HSBC Asset Management: Venture capital, investing in the age of AI

Public investors' AI money has mostly gone to sellers of tools and infrastructure, such as cloud services and chips, rather than to AI solution providers. The note argues that private markets are the more direct route. It names Anthropic, OpenAI and Databricks as examples of companies unavailable on exchanges. It also flags themes it expects to persist: agentic and vertical AI, data privacy, compute efficiency and tighter regulation. The concentration is visible in HSBC's companion VC piece. It cites PitchBook NVCA data showing AI at 65.4% of deal value but 39.4% of deal count in 2025. That gap implies a small number of very large rounds. HSBC Asset Management also invests in the area itself. Its VC strategy is a fund-of-funds program that commits to established venture funds and makes selective co-investments. One recent example is Model ML, a London-founded agentic platform for financial services. For founders, the signal is where institutional allocators say they want exposure: application and vertical-AI companies, not only infrastructure. Capital is concentrated in a few large rounds, so funding for everyone else is likely to depend on showing a clear use case and a route to liquidity.