Uncapped with Jack AltmanHow AI Is Rewriting Seed Stage Investing with Kevin Hartz & Bennett Siegel | Ep. 49
At a glance
WHAT IT’S REALLY ABOUT
A* Capital on seed investing incentives, bubbles, and AI talent shifts
- A* Capital positions itself as a seed-first, founder-partnered firm, arguing that mega-funds’ fee structures and option-value strategies distort early-stage incentives.
- They predict an AI-driven boom that may become a historic bubble, with much higher seed/A/B valuations and inevitable “carnage” for companies that can’t grow above their preference stacks.
- They see founders skewing younger and a new wave of AI researcher-founders, making “mapping talent, not markets” central to sourcing and underwriting at seed.
- They argue seed is persistently hard because outcomes are highly skewed and require heavy post-seed follow-on concentration rather than “peanut butter” pro-rata across the portfolio.
- They’re skeptical of AI rollups as a venture return strategy, believing culture/process change and margin work are difficult—and that most value comes from buying assets at the right multiple, which VCs are not optimized to do.
IDEAS WORTH REMEMBERING
5 ideasFee structures drive fund behavior more than most people admit.
They argue that if management fees were capped after a certain AUM, many large firms would change strategy; “two and twenty” at multi-billion scale incentivizes asset gathering and earlier-stage “toeholds” rather than craft-focused seed partnering.
Mega-funds invest in seed primarily to buy options, not to be long-term partners.
The described playbook is to build a basket at seed, then double/triple/quadruple down on winners; they note founders may accept this for higher valuations and more capital, but it can reduce hands-on support and increase rookie mistakes.
AI is inflating early-stage prices and compressing stage-to-stage milestones.
They cite seed moving from ~20–30 post-money to ~40–50, strong Series A shifting from ~100 post to ~250, and Series B expanding toward ~$500M–$1B—raising the risk of down rounds or being trapped under a heavy pref stack.
Seed underwriting is shifting toward talent identification amid platform uncertainty.
Because “nobody is differentiated at seed,” they look for obstacle-overcoming founders, talent-dense networks (top schools, accelerators), and “founder factory” companies (notably Palantir) rather than betting purely on clear market maps.
AI is producing a new founder archetype: researchers who can commercialize fast.
Researcher-founders were once an anti-pattern (high intellect, low commercial instinct), but foundation-model era dynamics let technical differentiation matter immediately—so investors must listen for precision, intent, and business articulation in how they speak.
WORDS WORTH SAVING
5 quotesI think, like in any platform shift, you're gonna have some incredible companies that come out of this vintage, and in fact, they're probably gonna be larger than anything we've ever seen, but you're also gonna have a lot of companies that aren't going to go the distance.
— Bennett Siegel
Seeds used to be 20 to 30 posts. They're now 40 to 50. A great Series A used to be 100 posts. Now they're happening at 250.
— Bennett Siegel
It's either too far one direction or too far the other.
— Kevin Hartz
I don't think founders wanna be thought of as an option, as part of a basket.
— Bennett Siegel
You work incredibly hard to build relationships with founders and earn the right to invest two, three, four, $5 million in a company that by all odds will likely not work, 'cause most companies do not become billion-dollar-plus companies, and if it's not a billion-dollar-plus company, it has almost no relevance to the perfo- uh, performance of your fund.
— Bennett Siegel
High quality AI-generated summary created from speaker-labeled transcript.
