digest
Short Summary
Main recommendation: Be cautious about the current AI boom, which Gurley argues looks like a real technology wave surrounded by speculative, bubble-like behavior, and if investing, favor AI-intersected opportunities built on deep industry expertise and proprietary workflows rather than broad bets that frontier labs could replicate. He also argues retail investors should be especially wary of SPVs and private-company exposure, and that everyone should actively learn to use AI tools to protect their careers.
Key Points
- Carlota Perez's book Technological Revolutions and Financial Capital (2002) is cited as a framework: real technology waves that create fast wealth inevitably attract speculators, carpetbaggers, and interlopers, so bubble-like behavior and real transformation tend to appear together.
- Gurley points to "circular deals" as a warning sign: Microsoft investing in OpenAI while OpenAI buys Microsoft services, and Nvidia giving money to CoreWeave while agreeing to buy its leftover services, are cited as questionable accounting-adjacent arrangements.
- Retail and private-market investing carries heavy risk: most VC-backed private companies go to zero, and private-company information is far less transparent or audited than public-market investors expect.
- Institutional investors currently show essentially zero interest in non-AI deals, which could leave non-AI angel-backed companies to die of neglect when raising future money.
- Gurley suggests the most defensible AI-era opportunities sit where a proprietary data set and multi-step, industry-specific workflows must be stitched together, which general frontier models are unlikely to handle quickly.
Details Worth Noting
- Concrete instructions:
- If angel investing, look for founders who combine deep domain expertise in a specific industry with heavy, hands-on use of AI tools, so they are the most AI-enabled practitioner in their vertical.
- Stay well away from the leading edge of what frontier labs like Anthropic or OpenAI publicly say they are building next; avoid startups aiming at the same core model capabilities.
- Prioritize niches the big model companies are unlikely to prioritize, such as "waste management" style verticals, where local workflows and customer-specific data must be integrated.
- Evaluate whether a company has proprietary data sets and enough workflows to build software around, rather than just answering questions a model can answer.
- Before investing in SPVs, assess your own risk tolerance honestly, since perceived tolerance often exceeds real tolerance without experience through major drawdowns; read the SPV terms, which typically state you could lose your entire investment.
- Protect your career by becoming the most AI-enabled version of yourself in your field, regardless of industry.
- Gurley's example of automatable workflows: at Zillow, a tool like ShowingTime books in-person home tours, while mortgage assembly and sign-offs involve many tasks that can be automated and integrated with AI.
- Gurley warns that SPVs are sometimes promoted by people who do not actually hold the underlying stock or merely hope to get it, which he calls the "wild, wild West."
- He notes that the 100x-plus returns from AI investments were largely made before the current wave, and that the odds of a new AI investment making money now are "really, really low," though he expects some will.
- A Silicon Valley CEO he spoke with reported having 50 angel investments, unreturned calls from founders, and regret, illustrating the unglamorous side of angel investing.
Bottom Line
Gurley's view is that AI is a real, important shift that is simultaneously attracting speculative excess, so the safest path is selective, expertise-driven exposure and skill-building rather than chasing hype or lottery-ticket private deals.
