What Anthropic’s $965B Valuation Assumes About the Future of AI
Anthropic valuation challenges refer to growing doubts over whether the company’s near-trillion-dollar market value can be sustained while enterprises increasingly prioritize cheaper, good-enough AI models over the absolute performance frontier. The question is whether revenue, margins, and competitive dynamics can support such a premium. Anthropic’s headline numbers are striking: revenue rose from USD 100 million (approx. RM460 million) in 2023 to USD 4.5 billion (approx. RM20.7 billion) by mid-2025, with an annualized run rate of USD 47 billion (approx. RM216.2 billion). Its valuation then accelerated from USD 61.5 billion (approx. RM282.9 billion) in early 2025 to USD 965 billion (approx. RM4.44 trillion) by May 2026. Those figures sit behind its recent confidential IPO filing, which seeks to convert breakneck growth into durable public-market confidence. The tension is that public investors will scrutinize not only top-line growth, but also how vulnerable those revenues are to AI pricing competition and open-source Claude alternatives.
Cost Shock: When Your Biggest Customers Say You’re Too Expensive
Pressure on enterprise AI costs is no longer whispered in corridors; it is now on the record. Microsoft AI CEO Mustafa Suleyman told Bloomberg that “Anthropic is extremely expensive, and I think many people are urgently looking for alternatives,” adding that Microsoft’s long-term goal is to “reduce and ultimately eliminate” what it pays Anthropic. That is a remarkable statement from a major customer and strategic partner. At the same time, Microsoft is investing heavily to become a “top four” frontier lab alongside Google DeepMind, OpenAI, and Anthropic, rather than depending on external providers. The message to the market is clear: even deep-pocketed buyers are no longer willing to accept premium API pricing indefinitely. If one of Anthropic’s most significant customers is plotting an exit from that spend, smaller enterprises are likely asking whether similar savings are within reach through Claude alternatives.
Startups Find Millions in Savings With Open Models Like DeepSeek
The cost squeeze is even sharper for startups whose products run models continuously. Flo Crivello, CEO of AI agent platform Lindy, said his company “switched 100% of Lindy traffic to DeepSeek v4, churning from Anthropic models” and that the move “saves us millions of $ and we’re seeing an increase in performance on many core use cases.” For Lindy, inference spend had been its number-one cost, higher than payroll, so a 2–5x reduction was described as transformative. DeepSeek V4-Pro is priced at USD 3.48 (approx. RM16.03) per million output tokens, and on one benchmark, costs USD 1,071 (approx. RM4,937) versus USD 4,811 (approx. RM22,178) for Claude Opus 4.7. When calls reach into the billions, those ratios deliver large savings. These concrete switching stories highlight how AI pricing competition is shifting from theoretical risk to live budget decisions.
Open Source Is Caught Up in Months, Not Years
Beneath these customer moves lies a structural threat to Anthropic’s margins: the narrowing gap between frontier and open-source AI. Epoch AI’s analysis of the Epoch Capabilities Index suggests open-source models trail the frontier by about four months, and DeepSeek’s own V4 technical report places its lead model roughly three to six months behind state-of-the-art systems. If model capabilities plateau or slow, open-source ecosystems would need only a short window to match whatever level the frontier reaches. At that point, AI models risk looking like commodities, with prices converging toward raw compute and infrastructure costs. The fact that many real-world tasks do not require the absolute frontier, but merely reliable and cheap models, tilts the calculus further. For a near-trillion Anthropic valuation, that scenario compresses the premium it can charge for Claude before customers migrate to cheaper Claude alternatives.
IPO Scrutiny: Can Anthropic Grow Fast Enough, Cheap Enough?
Anthropic’s confidential IPO filing lands in a market that now expects clear answers on how long extraordinary growth can continue. Public investors will weigh its rapid revenue climb and USD 47 billion (approx. RM216.2 billion) run rate against visible signs of price pressure and substitution risk. At the same time, going public reshapes incentives: quarterly expectations push companies to prioritize growth and monetization, even while Anthropic maintains a mission rooted in AI safety. With OpenAI also filing to list, buyers will compare their economics closely. If large enterprises keep signaling that current enterprise AI costs are too high, and startups continue to prove that open or cheaper models can meet their needs, Anthropic must either cut prices or supply more differentiated value. The company’s challenge is to prove that its premium position is durable, rather than a brief phase before AI pricing competition erodes its advantage.
