OpenAI and Anthropic's ARR row
- 01Theme 1: The AI Foundation Model IPO Race Is Distorting Key Metrics
- 02Theme 2: Anthropic Is Winning the Enterprise Battle Against OpenAI
- 03Theme 3: PE Deal Activity Is Resilient, But Fundraising Remains Deeply Challenged
- 04Theme 4: Physical AI and Radical Simplification as the Answer to "China Speed" in EVs
- 05Theme 5: AI Adoption Barriers Are Behavioral, Not Infrastructural
1. Key Themes
Theme 1: The AI Foundation Model IPO Race Is Distorting Key Metrics
Both OpenAI and Anthropic are preparing for what could be among the largest IPOs in history, but neither company is currently reporting revenue on a basis that would withstand public-market scrutiny.
"Essentially, both companies are in an ARR accounting arms race ahead of their IPOs, and neither is reporting on a basis that would survive a Big Four audit." — Harrison Rolfes, PitchBook Senior Research Analyst
The competitive anxiety is real: Anthropic has grown from ~$1B ARR a year ago to a reported $8B, while OpenAI has gone on offense with an internal memo accusing Anthropic of inflating numbers through revenue-share arrangements with Amazon and Google.
Theme 2: Anthropic Is Winning the Enterprise Battle Against OpenAI
Despite OpenAI's longer runway and broader consumer mindshare, Anthropic is now widely perceived as having built the more durable enterprise business — and OpenAI's own internal communications confirm this is a serious threat.
"OpenAI's enterprise sales motion has an Anthropic problem, and the Dresser memo is the internal acknowledgment of it, dressed up as competitive intelligence." — Harrison Rolfes, PitchBook Senior Research Analyst
Theme 3: PE Deal Activity Is Resilient, But Fundraising Remains Deeply Challenged
Despite macro headwinds — tariff uncertainty, private credit concerns, Iran war, and Goldman Sachs raising recession odds to 45% — PE dealmaking held firm in Q1 2026. However, the fundraising environment continues to deteriorate.
"Eighty-four PE funds reached a final close in the first quarter, collecting a total of $54.2 billion. This puts 2026 behind pace compared with last year, when 406 funds raised $286 billion, the lowest dollar amount since pandemic-disrupted 2020."
The bifurcation is stark: established names like Josh Harris' 26North closed a $6B debut fund while newer, lesser-branded firms struggle to raise at all.
"The fundraising fortunes of Josh Harris' 26North stand in stark contrast to those of other newer firms."
Theme 4: Physical AI and Radical Simplification as the Answer to "China Speed" in EVs
Slate Auto's $650M raise — backed by Jeff Bezos — is betting that stripping a product down to its essentials can compete on price with Chinese EV manufacturers. The thesis: physical AI (robotics, automation) may be the only sustainable long-term counter.
"Jeff Bezos-backed Slate Auto is betting radical product simplification can compete with Chinese EV makers on price, raising $650 million ahead of its 2026 production launch. Our analyst note explores why physical AI may be the only long-term answer to 'China Speed.'"
Theme 5: AI Adoption Barriers Are Behavioral, Not Infrastructural
A Gallup survey of 23,000 employees reveals that the primary obstacle to AI adoption isn't access to tools — it's skepticism about usefulness and moral objections.
"AI holdouts have declared their top reason for avoiding the tech: They don't see the point. A new Gallup survey of 23,000 employees finds the biggest barriers to adoption are skepticism about usefulness and moral objections, not a lack of tools."
This has major implications for enterprise AI vendors: the go-to-market problem is change management and demonstrated ROI, not product availability.
2. Contrarian Perspectives
OpenAI's Revenue Lead May Be More Fragile Than It Appears
The consensus view is that OpenAI, with its dominant consumer brand and ChatGPT ubiquity, is the clear revenue leader. But the article suggests OpenAI's enterprise trajectory is structurally weaker than the headline $6B ARR figure implied a year ago — and its own CRO's memo is the tell.
"OpenAI's enterprise sales motion has an Anthropic problem, and the Dresser memo is the internal acknowledgment of it, dressed up as competitive intelligence." — Harrison Rolfes
The fact that OpenAI felt compelled to attack Anthropic's accounting internally signals defensive posturing, not market dominance.
