How AI Is Rewriting the Power Law of Venture Capital

The episode discusses how AI is reshaping the venture capital landscape, emphasizing that only 20 out of 3,000 US capital firms have achieved consistent 3X net returns over the past two decades, underscoring the extreme power law in the industry. AI's influence extends across various sectors of the economy, with its addressable market potentially surpassing traditional software applications.

David George highlights that while excessive capital can hinder most startups, Frontier AI stands out as an exception where investment can be directly translated into computational power, enhancing product offerings and competitive positioning. The AI market has demonstrated remarkable growth, reaching $100 billion in revenue within four years, a stark contrast to the 15 years it took for SaaS to achieve similar figures.

The episode notes that companies are remaining private for longer periods, with top venture capital outcomes projected to increase from $10 billion to potentially $100 billion. AI's impact on healthcare is anticipated to be ten times greater than that of traditional healthcare IT, while labor spending in the U.S. is approximately 40 times higher than software spending. The speaker cautions against viewing AI merely as an evolution of software, pointing out the uncertainty surrounding the economic value it can generate.

The loss rates for early-stage and growth-stage investments in venture capital are significant, with early-stage investments facing around a 60% loss rate. The average returns for venture capital over the last decade have ranged between 1X and 2X net, making consistent returns particularly challenging compared to other asset classes. Larger funds tend to achieve larger outcomes due to founder-driven dynamics, while smaller firms must maintain consistent fund sizes to avoid competition with larger players.

The episode introduces the concept of the "death of the middle," referring to the difficulties faced by funds that do not fit into the extremes of the venture capital spectrum. Founders are increasingly prioritizing capital from partners who can effectively de-risk their ventures. The evolving market allows early-stage firms to engage in pre-seed and seed investments, which can provide advantages in terms of access and relationships at the growth stage.

AI complicates the venture capital landscape, leading to larger and faster funding rounds and making it challenging to differentiate between genuine traction and inflated metrics. Limited partners (LPs) are becoming more cautious about investing in a volatile market, and their role is shifting as they navigate the complexities introduced by AI. There is a misalignment of incentives between general partners (GPs) and LPs, as LPs are not penalized for missing out on high-potential investments.

AI is increasingly influencing investment decisions across all asset classes, with companies leveraging AI experiencing accelerated growth in public markets. Private equity is actively seeking AI-native companies for new investments, and the integration of AI is expected to reshape portfolio construction in venture capital. Notably, private equity exits in venture capital have surpassed those in traditional private equity, with significant exits this year reaching around $50 billion each for companies like Electronic Arts and Medline.

The episode also highlights a significant shift in market dynamics, with current valuations for software assets in public markets being considerably lower than those in private equity. AI is recognized as a transformative force across various industries, although its integration into knowledge work beyond coding may take longer than expected. The median U.S. company spends $12 per employee on AI monthly, while the top 1% invests $7,000 per employee.

Fast-growing companies are reportedly generating more revenue per month than established mega-cap tech firms. In terms of venture capital trends, CalPERS has adjusted its portfolio, reducing its venture and growth investments from 91% to 58% while increasing its private equity stake from 9% to 43%. Valuations for software deals during 2021-2022 averaged between 25 to 32 times EBITDA, but current valuations are likely around half that figure.

The episode concludes by noting that while AI can enhance business operations, simply adding AI does not guarantee success without effective integration into workflows. A correlation exists between a decline in Net Promoter Score and revenue loss, indicating that AI is not a foolproof solution for generating returns. The average unicorn remains private for over ten years, and the IPO process can take 12 to 24 months before achieving liquidity. Looking ahead, robotics is expected to surpass language AI in importance, and new funds addressing supply bottlenecks in AI could create opportunities worth over $100 billion.