Interview, Fireside Chat
Jake Saper, GP @ Emergence Capital: "We Sold Salesforce Early and Lost Out on Billions"
- Nine out of ten early-stage deals raised subsequent rounds, with 20% achieving valuations over $1 billion and 10% going public.
- Private equity is expected to consolidate the no-code web app sector, leaving only nimble players who reinvent themselves.
- Most companies raised 5 to 7 years ago with $50 million to $100 million ARR face an existential crisis regarding AI investment and survival.
- Top-tier companies must grow at 400% year-over-year and demonstrate net dollar retention of 120% or higher to be considered generational.
- The technology adoption cycle will experience a "trough of disillusionment" where experimental budgets convert to real budgets and non-value companies are cut.
- A potential "FTX moment" in B2B AI is feared if a large enterprise agent performs a harmful action due to insufficient guardrails.
- The IPO market is projected to reopen at the beginning of next year despite current macro and political uncertainties.
- Open source Large Language Models (LLMs) are anticipated to dominate the market, with potential market shares ranging from 60% to 90%.
- Value in AI is expected to accrue to focused startups rather than incumbents, as startups can execute faster on narrow problems.
- AI implementation is predicted to focus on growing businesses more efficiently with existing headcounts rather than immediate workforce reductions.
- The AI market will shift toward outcome-based pricing, though this remains difficult to implement due to causality issues in human-loop scenarios.
- IBM faces potential trouble if startups successfully migrate legacy COBOL code from mainframes to the cloud efficiently.
- Specialized open-source LLMs will emerge for domain-specific tasks, such as mortgage analysis.
- B2B software vendors will persist despite AI coding tools due to the necessity for opinionated perspectives, maintenance processes, and accountability.
- AI capabilities are expected to improve significantly over the next two to three years, rivaling or exceeding the exponential growth seen between GPT-2 and GPT-3.5.
- Rote and less creative tasks are predicted to disappear, shifting human focus toward interpersonal skills, emotional understanding, and persuasion.
- Founders who do not require VC funding are viewed as optimal, though incremental improvements in success odds remain significant given low base rates.
- Investment firms utilizing "what you have to believe" frameworks will be better equipped to identify deal requirements, with these frameworks refined during diligence as data supports or negates hypotheses.
- Retention cohorts for current AI businesses are expected to disappoint on average, despite the presence of performance outliers.