitive to compute costs. As Nvidia GPU prices and cloud compute rates evolve, API margin either expands or compresses.

Enterprise deployment contracts: Multi-year, multi-seat contracts with large enterprises for ChatGPT Enterprise, Codex deployment, and the new Deployment Company structure (the partnership with TPG, McKinsey, and Goldman for enterprise AI deployment). These contracts are high-ACV, moderate-margin, and provide revenue visibility.

The S-1 will reveal the revenue split and gross margin per segment. The number that institutional investors will focus on: compute COGS as a percentage of API revenue. If OpenAI is spending $0.60 of compute for every $1 of API revenue, the margin profile is very different from spending $0.20.

Anthropic Filed First — The Dual-IPO Dynamic

Anthropic submitted its confidential S-1 on June 1, 2026. OpenAI filed approximately 13 days later. The back-to-back filings create a competitive dynamic in public markets that has no direct precedent in the AI era.

Both companies are targeting overlapping investor pools. Both are being priced on AI fundamentals that retail investors have been tracking for three years. The company that lists first gets the cleaner narrative — it defines the comparable for the second listing. The company that lists second gets the benefit of watching the first listing's reception and adjusting pricing accordingly.

The strategic calculus: Anthropic at $380B valuation is the "safer" AI bet — revenue-generating, Claude Code is a breakout product, government contracts through Glasswing. OpenAI at $850B is the brand premium bet — ChatGPT is the consumer AI brand, Codex is the developer brand, GPT-5.5 leads benchmarks. Two different investment theses going public in the same window.

For developers, the dual-IPO dynamic matters because both companies will face quarterly earnings pressure immediately after listing. That pressure historically pulls companies toward high-margin, high-revenue enterprise deals and away from low-margin, community-beneficial projects. Expect API pricing pressure to increase, free tier limits to tighten, and open-weight or open-API commitments to receive less internal prioritization.

We covered Anthropic's IPO filing in Anthropic's confidential IPO: $380B valuation, $47B revenue target, what developers need to know.

What the IPO Means for API Pricing and Model Access

The most direct developer impact of OpenAI going public is the constraint it places on pricing decisions.

A private company can decide to cut API prices by 50% to win market share, accepting short-term revenue decline in exchange for developer adoption. OpenAI has done this multiple times — the GPT-4 Turbo price cuts in 2024, the GPT-4o pricing at GPT-3.5 price points, the GPT-5.3 Garlic pricing to undercut Claude Sonnet.

A public company faces a quarterly earnings call where those pricing cuts show up as revenue misses. Wall Street analysts will model a price-per-token benchmark and expect it to hold or grow. Pressure to maintain pricing will increase. The cheapest model tier (GPT-5.5 Instant) is likely protected — it's the volume driver that justifies developer adoption numbers. The premium API tier (GPT-5.5 Pro) and enterprise contracts will face upward pressure.

For developers currently on OpenAI API contracts: price predictability should improve because public company revenue commitments require stable contract terms. But the era of aggressive price-cutting to win developer share will compress.

Microsoft's Stake and the Ownership Structure

Microsoft's position in OpenAI is approximately 49% of the for-profit entity, acquired through a multi-year commitment of $13 billion in Azure credits, cash, and infrastructure. That stake is now worth approximately $420 billion at the $850B IPO valuation — a return of roughly 32x on the cash invested.

The conversion from OpenAI LLC to a Delaware public benefit corporation changes what Microsoft owns. The precise share structure, conversion ratio, and lockup periods will be in the S-1. What's known: Microsoft negotiated a perpetual IP license to OpenAI's model technology as part of its investment structure. Even if Microsoft's equity stake dilutes through IPO and secondary offerings, the IP license remains — Azure AI customers will continue to have access to OpenAI models through Azure OpenAI Service regardless of equity structure.

For developers on Azure: the IPO changes nothing about Azure OpenAI Service access in the short term. The commercial relationship between Microsoft and OpenAI is contractual, not equity-dependent.

Sam Altman's stake has been a subject of reporting. After the for-profit conversion, Altman negotiated approximately 7% equity — a grant delayed from the original cap structure where he held no equity. At $850B, 7% = $59.5 billion. The S-1 will confirm the exact figure.

Our Analysis: Public Markets Will Price Risk Differently Than VCs

Venture capital investors in OpenAI's last round paid $852 billion for a company burning cash on compute, with no guarantee of profitability timelines, in a sector where the product is replicable by competitors with sufficient resources. They accepted that risk because the upside of owning a piece of the defining AI platform justified the uncertainty.

Public market institutional investors operate differently. Pension funds, index funds, and mutual funds buy companies they can model with earnings multiples, not just revenue. The first quarterly earnings call after OpenAI lists will be the defining moment: if the company shows a credible path to positive operating income within 6-8 quarters, the stock holds. If compute costs are growing faster than revenue, the valuation compresses fast.

The OpenAI IPO is also a referendum on whether AI-native companies can sustain the kind of growth trajectory that justifies frontier valuations. Nvidia sustained it because it had hardware margins, not software margins. OpenAI's margins depend on how the compute cost curve evolves — if inference gets cheaper faster than prices fall, margins expand. If GPT pricing wars compress revenue while compute stays expensive, they don't.

For developers, the most important sentence in the S-1 will be the one that describes gross margin by product segment. That number determines whether OpenAI can afford to keep developer-friendly pricing or needs to extract margin from the API tier. Watch for it.

Key Takeaways

• OpenAI filed a confidential IPO with Goldman Sachs and Morgan Stanley, targeting September 2026 at $730B-$850B private market pricing; public market consensus expects $1T+ at listing
• $20B+ annualized revenue confirmed by CFO Sarah Friar in January 2026; full-year 2025 was approximately $13.1B — the acceleration implies ChatGPT Enterprise and Codex are both growing fast
• $122 billion last raised at $852B valuation (2025/early 2026) — the IPO pricing continues that valuation arc
• Filed 13 days after Anthropic's S-1: dual-IPO dynamic creates competition for the same investor pool; first-to-list company defines the comparable for the second
• Microsoft owns ~49% of the for-profit entity being listed, worth approximately $420B at IPO pricing — IP license to Azure AI Service is separate from equity and survives dilution
• Sam Altman's stake: approximately 7% equity post-conversion, confirmed in reporting; worth ~$59.5B at $850B valuation
• Developer impact: quarterly earnings pressure will reduce aggressive API price-cutting; free tier limits may tighten; enterprise contracts will face upward pricing pressure post-IPO
• Watch the S-1's gross margin by segment: compute COGS as % of API revenue is the number that determines whether OpenAI can sustain developer-friendly pricing at scale

Sources

• TechTimes — OpenAI Targets IPO as Soon As September at $850B
• AI Weekly — OpenAI Files Confidential IPO Targeting $850B Valuation
• Yahoo Finance — OpenAI Files for IPO After Raising $122 Billion
• Enterprise DNA — OpenAI Files Confidential IPO Papers, Eyes September
• Crypto Briefing — OpenAI files for IPO with potential $1 trillion valuation