The dawn of a new startup era
๐ Abstract
The article discusses the changing landscape of the startup ecosystem and the implications for venture capital performance in the next decade. It explores three broad categories of software startups being founded today - AI frontends, AI infrastructure, and AI full-stack products - and analyzes the expected value and defensibility of each category.
๐ Q&A
[01] The Changing Startup Landscape
1. What are the key factors driving the changing dynamics in the startup ecosystem?
- The article argues that the startup era of the last couple of decades has been about betting on the depth of market penetration of existing technologies, rather than technical invention. This has led to high consensus and competition, reducing the likelihood of delivering outsized returns.
- Factors contributing to this include:
- Increased accessibility of tech tools and education, reducing barriers to entry
- Incumbents becoming more adept at building digital products, increasing competition
- Tech attracting a different type of talent focused on extraction rather than value creation
- Deflationary pressures on margins due to increased supply and demand
2. How does the author view the prospects of "AI frontends" or "wrappers" as a category of startups?
- The author believes that AI frontends, which embed third-party language models, will face high consensus and low defensibility, as the monetization and distribution techniques on the internet are mostly figured out.
- While these startups can build successful $1M ARR businesses, the author does not expect them to yield 100x market returns, as the expected value is packed around the market average.
[02] AI Infrastructure and Full-Stack AI
1. What is the author's perspective on AI infrastructure startups?
- The author sees AI infrastructure startups as having higher expected value, as they are making bets on what the post-AI world will look like and building the necessary tools.
- However, the author also sees these startups as having high variance in outcomes, similar to the dot-com era, as the market demand is still undefined.
- The author advises founders to be cautious about this category, as success will largely depend on factors outside their control.
2. How does the author view the potential of "AI full-stack" startups?
- The author is most optimistic about startups that are willing to take on more technical risk and control the entire data feedback loop, developing proprietary AI models.
- These companies will need to make AI a core competency and hire specialized talent to leverage their data advantages.
- The author believes these "extreme technical bets" have the potential to be the true market crushers of the next two decades, as they can build defensible positions by accessing data not available on the internet.
[03] Funding Dynamics for the New Startup Era
1. How does the author expect the funding dynamics to change for the new breed of startups?
- The author suggests that the funding for these "extreme technical bets" will be different from the software startup factory model, as they are more capital-intensive and require longer incubation periods.
- Successful examples have often involved partnerships with large tech companies, government subsidies, or concentrated bets from individual investors, rather than the traditional VC funding model.
- Investors will need to be more comfortable with technical risks rather than just market risks, and will need to do more scientific due diligence.
2. What is the author's overall outlook on the future of the startup ecosystem?
- The author believes we are witnessing the dawn of a new startup era, where the bulk of returns will shift from growth-focused internet startups to those willing to take on more technical risk and develop defensible positions through proprietary data and AI models.
- Success will be more unevenly distributed, as the new opportunities may not be easy to seize for most people.