Version 2

SemiAnalysis ran empirical tests on top-tier AI subscriptions and found heavy users pay 40–70x less than equivalent API rates [3][2]. The Wall Street Journal reported OpenAI is considering deep price reductions driven by Anthropic's growth in developer and coding workflows [6][7]. Anthropic raised $65 billion in a Series H at a $965 billion post-money valuation [11][13] and filed confidentially for an IPO [14], while OpenAI followed with its own confidential IPO filing at an $852 billion valuation approximately one week later [16][17].

Both companies are moving toward public markets while carrying a subscription model where heavy users extract far more compute value than their fees imply at API rates. The pricing decisions each company makes now — on subscriptions, API rates, and the gap between them — will shape the financial narratives they present to public investors.

SemiAnalysis published a four-part analytical thread in June 2026 comparing subscription and API business models for AI labs [1]. The firm purchased top-tier plans from both Anthropic and OpenAI and ran long-horizon coding tasks until hitting weekly usage limits, then compared the implied token cost to API rates for the same volume [2]. The finding: heavy users pay 40–70x less on subscriptions than they would via the API; the API's primary advantage is automation and product-integration use cases, not interactive or coding workloads [3]. SemiAnalysis frames subscription margin as a function of average utilization across the full subscriber base and argues that explicitly throttling subscriptions would generate backlash — but that falling inference costs will eventually close the economics gap, making frontier-quality models profitable at $20/month without throttling [4][5].

The Wall Street Journal reported on June 11 that OpenAI is considering deep price reductions, with Anthropic's growth in developer and coding workflows — particularly Claude Code — identified as the direct competitive pressure [6][7]. Agentic coding sessions generate unusually high token volumes per user, making subscription pricing especially consequential for both companies' revenues. The WSJ reporting was picked up broadly across the tech press [8][9][10].

The competitive and financial context sharpened further with a series of corporate disclosures in late May and early June 2026. Anthropic closed a $65 billion Series H at a $965 billion post-money valuation [11][12][13], surpassing OpenAI's publicly known valuation, and filed confidentially for an IPO dated June 1, 2026 [14][15]. OpenAI filed confidentially for its own IPO at an $852 billion valuation approximately one week after Anthropic's filing [16][17]. Anthropic's revenue reportedly grew from $87 million to $30 billion over 27 months [18], and broad capital-market appetite for the sector is evident in $159 billion of AI tech bond issuance through early June 2026 [19].

Taken together, these threads describe a market in which subscription pricing is strategically aggressive — delivering heavy users far more compute value than their fees imply — while both companies simultaneously prepare to justify their valuations to public-market investors. SemiAnalysis argues the economics will self-correct as inference costs fall; OpenAI's reported price-cut deliberations suggest it views the gap as requiring an active competitive response rather than a natural one.