Anthropic's IPO Surge, OpenAI's Public Market Retreat.

Mazen Al Sabbagh

— 7 min read
Share

TL;DR

  • OpenAI IPO delay: CFO Sarah Friar is reportedly advocating to postpone OpenAI's public offering from 2026 to 2027, citing unmet revenue goals and reporting standards. This occurs as Anthropic drives toward an October listing at over $900 billion. (Gizmodo, Morningstar)
  • Altman extends invitation to Musk: Sam Altman posted an invitation to Elon Musk for the GPT-5.5 launch party, stating "the world needs more love," amidst their ongoing legal proceedings now entering Week 2. (Business Insider, Benzinga)
  • US AI/crypto skepticism: A POLITICO poll reveals widespread American distrust of AI and cryptocurrency, creating tension for candidates benefiting from record super PAC spending in the 2026 midterms. (POLITICO)
  • Hong Kong tech inflow: Morgan Stanley forecasts $1.25-1.75 billion in passive investment into Hong Kong tech stocks when Chinese AI firms Zhipu AI and MiniMax join the Hang Seng Tech Index on June 8. (CNBC)
  • Oscars ban AI: The Academy of Motion Picture Arts and Sciences has formally disqualified AI-generated acting and writing from Oscar eligibility, mandating human performance for all award-worthy contributions. (Reuters, TechCrunch)

Lead Story: The Public Market Inversion: OpenAI Stalls as Anthropic Accelerates

The weekend's most critical financial narrative is not an emergence but a reversal: OpenAI's public market trajectory is faltering just as Anthropic's accelerates. This dynamic shifts the competitive landscape and redefines investor expectations for AI enterprises.

Multiple reports indicate OpenAI CFO Sarah Friar has privately advised delaying the company's IPO from late 2026 to 2027. This recommendation stems from a confluence of factors: failure to reach 1 billion weekly active ChatGPT users by the close of 2025, revenue growth lagging projections, and the significant capital drain from a planned $1.4 trillion data center investment. PitchBook's analysis, cited by Crowdfund Insider, concluded a Q4 2026 listing is now "unattainable." Morningstar underscored the strategic implication: the first AI company to IPO will "define what good looks like" for public market participants, a benchmark OpenAI risks relinquishing to its rival.

In stark contrast, Anthropic is demonstrating aggressive market confidence. TechCrunch reported a 48-hour window for investor allocations in a $50 billion round at a valuation exceeding $900 billion, potentially its final private raise before an anticipated October 2026 IPO. Goldman Sachs and JPMorgan are already retained as underwriters. BetaNews notes that a successful close at this valuation would formally position Anthropic ahead of OpenAI's $852 billion, establishing it as the world's most valuable AI firm.

This divergence in market positioning is rooted in fundamental business models. Counterpoint Research's Q1 2026 data revealed Anthropic surpassing OpenAI in LLM revenue share (31.4% vs 29%), driven by substantially higher per-user monetization ($16.20 ARPU vs $2.20). While OpenAI commands a larger user base, Anthropic converts its users into greater revenue. This underpins Friar's caution and fuels Anthropic bankers' confidence in their October window.

OpenAI's recent strategic shifts, including the default enabling of advertising cookies for free users (effective April 30) and a reported 80% decline in ChatGPT Plus subscriptions, suggest a consumer-focused business grappling with its revenue architecture. This indicates a company less prepared for the rigorous financial scrutiny inherent in public market entry.

In Other News

Altman to Musk: "World Needs More Love" -- An Invitation Mid-Trial. The Musk v. OpenAI litigation produced a notable social moment Saturday: Sam Altman extended a public invitation to Elon Musk for the GPT-5.5 launch party on May 5, declaring "the world needs more love." This conciliatory gesture emerged as Business Insider detailed Week 1, dominated by Musk's testimony, his admission of model distillation, and the Shivon Zilis revelation. The trial resumes Monday, with the liability phase anticipated to conclude by May 21. The SF Chronicle confirmed live audio broadcasts for Week 2. Business Insider also profiled William Savitt, OpenAI's lead counsel, highlighting his previous legal victory against Musk in the 2022 Twitter acquisition. (Business Insider, Benzinga, SF Chronicle)

POLITICO Poll: Americans Don't Trust AI or Crypto -- But the Money Keeps Flowing. A POLITICO poll published Saturday revealed broad American skepticism towards both AI and cryptocurrency, despite industry super PACs channeling record capital into the 2026 midterm elections. This creates a palpable tension for candidates benefiting from significant financial support, as their voter base expresses unease with the very technologies funding their campaigns. A prior poll found three-quarters of Trump voters support AI oversight, even as the current administration advocates for deregulation. This widening chasm between public sentiment and policy direction is notable: the same week WIRED exposed the "Build American AI" dark-money campaign utilizing TikTok influencers to generate fear about Chinese AI, voters are signaling a desire for greater regulation, not less. This poll surfaces as the EU AI Act's August 2 compliance deadline approaches, with trilogue talks having collapsed, leaving companies facing enforcement without clear amendment guidance. (POLITICO, POLITICO)

Morgan Stanley: Chinese AI Stocks About to Reshape Hong Kong Markets. Morgan Stanley issued a Saturday note forecasting $1.25-1.75 billion in passive investment inflows to Hong Kong tech stocks, driven by the June 8 inclusion of Zhipu AI (via Knowledge Atlas Technology) and MiniMax in the Hang Seng Tech Index. The bank asserts that AI and large language model companies are poised to become a dominant force in Hong Kong equity markets, influencing index composition, performance, liquidity, and fund flows. This projection holds despite the Hang Seng Tech Index experiencing an 11% year-to-date decline. Morgan Stanley identified Alibaba as its preferred China internet selection, viewing it as a comprehensive AI play across its stack from cloud infrastructure to model development. Analysts also observed a rise in Chinese AI model pricing, with first-quarter costs reaching at least 17% of US peer pricing, diminishing the cost advantage that previously spurred adoption. (CNBC)

