bnnbloomberg.ca via Reddit

OpenAI Disputes Anthropic Revenue by Billions

6 sources tracking this story

Key insights

  • Anthropic's Goldman-Morgan Stanley-JPMorgan underwriter team targets an October 2026 IPO, potentially beating OpenAI's September listing to set sector accounting standards.
  • At OpenAI's preferred net figure of $22 billion, Anthropic's 15-month 22x growth run still dwarfs OpenAI's 25% annualized expansion.
  • One OpenAI investor requires a $1.2 trillion IPO valuation to justify the $852 billion private round, a threshold the accounting dispute directly threatens.

Why this matters

OpenAI and Anthropic filed confidential S-1s seven days apart in June 2026, forcing institutional investors to evaluate conflicting revenue figures at the same time. OpenAI CRO Denise Dresser's April memo, confirmed by The Next Web to include a warning against becoming a 'single-product company in a platform war', positioned the gross-versus-net accounting dispute as a deliberate competitive weapon before either S-1 was formally submitted. Whichever company reaches the public market first sets the accounting template that will govern how investors value every subsequent AI IPO, a structural advantage DA Davidson analyst Gil Luria called 'defining how a frontier AI lab reports its financials in public markets.' At least one OpenAI investor already requires a $1.2 trillion public valuation assumption to justify the current $852 billion private round, making the revenue comparison a live stress test on investor confidence in both companies.

Summary

OpenAI claims Anthropic overstates its revenue by billions of dollars, because Anthropic books gross customer revenue while OpenAI reports net after payments to partners like Microsoft. Both companies filed for IPO within days of each other: Anthropic on June 1, OpenAI on June 8, 2026. OpenAI is targeting a valuation of approximately $1 trillion, putting enormous pressure on how both companies present their financial metrics to investors at the same moment. Essentially: (OpenAI, Anthropic) are fighting over the definition of revenue at the exact moment investor scrutiny is at its peak. - Anthropic books gross customer revenue; OpenAI reports net after partner cuts, making the two figures structurally incomparable - Altman called Anthropic's Super Bowl advertisements 'deceptive' in February 2026, one front in a rivalry that now spans product launches, public statements, and accounting methodology - Both CEOs refused to join hands at an India summit, a moment captured in viral video The accounting dispute could determine which company appears larger to investors before two of the most consequential AI IPOs in history.

Potential risks and opportunities

Risks

  • Anthropic's IPO valuation could be materially reduced if the SEC requires revenue restatement on a net basis before the offering clears
  • OpenAI faces reputational risk if its public attacks on a competitor's accounting are perceived as a coordinated pressure campaign to undercut a rival's IPO
  • Investment banks managing both IPOs simultaneously face conflict-of-interest scrutiny from regulators if information barriers between their deal teams prove insufficient

Opportunities

  • Accounting firms specializing in tech-platform revenue recognition gain leverage advising both companies and future AI startups on gross-vs-net treatment ahead of and after IPO
  • Institutional investors who can accurately model net revenue figures for both companies hold a meaningful analytical edge in the IPO allocation process
  • Rival AI companies can preemptively clarify their own revenue recognition methodology, differentiating themselves to institutional investors watching this dispute play out

What we don't know yet

  • Which specific accounting standard each company applies and whether the SEC has formally weighed in on the dispute before the IPO filings clear review
  • The exact revenue figures Anthropic reports versus what OpenAI claims the net equivalent would be under its own methodology
  • Whether other AI companies with cloud-distribution deals face the same gross-vs-net reclassification risk as institutional investor scrutiny intensifies around the sector

What others are reporting

Coverage cluster as of 8h after publish

  1. The Next Web Read →

    Reports one investor requires a $1.2T public valuation to justify OpenAI's private round, and includes Dresser's warning against becoming a 'single-product company in a platform war.'

    You do not want to be a single-product company in a platform war
  2. TechTimes Read →

    Covers Anthropic's underwriter selection and DA Davidson's argument that the first AI company to IPO defines valuation reporting standards for the entire sector.

    The company that files first gets to define how a frontier AI lab reports its financials in public markets.
  3. Let's Data Science Read →

    Primary source analysis of the Dresser memo: adds the Amazon distribution pivot signal, a new model called Spud, and Dresser's direct language on Anthropic's 'fear and restriction' framing.

    Their story is built on fear, restriction, and the idea that a small group of elites should control AI.
  4. AI Afterhours Read →

    Introduces revenue per gigawatt of compute as an accounting-neutral metric: Anthropic at $21.4B/GW versus OpenAI's $12.6B/GW, a 70% efficiency gap independent of gross-versus-net treatment.

    Revenue tells you how much money is coming in but nothing about how efficiently it's being generated.
  5. remio.ai Read →

    Argues growth trajectory matters more than the accounting dispute: Anthropic's 30x expansion in 15 months versus OpenAI's 25% annualized growth reflects a structurally different business model.

    Even using OpenAI's preferred figure of $22 billion for Anthropic, Anthropic got to $22 billion from $1 billion in fifteen months.