Neither AI Giant's Financials Are IPO-Ready Despite Massive Valuations
Markets and media are treating the impending OpenAI and Anthropic IPOs as near-certainties at enormous valuations. But the article's core assertion is that the foundational revenue metric — ARR — is unreliable for both companies.
"Neither is reporting on a basis that would survive a Big Four audit." — Harrison Rolfes
Investors pricing these companies on current ARR may be underwriting accounting conventions that will look very different once public-company disclosure standards apply.
"Idling" Is a Skill Worth Investing In, Not Eliminating
Against the dominant narrative that AI will maximize human productivity by eliminating idle time, Anthropic's own co-founder argues the opposite: unstructured thinking time becomes more valuable, not less, in an AI-saturated environment.
"'Idling' is the human skill that will be worth 10x more in the age of AI." — Jack Clark, Anthropic co-founder, at Semafor's World Economy Summit
3. Companies Identified
OpenAI
- Description: Leading AI foundation model company, ChatGPT maker
- Why mentioned: Central to the ARR accounting dispute; its CRO issued an internal memo attacking Anthropic's revenue reporting ahead of its IPO
- Quote: "OpenAI chief revenue officer Denise Dresser writes that Anthropic's revenue-share arrangements with Amazon and Google...are allegedly being accounted for in ways that artificially boost the company's headline numbers."
Anthropic
- Description: AI safety-focused foundation model company, maker of Claude
- Why mentioned: Reported $8B ARR (up from $1B a year ago); identified as having built a "bigger and more durable enterprise business" than OpenAI
- Quote: "Not only has Anthropic reached a reported $8 billion in ARR most recently, but it's widely seen as having built a bigger and more durable enterprise business than its rival."
Slate Auto
- Description: EV startup backed by Jeff Bezos
- Why mentioned: Raised $650M ahead of 2026 production launch; positioned as a case study in product simplification as competitive strategy against Chinese EV makers
- Quote: "Slate Auto is betting radical product simplification can compete with Chinese EV makers on price."
26North
- Description: PE firm founded by Josh Harris, co-founder of Apollo
- Why mentioned: Closed a $6B debut PE fund, a standout fundraising success in an otherwise difficult capital formation environment
- Quote: "The fundraising fortunes of Josh Harris' 26North stand in stark contrast to those of other newer firms."
Meta
- Description: Social media and AI technology conglomerate
- Why mentioned: Mark Zuckerberg is building a 3D AI avatar of himself to interact with employees, part of a "personal superintelligence" push
- Quote: "Mark Zuckerberg is building an AI avatar of himself to interact with Meta employees...amid Meta's push to develop 'personal superintelligence.'"
Neomorph
- Description: Biotech startup focused on molecular glue degraders
- Why mentioned: Raised $100M Series B led by Deerfield Management — notable deal in biotech VC
- Quote: "Neomorph, a biotech startup focused on developing molecular glue degraders, raised a $100 million Series B led by Deerfield Management."
Kailera Therapeutics
- Description: Biotech startup developing GLP-1 drugs
- Why mentioned: Filed for IPO — notable given the continued investor interest in GLP-1/obesity drug space
- Quote: "Janus Henderson-backed Kailera Therapeutics, a biotech startup developing GLP-1 drugs, filed for an IPO."
PolyPeptide
- Description: Swiss contract drugmaker focused on peptides and oligonucleotides
- Why mentioned: EQT, KKR, and Advent International are among suitors — signals continued large-cap PE appetite for pharma CDMO assets
- Quote: "EQT, KKR and Advent International are among suitors for a potential acquisition of PolyPeptide."
Ping An
- Description: Chinese insurance group
- Why mentioned: Tapping Campbell Lutyens to manage a $1B secondary sale of software-focused PE stakes — significant secondary market signal
- Quote: "Chinese insurance group Ping An has tapped Campbell Lutyens to manage a secondary sale of $1 billion in software-focused private equity stakes."
4. People Identified
Harrison Rolfes
- Description: Senior Research Analyst at PitchBook covering OpenAI and Anthropic
- Why mentioned: Provided the sharpest analytical commentary in the issue, calling out both companies' pre-IPO accounting games and diagnosing OpenAI's enterprise vulnerability
- Quote: "Essentially, both companies are in an ARR accounting arms race ahead of their IPOs, and neither is reporting on a basis that would survive a Big Four audit."