The Oscars Draw the Line: AI Acting and Writing Banned from Eligibility. The Academy of Motion Picture Arts and Sciences has codified new eligibility rules for the 99th Oscars (2027 ceremony), explicitly excluding AI-generated performances and scripts. Acting nominations now mandate a performance "demonstrably performed by a human" and credited accordingly. Writing categories require human authorship. The Guardian noted these rules also permit double acting nominations for the first time. This move formalizes previous implicit standards as AI-generated content increasingly integrates into pre-production workflows. TechCrunch clarified that these rules do not prohibit AI as a production tool—its use in visual effects, pre-visualization, or editing is permitted—but the core creative labor of writing and acting must remain human. This distinction positions the Academy as technologically discerning rather than entirely anti-AI, establishing clear boundaries for artistic recognition. (Reuters, The Guardian, TechCrunch, BBC)

X / Social Pulse

Altman's "world needs more love" post dominated Saturday's social discourse, eliciting divided reactions. Some perceived genuine magnanimity, while others interpreted it as a calculated PR maneuver mid-trial. Legal commentators frequently highlighted the optical advantage: Altman projected conciliation while Musk concluded three days on the stand admitting to model distillation and facing a skeptical judge. "Peak psychological warfare disguised as kindness" emerged as a prevalent sentiment.

The Apple CLAUDE.md leak continued to fuel developer discussions over the weekend. The May 1 revelation that Apple's Support app v5.13 contained internal Claude Code instruction files, referencing Apple's "Juno" AI system, confirmed suspicions regarding Apple's reliance on Anthropic's tools, even as it positions Apple Intelligence as a proprietary offering. Medium posts dissecting these leaked files garnered substantial attention, with developers noting the irony of Apple's notorious secrecy being undermined by a build-pipeline oversight.

Uber's CTO confirming the company expended its entire 2026 AI coding budget within four months became a central topic on Hacker News and r/artificial. Engineers debated whether 70% AI-originated code justified the token expenditure or indicated an impending cost crisis for enterprises deploying agentic workflows at scale.

A Psychiatric Times piece on escalating lawsuits against AI chatbot developers—addressing claims of suicide, addiction, and psychosis—circulated within health policy circles. Several practitioners noted that the Daily Mail's proposition to classify AI chatbot addiction as a mental illness signifies a notable shift from a fringe concern to a mainstream medical discourse.

One to Watch

The IPO Race as Industry Bellwether. The divergent IPO timelines for Anthropic and OpenAI transcend corporate finance; they are poised to establish the foundational valuation model for AI companies in the public markets. Morningstar's assertion holds true: the first to list sets the template. If Anthropic proceeds with an October IPO, potentially at $400-500 billion (as cited by BigGo Finance), it will affirm that an enterprise-focused AI model, characterized by high ARPU and disciplined capital expenditure, is more attractive than a consumer-scale model reliant on advertising revenue. Conversely, if OpenAI lists first despite its CFO's reservations, it will test whether user volume and brand recognition can counterbalance less robust monetization metrics. The implications extend beyond these two entities; every AI startup's future fundraising, every venture capitalist's portfolio strategy, and every public market investor's allocation model will be calibrated against the performance of this inaugural AI public offering. This race is less about valuation figures and more about seizing narrative control. Anthropic currently holds the strategic advantage.

Quick Hits

  • Apple's internal use of Anthropic's Claude Code was inadvertently confirmed when CLAUDE.md instruction files shipped with the Support app v5.13 update on April 30. Apple issued a patch within 24 hours. The files referenced an internal "Juno" AI system. (BeInCrypto, News9Live)
  • Goldman Sachs prohibited its Hong Kong bankers from utilizing Anthropic's Claude, citing data security and cyber risk concerns, particularly following China's persistent ban on Western AI tools. (Reuters, Bloomberg)
  • Health experts advocate for classifying AI chatbot addiction as a mental illness, citing reported cases of users experiencing suicidal ideation upon separation from chatbots. Psychiatric Times published an advance look at escalating legal challenges against AI companies. (Daily Mail, Psychiatric Times)
  • Uber exhausted its entire 2026 AI coding budget in four months following a surge in Claude Code adoption from 32% to 84% across its 5,000-engineer organization. CTO Praveen Neppalli Naga confirmed 70% of committed code is now AI-generated. (Startup Fortune, The Information)
  • Mayo Clinic's AI model identifies pancreatic cancer up to three years before clinical diagnosis using routine CT scans, detecting subtle indicators prior to tumor visibility. A clinical trial is currently underway. (Geo.tv, BioQuick News)

The brief pause in the Musk trial offers both parties an opportunity for strategic reassessment before a critical Week 2. However, the more enduring narrative, one that will profoundly shape the global AI market in the coming months, is unfolding not in a courtroom but in the boardrooms where Anthropic and OpenAI are enacting fundamentally divergent visions of an AI enterprise. One prioritizes robust enterprise revenue models and a streamlined IPO path. The other is implementing advertising cookies and contending with internal debates over its readiness for public market scrutiny. The market will ultimately render its judgment on which strategy prevails, and that judgment may arrive sooner than OpenAI's CFO would prefer.

Sources

Lock in. M. mazen@thorterminal.com

Read more