PitchBook’s Q2 2026 research examines the imminent arrival of public market pricing for frontier AI, using SpaceX’s recent listing as the sector’s first live comparable. The analysis from PitchBook now centers on Anthropic and OpenAI as they approach potential IPOs, weighing business fundamentals against narrative-driven scarcity premiums.
The research report from PitchBook introduces the AIBQ framework, which scores frontier AI companies on a 0–10 scale across capital efficiency, revenue quality, compute independence, governance flexibility, and competitive durability.
Anthropic earns the highest composite of 8.20 (Strong), while OpenAI scores 4.53 and the xAI operations embedded in SpaceX score 4.49 (both Developing).
Despite superior quality metrics, Anthropic’s latest private valuation of roughly $965 billion translates to about $118 billion per AIBQ point—well below OpenAI’s $188 billion and the xAI segment’s estimated $345+ billion per point.
This positioning frames Anthropic as higher-quality operations available at a relative discount if public markets ultimately reward fundamentals.
Public valuation ranges remain wide and regime-dependent.
A quality-disciplined approach produces estimates near current private marks, while scarcity or narrative pricing benchmarked to SpaceX’s AI infrastructure block lifts values sharply.
The resulting bands span roughly $970 billion to $2.83 trillion for Anthropic (probability-weighted expectation around $1.43 trillion) and $540 billion to $1.56 trillion for OpenAI (centered near $950 billion).
Roadshow indications in the coming months will determine which regime clears.
SpaceX supplies the initial public benchmark but carries structural caveats. Its AI infrastructure block is valued at approximately $1.67 trillion within a sum-of-the-parts enterprise value near $2.6 trillion, supported by Starlink’s profitable connectivity business.
Starlink generated $11.4 billion in revenue and $4.4 billion in operating income, helping offset AI-segment losses of $6.4 billion on $3.2 billion of revenue.
This cross-subsidy enabled SpaceX’s record $75 billion IPO, yet standalone AI labs lack any comparable internal cash engine and must fund ongoing burn entirely through external capital.
Revenue recognition adds further uncertainty. Anthropic’s headline run-rate multiple of roughly 20.5x could rise to 50x–60x once restated on recognized revenue in public filings—levels approaching the richest sustained public AI multiples.
Upcoming S-1 disclosures will clarify any gross-to-net adjustments, potentially resetting cohort-wide perceptions before pricing.
PitchBook outlines three scenarios. In the constructive case, investor appetite extends to high-burn, later-profitability AI models with cleaner standalone stories.
The cautionary view highlights vulnerability: even SpaceX, with its strongest capitalization, required massive public capital; pure-play labs face longer paths to breakeven without profitable subsidiaries.
The base case expects SpaceX’s precedent to test broader tolerance for frontier AI’s extended timelines, with Anthropic likely pricing first and setting the initial comp for OpenAI.
The research report from PitchBook concludes that funding capacity is not the constraint—ample dry powder exists against modest new supply—but buyer appetite and rotation dynamics among existing AI holdings will prove decisive.
With key milestones including S-1 filings and roadshows expected through October 2026, the coming period will reveal whether public markets assign premiums to operational quality or continue to price scarcity and optionality in frontier AI. The PitchBook update has concluded that Anthropic’s quality edge and nearer profitability trajectory position it as relatively more resilient under either outcome.
Have a crowdfunding offering you'd like to share? Submit an offering for consideration using our Submit a Tip form and we may share it on our site!