Interview
Alex Kolicich, 8VC | AI & Defense Tech Renaissance
- The winners of the current AI phase shift will be determined within a two to three-year timeframe, necessitating a transition from traditional SaaS metrics as business models and durable moats remain uncertain.
- Foundation model lifespans are projected to be on the order of months or years, where failure to sustain innovation cycles results in a total loss of value for the company.
- Value capture is expected to reside in incumbent software systems and data gravity locations rather than the base model layer, driving significant revenue growth as these systems integrate AI workflows.
- The current elevated AI multiples are contingent on an oversupply of venture capital and may normalize, creating fundraising difficulties if capital dries up.
- The AI abilities gap is unlikely to close soon without fundamental transformer research advances, with models potentially remaining capped at an 80% accuracy threshold requiring human oversight.
- The future software paradigm is predicted to evolve into man-machine symbiosis where AI automates 80% of work under human supervision rather than achieving full AGI autonomy.
- Decreasing returns on investment are anticipated due to scaling laws where order-of-magnitude compute increases yield only sublinear performance gains.
- The market is currently viewed as potentially at or above the peak of inflated expectations and is expected to eventually enter a trough of disillusionment when revenue fails to match infrastructure investment.
- A large number of venture capital managers may retire or be terminated over the next decade for failing to adapt investment strategies to the new AI paradigm.
- IPOs in software and biotech are projected to remain in the single digits for the foreseeable future, while M&A activity is expected to cool as acquirers wait for value clarity.
- A secondary market opportunity exists for sovereign wealth funds to acquire positions in 10-year-old venture funds that lack large exit options.
- Defense startups will require significant capital with minimal commercial traction due to long deployment and procurement cycles, though the total number of new defense primes will be limited to the order of tens.
- High-profile leadership departures from major AI companies prior to an IPO could signal instability and irregular operational conditions.