The Civic Node

OpenAI

9 min read

Overview

OpenAI is the San Francisco-based AI company behind ChatGPT and the GPT model series. It is the market leader in consumer AI with a valuation of $157B as of early 2025 (with a reported $340B round pending). Founded as a nonprofit and restructured as a for-profit, it maintains a Microsoft partnership providing deep enterprise integration.

Key Facts

  • Creator of ChatGPT and GPT-4/GPT-5 series; dominant consumer AI brand; 82.65% of consumer generative AI market as of July 2025 (HHI: 6,916)
  • Valuation: $300B (as of August 2025); $500B reported as of October 2025
  • 2024 actual revenue: ~$3.6B (Zitron analysis); ARR reported as $12–13B (annualized, not actual) as of mid-2025
  • Burns $5B+/year; not profitable
  • Microsoft exclusivity clause for Azure compute was used to artificially inflate ChatGPT prices 100–200x versus competitors; resolved June 2025 when OpenAI gained Google compute access; prices immediately dropped 80%
  • Class action antitrust lawsuit filed October 2025 alleging Microsoft-OpenAI deal restricted compute supply
  • Aligned with Trump administration from day one: Stargate joint venture (Jan 21, 2025); White House relationship maintained through 2025
  • EU AI Act classifies GPT-4-level models as “high-impact GPAI” requiring systemic risk evaluation
  • Former Scale AI client for data labeling (ended due to competitive conflicts after Meta’s Scale investment)
  • DRAM crisis (Oct 2025): Signed simultaneous letters of intent (LOIs, non-binding) with Samsung and SK Hynix for ~900,000 DRAM wafers/month — roughly 40% of global supply (~2.2-2.3M total). Neither supplier knew about the other’s deal. LOI valued at $71.3B over four years (analyst estimate). Triggered a global memory panic: DRAM contract prices rose 171% YoY by Q3 2025; 64GB DDR5 went $180→$710 (294%); quote windows compressed to 7 days. Apple, Amazon, Microsoft, and AMD flew executives to South Korea and locked in 2-3 year binding deals at peak prices. Sam Altman’s Dirty DRAM Deal How Sam Altman’s OpenAI may have caused the worst consumer hardware crisis The Letter That Moved a Market — Medium The Dirty DRAM Deal How OpenAI Just Locked Up 40 Percent of Global RAM — Bizety
  • Spending cut (Feb 2026): Cut compute spending target from $1.4 trillion to ~$600 billion through 2030 — a 57% reduction. 2025 revenue was $13.1 billion against $8 billion cash burn. OpenAI Massively Cuts Spending Plan as Reality Closes In — Futurism
  • Stargate cancelled (Mar 2026): Flagship data center expansion in Abilene, TX killed after Oracle financing negotiations collapsed and OpenAI’s demand forecasts shifted. The 1.2GW→2.0GW expansion was scrapped; Microsoft took over the site. Stargate Data Center Expansion Cancelled — Oracle and OpenAI
  • Florida criminal investigation (Apr 2026): Florida AG James Uthmeier escalated a civil probe into a criminal investigation over ChatGPT’s alleged role in the April 2025 FSU mass shooting. Claims ChatGPT advised the shooter on weapon type, ammunition, and tactics. Uthmeier: “If this were a person on the other side of the screen, we would be charging them with murder.” Subpoenas seek internal policies on user threats of harm and a full ChatGPT employee list — suggesting individual criminal liability theory. OpenAI responded that ChatGPT provided “factual responses” from publicly available information. Florida’s attorney general announces criminal investigation into OpenAI
  • TBPN acquisition (Apr 2026): Acquired podcast TBPN for a reported sum in the “low hundreds of millions” (FT); WSJ reported TBPN on track for $30M+ revenue. TBPN had 58K YouTube subscribers. Chris Lehane (OpenAI VP of Global Affairs) oversees it. Ben Thompson called OpenAI “the short bus at the end of the rainbow” in response. OpenAI Acquires TBPN — TechCrunch Big A — The Crisis Got Weirder (RAM Apocalypse Update)
  • Sora shuttered (Mar 2026): Discontinued the consumer video app and API; Sora research team redirected to “world simulation research” for robotics. Part of broader cost-cutting. ChatGPT image generator unaffected. OpenAI Pulls the Plug on Its Sora AI Video App OpenAI Shutters Sora App — CNBC
  • GPT 5.4 Cyber: OpenAI has a cyber-security-focused model (GPT 5.4 Cyber) described as highly capable — parallel to Anthropic’s gated Claude Mythos. Context for the Mythos breach investigation. Claude Mythos Unauthorised Access — BBC
  • Musk v. Altman verdict (May 18, 2026): Federal jury in U.S. District Court (Oakland) unanimously rejected Elon Musk’s $150B suit after less than 2 hours of deliberation. Jury found Musk filed after the relevant statutes of limitations expired (3-year deadline Aug 5, 2021 for breach of charitable trust; 2-year deadline Aug 5, 2022 for unjust enrichment). Judge Yvonne Gonzalez Rogers dismissed claims on the spot. Verdict removes one of the final roadblocks to OpenAI’s expected IPO this year, which could be one of the largest in history; valuation cited at $730B during trial. Musk’s antitrust claims (added November 2024) technically remain alive but judge has signaled they are “not very good claims.” Musk-OpenAI Jury Verdict — NYT - 2026-05-18 Musk-OpenAI Jury Verdict — Ars Technica - 2026-05-18
  • Sam Altman credibility exposure from trial: three weeks of testimony attacking Altman’s truthfulness; New Yorker’s pre-trial 16,000-word “Sam Altman May Control Our Future — Can He Be Trusted?” piece (Apr 13, 2026); Steven Molo’s closing argument that no one would walk a bridge “built on Sam Altman’s version of the truth.” Musk lost the case but the credibility frame is now on the public record going into the IPO road show.
  • Catherine Bracy / EyesOnOpenAI institutional vector: post-verdict statement urged CA AG Rob Bonta to revisit OpenAI’s restructuring agreement, order an independent valuation of the nonprofit’s assets, and compel their transfer to a truly independent charitable entity. The verdict closed the Musk vector but opens the AG-revisit vector.

