The stack spans foundation models, inference and training silicon, cloud orchestration, and the application layer. Listed exposure is concentrated in silicon and cloud names already priced for AI scenarios. The most differentiated pre-IPO opportunities sit in the foundation-model and specialist-silicon layers.
OpenAI is the category-defining business in foundation models and the closest available proxy for the broader generative-AI thesis as a single holding.
Anthropic operates as the principal differentiated alternative at the frontier-model layer, with Claude gaining traction in enterprise coding and analytical workloads. Most allocators hold both.
Groq is one of the most credible challengers to incumbent GPU silicon for high-throughput inference. Narrower thesis, asymmetric payoff.
Regulation across multiple jurisdictions, model commoditisation as the gap between leading and second-tier models narrows, and the capital intensity of successive training generations driving pre-listing dilution.
OpenAI Pre-IPO Anthropic Pre-IPO Groq Pre-IPO IPO Index Investor Platform Pre-IPO Due Diligence Private Equity Access All Insights