This conversation features Kevin and Bennett from A-Star, a $1 billion AUM venture capital firm focused on seed-stage investing. The firm was founded five years ago and has just closed its third fund, emphasizing partnership with founders at the earliest stages.

The discussion covers the current venture landscape, including the impact of large multistage funds moving into seed, the rise of young AI-native founders, and the challenges of building application-layer companies while competing with AI labs. They also explore talent mapping strategies, the difficulty of maintaining focus at seed stage, and their skeptical view of AI-enabled roll-ups.

The Venture Landscape: Big Funds vs Boutique Seed Firms

Large funds charging 2% management fees on $5-10 billion create misaligned incentives compared to smaller seed-focused firms like A-Star's $450 million fund

Big funds use seed investing for 'option value' - building baskets of companies to see what pops for follow-on investments

"I don't think founders want to be thought of as an option, as part of a basket" - Bennett on the boutique advantage

Seed valuations have inflated dramatically: "Seats 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" - Kevin

The AI-Native Founder Revolution

Young founders are increasingly successful because "who cares if you know how to build a SaaS product... This is new for everyone" - Bennett

The trend mirrors historical patterns: "Steve Jobs founded Apple and Bill Gates founded Microsoft, they were 19" with acceleration through programs like Thiel Fellows

A new category of successful founders has emerged from research backgrounds, breaking the traditional anti-pattern of academic founders

Palantir produces exceptional founders because "Everyone's a mini CEO. Everyone had to build and launch a product and find product market fit"

Talent Mapping and Proprietary Deal Flow

A-Star maps talent rather than markets, focusing on "high-quality people want to hang out with high-quality people" in talent-dense nodes

More than half of A-Star's seeds are "pseudo-proprietary" - not cold outreach during fundraising processes but pre-existing relationships

The Decagon investment exemplifies this approach: Kevin had prior history with founder Ashwin from a previous company, making the seed investment "kind of a no-brainer"

"The half that we had some prior relationship or somewhat proprietary have disproportionately generated returns" - Kevin on deal sourcing effectiveness

The Concentration Strategy: Why Seed Funds Must Double Down

More than 50% of A-Star's dollars come after seed rounds, with concentration in 3-5 companies driving fund returns

"If all you're writing is the first check in seed, it's exceptionally hard to drive returns" - Bennett on the necessity of follow-on investing

The firm always supports companies through Series A pro rata, then makes concentration decisions in later rounds based on performance

Successful seed investing requires knowing "what great looks like at every level" to make proper follow-on decisions as companies mature

AI Market Dynamics and Application Layer Challenges

"The labs are not just infrastructure companies now. They are legitimate competitors at the application layer" - Bennett on the competitive threat

Traditional software companies face existential questions: "If you can vibe code an app in a matter of minutes, why do you need so many engineers?"

Successful application layer companies are building "systems of intelligence" and "systems of action" rather than traditional workflows

The best incumbent software teams can adapt by re-architecting "everything from the ground up" like portfolio company Ramp has done with AI agents

The AI Roll-Up Skepticism

AI-enabled roll-ups are "phenomenal for founders" who keep 80% ownership while VCs get diluted to 20% of acquired assets

"Unless these companies appreciate dramatically in value, the venture capitalist owns a relatively small percentage" - Bennett on structural challenges

The most promising roll-up targets are "services-oriented professions that have largely recurring revenue sources" like accounting and IT services

"90% of the value created is going to be buying the asset for the right multiple, which venture capitalists are historically very bad at" - Bennett

Resources Mentioned

Color by Number 80's Vibes | Black Background Edition | Mystery Mosaics Coloring Book for Adults | Pixel Art Coloring for Relaxation and Mindfulness ... (Color by Number - Black Edition | Pixelart)

eration through programs like Thiel Fellows A new category of successful founders has emerged from research backgrounds, breaking the traditional anti-pattern of academic founders Palantir produces

team out of Stanford called Simley

that's very high value, and that differs from going after a full horizontal market. Or we backed a research team out of Stanford called Simley, which is focused on market research, which went on to r