Newsletter Relevance

OpenAI’s scale and Microsoft relationship create a structural advantage that smaller AI companies (Anthropic, Mistral) are racing to overcome. The Microsoft deal is simultaneously the source of its enterprise dominance and a target of antitrust scrutiny. OpenAI is also the primary subject of EU AI Act high-impact model provisions — the first company to face systemic risk assessment requirements.

Connections

  • Anthropic — primary competitor; differentiated on safety branding
  • Microsoft — investor and enterprise distribution partner; also sued over the relationship
  • Meta — competitor; open-source strategy contrasts with OpenAI’s closed approach
  • EU AI Act — GPT-4 cited as model requiring systemic risk evaluation
  • Platform Antitrust — Microsoft-OpenAI deal under antitrust scrutiny
  • Tech-State Conflict — EU regulation and US antitrust proceedings

Source Appearances

Open Questions

  • What is the status of the Microsoft antitrust lawsuit?
  • Does OpenAI comply with DoD requests that Anthropic refused? Partially answered: OpenAI signed a DoD contract claiming shared red lines (mass surveillance, autonomous weapons, social credit). But the key question from American Progress: why would the government honor OpenAI’s restrictions after trying to destroy Anthropic for similar ones?
  • How does OpenAI’s for-profit restructuring affect its regulatory relationships?
  • What does the $340B valuation round mean for AI market concentration?
  • Can OpenAI reach $280B revenue by 2030 to justify even the reduced $600B spending target?
  • What happens to the non-binding LOIs with Samsung/SK Hynix as OpenAI scales back? Are they quietly being renegotiated?
  • GPT-6: if it doesn’t represent a major capability leap, does OpenAI have a viable path to profitability?
  • What is the legal exposure from the Florida criminal investigation? If criminal charges are filed against employees, does that change hiring and deployment decisions across the industry?

Graph View

Backlinks