Anthropic IPO rumors and pre-IPO investor access

The headline access story isn't equity — it's a derivative. On May 25, 2026, BingX opened USDT-margined perpetual futures tied to ANTHROPIC (Pre-IPO) and OPENAI (Pre-IPO), letting traders take leveraged long or short positions without owning a single share ^[4]. These contracts settle entirely in stablecoin and reference a private valuation mark rather than a tradable security: the Anthropic contract priced near $1,800 against its February ~$380B raise, the OpenAI contract near $1,600 against an ~$852B round ^[5]. BingX joins Binance, OKX, and HTX in listing these instruments, a sign that exchanges have found genuine retail demand for AI-startup exposure that public and secondary markets simply don't supply ^[5]. The catch is that 'access' here means tracking a number, not buying into a cap table — and because the contracts are leveraged perpetuals, a sharp move in either direction can liquidate a position before any actual IPO ever exists ^[5]. It is exposure to a story about Anthropic, dressed as exposure to Anthropic.

Anthropic's valuation curve is the real protagonist. It went from roughly $61.5B in March 2025 to $183B that September, to near $350B after a Microsoft-Nvidia package in November, to $380B in February 2026, and finally to $965B post-money in the $65B Series H that closed in late May ^[2]. Revenue moved too — run-rate crossed $47 billion in May 2026, up from about $9 billion at the end of 2025 ^[3]. But the gap between price and proof is exactly what the IPO would expose: an S-1 forces public-company disclosure in place of private run-rate marketing, and reporting notes all three major AI issuers have been losing more than they make ^[8]. That tension is precisely what fuels the dominant skeptical voice on social platforms, where one widely shared post juxtaposed Anthropic's near-$1T valuation against Samsung's larger revenue and smaller market cap to argue the AI complex is a bubble. Yahoo Finance frames the stakes bluntly: the 2026 window is either the most consequential IPO cycle since the dot-com era or the most expensive lesson in narrative-versus-fundamentals markets have ever taught ^[7].

Filing first is a procedural lead, not a verdict. Anthropic confidentially filed its preliminary S-1 on June 1, 2026, ahead of OpenAI ^[1], and prediction-market traders assigned roughly a 76% implied probability that Anthropic reaches public markets before OpenAI ^[6]. Yet commentary stresses that being first to file doesn't mean being first to actually IPO — OpenAI has worked its process for years, engaged Goldman Sachs and Morgan Stanley, and could still list first ^[6]. The deeper contrast is in the revenue base: over 80% of Anthropic's revenue comes from enterprises, a structurally stickier mix than OpenAI's consumer-heavy ChatGPT footprint, which analysts treat as an advantage rather than a finish-line guarantee ^[6]. Amodei's own framing is that this is fundamentally about capital — 'a really big upfront cost to train the models and to serve inference on them' — and that the public market is simply where frontier labs will have to go for funding ^[3]. With Anthropic, a coming OpenAI filing, and SpaceX's ~$1.77T target stacking up, the window could be the largest simultaneous capital raise ever brought to market ^[7].

Strip away the valuation headline and the cost structure tells its own story. Enterprises supply over 80% of Anthropic's run-rate revenue, the stickiest kind of demand for a frontier lab, yet a single compute partnership reportedly runs $1.25 billion per month ^[3]. That is the engine and the anchor at once: the same enterprise traction that justifies a steep revenue ramp is being chased with a burn rate large enough to keep the company perpetually capital-hungry. It reframes Amodei's 'access to capital' argument as less a choice than a treadmill — the public market isn't just convenient, it is structurally necessary to sustain spending of that magnitude. Bulls read the enterprise mix and the revenue slope as proof the spend converts; skeptics read the monthly compute bill and ask what happens to margins the first time growth merely slows.

Beneath the finance-media excitement runs a sharply skeptical community current. On Claude-focused forums the reaction to the filing skewed overwhelmingly negative, with users predicting 'enshittification' — the worry that quarterly public-market KPIs will pressure product quality — and dismissing retail IPO participation as a sucker's bet, alongside a contrarian minority buying precisely because sentiment is so negative. A related bearish theory holds that Anthropic is rushing to list before OpenAI is 'forced' to IPO and exposed as a cash bonfire, with the framing that retail buyers become 'exit liquidity' for earlier investors. The access discussion is the most telling part: when retail traders ask how to actually buy pre-IPO Anthropic shares, the practical answer is that they largely can't — secondary marketplaces had no inventory, Anthropic invalidated external SPVs as of mid-May, and the suggested workarounds collapse into proxy plays, expensive secondaries, prediction markets, and exactly the kind of crypto perpetuals BingX is selling. The gap between locked equity and open synthetic leverage is the whole story.