Denise Dresser
- Description: Chief Revenue Officer at OpenAI
- Why mentioned: Authored the internal memo obtained by The Verge accusing Anthropic of inflating ARR through cloud partner revenue-share arrangements
- Quote: "OpenAI chief revenue officer Denise Dresser writes that Anthropic's revenue-share arrangements with Amazon and Google...are allegedly being accounted for in ways that artificially boost the company's headline numbers."
Jack Clark
- Description: Co-founder of Anthropic
- Why mentioned: Made a provocative claim at Semafor's World Economy Summit that "idling" — unstructured human thinking — will be worth 10x more in the AI era
- Quote: "'Idling' is the human skill that will be worth 10x more in the age of AI."
Josh Harris
- Description: Co-founder of Apollo Global Management; founder of 26North
- Why mentioned: Closed a $6B debut PE fund for 26North, bucking the broader PE fundraising slump and demonstrating the continued LP premium placed on brand-name founders
- Quote: "Apollo's co-founder closed a $6 billion debut PE fund. The fundraising fortunes of Josh Harris' 26North stand in stark contrast to those of other newer firms."
Jon Winkelried
- Description: CEO of TPG
- Why mentioned: Publicly pushed back on broad private credit fears, arguing existential risks are limited to the software sector specifically
- Quote: "TPG CEO Jon Winkelried says the existential risks to private credit are limited to the software sector."
Nir Zuk
- Description: Founder of Palo Alto Networks
- Why mentioned: Seeking regulatory approval to acquire a majority stake in California lender Liberty Bank — notable as a tech founder moving into financial services
- Quote: "Palo Alto Networks founder Nir Zuk is seeking regulatory approval to buy a majority stake in California lender Liberty Bank."
Mark Zuckerberg
- Description: CEO of Meta
- Why mentioned: Building a 3D AI avatar of himself to interact with employees as part of Meta's "personal superintelligence" strategy
- Quote: "Mark Zuckerberg is building an AI avatar of himself to interact with Meta employees."
5. Operating Insights
Pre-IPO Metrics Need Audit-Ready Rigor Before Investors Rely on Them
For founders and CFOs preparing for public markets, the OpenAI/Anthropic episode is a cautionary tale. Strategic cloud distribution partnerships (like those with Amazon and Google) can inflate ARR in ways that won't survive GAAP/Big Four scrutiny. Build your reporting architecture to public-company standards before you need it.
"Neither is reporting on a basis that would survive a Big Four audit." — Harrison Rolfes
Enterprise AI Sales Win Requires Proving Durable Business Value, Not Just Product Access
The Gallup finding that AI non-adopters cite "not seeing the point" rather than lack of tools means enterprise AI vendors must prioritize use-case ROI demonstration and change management over feature delivery. Anthropic's enterprise success over OpenAI further underscores that trust, reliability, and tailored business solutions matter more than brand recognition.
"The biggest barriers to adoption are skepticism about usefulness and moral objections, not a lack of tools."
Fundraising Preparation Must Be Continuous, Not Episodic
In a market where PE fundraising has fallen to near-2020 lows, the difference between firms that close and those that don't increasingly comes down to operational readiness — clean cap tables, consistent investor updates, and diligence materials available on demand, not assembled last-minute.
"Most founders don't struggle with fundraising because of vision or ambition. They struggle because the mechanics are unclear, fragmented, and learned too late." — Fidelity Private Shares (sponsored)
6. Overlooked Insights
The Secondary Market Is Beginning to Reflect Stress in Software-Focused PE Portfolios
Ping An's $1B secondary sale of software-focused PE stakes — in a period when TPG's CEO is specifically flagging software as the sector with existential private credit risk — may signal broader LP-level rebalancing away from software PE exposure. This is a quiet but potentially meaningful indicator of where secondary market pricing pressure is building.
"Chinese insurance group Ping An has tapped Campbell Lutyens to manage a secondary sale of $1 billion in software-focused private equity stakes." "TPG CEO Jon Winkelried says the existential risks to private credit are limited to the software sector."
European AI Startup Geography Mirrors Conventional VC Ecosystems
Buried in the Chart of the Day, the State of European AI data shows that AI startup concentration in Europe follows existing VC infrastructure (UK, Germany, France lead) rather than emerging in new geographies. This suggests European AI investment opportunity may be more concentrated — and more competitive in established markets — than the "rising European AI scene" narrative implies.
"In absolute terms, the number of AI startups by country follows the pattern of conventional venture ecosystems, with the UK, Germany, and France holding top positions."