A financial analysis published March 15, 2026 by Joseph Blumenfeld on Substack puts a number on what the AI industry already suspects: 2026 is shaping up to be a historically unprecedented year for public market debuts. Five companies — the pending SpaceX/xAI merger (~$1.75T, though the deal has not yet closed), OpenAI, Anthropic (the two LLM giants combining for ~$1.2T), Stripe, and Databricks — are each expected to go public this year, collectively injecting over $3 trillion in market capitalization into public equities. If realized, that would place four of these listings among the five largest IPOs in history.
The Anthropic and OpenAI filings are the centerpiece for the AI agent ecosystem. Blumenfeld notes that Anthropic, founded in 2021, is currently leading in key AI capabilities including coding — a direct reference to Claude's performance on software engineering benchmarks — while OpenAI remains the cultural touchstone credited with transforming how people interact with AI through ChatGPT. The SpaceX/xAI combination, if it closes, turns what was previously a pure space and satellite play into an AI, space, and data conglomerate. Starlink's reported $10B-plus in 2025 revenue gives it a tangible business foundation that neither OpenAI nor Anthropic can yet claim at comparable scale.
Blumenfeld's historical backtest is what separates the piece from a straightforward valuation story. He examined the 1980 IPO class — which included Apple (valued at $1.78B at IPO) and Nike — and found that equal-weight investing across all 234 companies that went public that year would have returned roughly 32x a $1 investment. The S&P 500 with dividends reinvested returned approximately 160x over the same period, beating even a vintage that happened to include two of the greatest wealth-creating stocks ever listed. The pattern held across the early 1980s: outside exceptional years anchored by generational companies, broad IPO investing consistently underperformed passive indexing.
Blumenfeld concedes the 2026 cohort looks nothing like prior IPO waves — these are mature, revenue-generating businesses rather than speculative startups. But his conclusion is blunt: <a href="/news/2026-03-14-ai-valuation-investor-uncertainty">picking individual IPO winners is extremely difficult</a>, conviction to hold through multi-decade underperformance is rarer than investors assume, and the S&P 500 has an uncomfortable habit of winning even in banner vintage years. The companies may reshape their industries, but that does not automatically translate into market-beating returns for public shareholders.