a16zThe Frontier Ideas a16z Wants to Fund Now | Marc Andreessen, Ben Horowitz & Chris Dixon
At a glance
WHAT IT’S REALLY ABOUT
a16z launches Frontier to back AI-, crypto-, and agent-driven leaps
- a16z is rebranding and expanding its crypto fund into “a16z Frontier” to invest in cross-cutting, early, high-variance technologies that don’t fit existing vertical practices.
- The speakers argue that AI is accelerating technical feasibility and compressing timelines, creating founders with unprecedented leverage and widening the gap between AI-native startups and incumbents.
- They expect a surge in startup formation as AI is adopted first by individuals rather than institutions, enabling small teams to build full-stack products across software, hardware, and media.
- They foresee AI reshaping the internet into an agent-heavy ecosystem and disrupting current attention and distribution economics, pressuring legacy content and marketplace models.
- They make the case that open source AI and crypto rails (e.g., stablecoins) can decentralize power, broaden education and safety research, and enable new agent-to-agent economies.
IDEAS WORTH REMEMBERING
5 ideasa16z Frontier expands the crypto fund into a broader ‘outer planets’ mandate.
a16z is broadening its crypto investing mandate into “Frontier” because crypto is increasingly interwoven with fintech, capital markets, AI, and other emerging tech. The goal is a vehicle and team explicitly designed to back out-of-vertical, high-uncertainty bets rather than forcing them into existing practices.
They’d rather be early than perfectly timed—and price that into the strategy.
They argue the hardest variable in breakthrough tech is not whether it works, but when it becomes viable, because timing depends on intertwined technical, cultural, regulatory, and market factors. Frontier is explicitly willing to accept ‘category risk’ and be early, even if adoption takes much longer than expected (e.g., VR).
AI is creating the widest early-adopter vs. mainstream capability gap they’ve ever seen.
Andreessen describes an unusually large and widening gap between what AI-native founders believe is possible and what incumbents (and even many tech companies) understand or can operationalize. This gap, plus accelerating tooling, is expected to drive faster company formation and more ambitious, multi-domain “full stack” startups.
AI adoption is bottom-up: employees move first, institutions lag, startups proliferate.
A key pattern they highlight is individuals adopting AI faster than institutions, sometimes without management’s visibility, which lowers the barrier to entrepreneurship. That dynamic suggests both productivity leaps inside firms and a surge in employees leaving to start AI-native companies.
The next internet may be dominated by AI agents, not humans.
They anticipate the internet evolving toward an “agent ecology” where bots/agents become the most numerous actors online. This creates new categories—agent-to-agent communication, payments, trust networks, and ‘multiplayer AI’ (swarms/fleets)—and raises questions about identity, coordination, and transaction rails.
WORDS WORTH SAVING
5 quotesThis is the biggest delta between what I would say what the early adopters understand is possible now, uh, versus what kind of everybody, kind of normal people understand.
— Marc Andreessen
We used to kind of ask the question is, is this idea too ambitious? And I, and I think, I think we're much more likely to ask this, is this ambitious enough?
— Ben Horowitz
I think startup formation is gonna skyrocket from here.
— Marc Andreessen
Basically, I don't know of a startup that wasn't a shit show at some point.
— Chris Dixon
Close your eyes and picture the concept of government sushi. Um, and then keep your eyes closed and picture government AI.
— Marc Andreessen
High quality AI-generated summary created from speaker-labeled transcript.
