Key Takeaways
- • Venture firm a16z (Andreessen Horowitz) is expanding its crypto fund and rebranding it as 'a16z Frontier' to invest in what Chris Dixon calls 'outer planets,' areas outside the firm's core verticals.
- • Dixon cited two motivations: crypto has become part of mainstream tech as stablecoins are adopted by companies such as Stripe and Wall Street engages in activities such as tokenizing stocks, and other a16z teams have ample opportunities in their core fields.
- • Frontier will operate with a 'lean toward being too early' stance, accepting the risk that a category may never materialize; Dixon cited Oculus and VR as an example of this risk.
- • Andreessen said the gap between what early AI adopters understand and what the public understands is the largest of his career, and argued that incumbents face disruption from tech-savvy startups challenging the broader economy.
- • Andreessen warned that government moves to regulate or ban open source pose a threat, while Horowitz argued for open research and broader access to model weights to address AI safety issues such as reward hacking.
Summary
Venture capital firm a16z (Andreessen Horowitz) has announced that it is expanding the scope of its crypto fund and rebranding it as 'a16z Frontier.' Chris Dixon, Marc Andreessen, and Ben Horowitz discussed the announcement and the firm's strategy in a podcast lasting about an hour. According to the source, Dixon described two motivations for the expansion. First, the positioning of crypto has changed: stablecoins are being used by companies such as Stripe and are becoming a standard fintech tool, and Wall Street is deeply involved through activities such as the tokenization of stocks, making it difficult to separate crypto companies from tech as a whole. Second, other teams at the firm have no shortage of opportunities in their core fields. Martin Casado refers to these core areas as 'inner planets' and to more eccentric areas as 'outer planets,' and Dixon sees potential coverage gaps there.
Horowitz explained that the firm has historically had Dixon investigate new areas and build verticals for each emerging field, but that projects not fitting a specific vertical were somewhat ambiguous as business initiatives. Frontier is intended to be a mechanism for investing in the most interesting founders in those areas, and it may eventually become an independent fund. Horowitz also noted that Dixon's team includes mathematicians and cryptography researchers, and that the team combines technical, business, and market judgments. Dixon said the firm has also invested in policy and regulation, which he expects to matter for autonomous driving, drones, and AI.
On investment philosophy, Horowitz described Dixon's framework of seeking 'good ideas that look like bad ideas,' meaning projects where the entrepreneur holds a non-obvious insight. Dixon said many successful technologies had predecessors that were roughly a decade too early, and that Frontier will lean toward being early, accepting the risk that a category does not materialize. He cited Oculus, noting that VR has not become as mainstream as expected 13 years on, though his conviction remains unchanged. Dixon also argued that AI is turning former 'science projects' into 'engineering projects,' allowing a 30-person team to build a full-stack device company that once would have required around 1,000 people.
Andreessen stated that the gap between AI understanding among early adopters and the general public is the widest of his career, and that AI-native startups view the technology differently from incumbents. He argued that many industries will enter a period of scramble, and suggested that companies are increasingly ambitious. Andreessen said that while many companies have not formally adopted AI, employees are using it without management's knowledge, and that individuals will increasingly found startups. He also argued that Clayton Christensen's disruptive innovation framework is harder to apply, describing a shift from competition among big tech firms to startups challenging the broader economy, and warned incumbents that 'the clock is ticking.'
On open source, the moderator referenced Peter Thiel's view that crypto is libertarian while AI is centralized. Andreessen assessed open-source AI as 'actively competitive,' with a gap of roughly 6 to 12 months versus closed models, but expressed concern that government moves to regulate or ban open source over the past three to four years are a threat. Horowitz argued that concentration of power has historically produced poor outcomes and that research labs are too focused on competition to address reward hacking adequately. These are the individual views of the a16z partners and should be weighed in that context.
Horowitz explained that the firm has historically had Dixon investigate new areas and build verticals for each emerging field, but that projects not fitting a specific vertical were somewhat ambiguous as business initiatives. Frontier is intended to be a mechanism for investing in the most interesting founders in those areas, and it may eventually become an independent fund. Horowitz also noted that Dixon's team includes mathematicians and cryptography researchers, and that the team combines technical, business, and market judgments. Dixon said the firm has also invested in policy and regulation, which he expects to matter for autonomous driving, drones, and AI.
On investment philosophy, Horowitz described Dixon's framework of seeking 'good ideas that look like bad ideas,' meaning projects where the entrepreneur holds a non-obvious insight. Dixon said many successful technologies had predecessors that were roughly a decade too early, and that Frontier will lean toward being early, accepting the risk that a category does not materialize. He cited Oculus, noting that VR has not become as mainstream as expected 13 years on, though his conviction remains unchanged. Dixon also argued that AI is turning former 'science projects' into 'engineering projects,' allowing a 30-person team to build a full-stack device company that once would have required around 1,000 people.
Andreessen stated that the gap between AI understanding among early adopters and the general public is the widest of his career, and that AI-native startups view the technology differently from incumbents. He argued that many industries will enter a period of scramble, and suggested that companies are increasingly ambitious. Andreessen said that while many companies have not formally adopted AI, employees are using it without management's knowledge, and that individuals will increasingly found startups. He also argued that Clayton Christensen's disruptive innovation framework is harder to apply, describing a shift from competition among big tech firms to startups challenging the broader economy, and warned incumbents that 'the clock is ticking.'
On open source, the moderator referenced Peter Thiel's view that crypto is libertarian while AI is centralized. Andreessen assessed open-source AI as 'actively competitive,' with a gap of roughly 6 to 12 months versus closed models, but expressed concern that government moves to regulate or ban open source over the past three to four years are a threat. Horowitz argued that concentration of power has historically produced poor outcomes and that research labs are too focused on competition to address reward hacking adequately. These are the individual views of the a16z partners and should be weighed in that context.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.