OpenAI vs Anthropic Revenue Math Differs, Muddying Valuations

Bloomberg reports that OpenAI and Anthropic calculate revenue in different ways, which makes side-by-side comparisons unreliable for investors. The gap matters for valuations and deal-making at two of the largest private AI companies.

Key points

• Bloomberg says differing revenue measures keep investors from reliably comparing OpenAI and Anthropic.
• The confusion complicates valuations and deal-making, per the same report.
• A separate Bloomberg item describes a $70 billion annualized run rate tied to enterprise growth and infrastructure strain. Our earlier coverage noted this figure conflicts with another report.
• Bloomberg also reports the departure of OpenAI’s top Asia-Pacific policy executive after six months.

Why do revenue definitions matter?

Revenue is the anchor for most private-market valuations of AI companies, usually expressed as a multiple of annualized or run-rate figures. If two companies define the number differently, the multiples attached to them are not comparable, even when both are labeled “revenue.”

The Bloomberg summary does not, in the material available to us, spell out every difference in method. We therefore cannot say which company’s approach is more conservative, or how large the gap is. Readers should treat any headline comparison of the two firms’ revenue as provisional until the definitions are disclosed.

This also bears on the $70 billion run-rate figure. A run rate extrapolates a recent period to a full year, and its meaning depends on what is counted and over what window. The Bloomberg video summary links the figure to enterprise growth and says it highlights infrastructure strain that could affect customers, market sentiment and public offering plans. We have no further detail on the underlying calculation from these sources.

What does it mean for an IPO?

Public-market investors would eventually get audited financials under standard accounting rules, which would settle many definitional questions. Until then, private rounds, secondary trades and bank commentary rely on figures the companies or their backers choose to share. That leaves room for inconsistent framing.

Bloomberg says the inconsistency complicates deal-making. For OpenAI, which has a public offering as a stated area of interest for observers of the company, a clear and defensible revenue definition would likely be a prerequisite for pricing. That is our inference, not a claim from the sources.

What about the Asia-Pacific policy exit?

Bloomberg reports that OpenAI’s top public policy executive for Asia-Pacific is leaving after six months in the role. The report frames the exit as raising concerns about how well the company can handle increased regulatory scrutiny in the region. The sources do not give a reason for the departure, name a successor, or say whether the move was planned. Short tenures in policy roles can reflect many things, so the signal is limited until more is known.

What to watch

• Whether OpenAI or Anthropic publish or clarify how they define revenue and run rate.
• Any reconciliation of the $70 billion figure with other reported numbers.
• Announcement of a replacement for the Asia-Pacific policy role.

This site is independent and not affiliated with OpenAI. Nothing here is investment advice.