Why the AI Boom Is Just Getting Started

The conversation centers on Whale Rock Capital's investment thesis around AI, with a particular focus on their highest-conviction position in Anthropic. After ChatGPT launched in November 2022, Whale Rock conducted a deep dive and initially focused on chips and infrastructure, reasoning that compute demand would be certain regardless of which AI companies won. Over time, they developed conviction that the foundational model layer would consolidate into an oligopoly of three players — Anthropic, OpenAI, and Google/Gemini — similar to how cloud computing consolidated around AWS, Azure, and GCP.

The key inflection point for Anthropic came with the explosion of AI coding tools in 2025. When Claude Code went largely agentic and Andrej Karpathy publicly stated that AI now writes 80% of his code (a reversal from the prior year's 20%), the investor recognized a massive market opportunity. With approximately 20 million coders worldwide potentially spending $20,000–$30,000 per year on tokens, the coding market alone represents a ~$500 billion opportunity. Whale Rock invested at a $180 billion valuation after preparing a 90-page deck using Claude Code to analyze the competitive landscape.

The S-curve framework is central to Whale Rock's investment philosophy, combined with identifying competitive advantages and underappreciated earnings power. The investor explains that technology adoption follows predictable S-curve patterns, but the key variables are: when barriers to adoption are removed (triggering the inflection), how tall the curve is (determining when to sell), and the slope of adoption (B2B is slower like a dishwasher requiring installation; consumer can be fast like radio). AI is described as an 'L-curve' — straight up — because unlike cloud/SaaS, users simply open a browser and it's there, with enterprise penetration still under 1%.

On hardware and chips, the investor argues we are in a 'renaissance of decommoditization.' AI workloads are growing 10x annually and pushing hardware to physical limits, creating innovation cycles across every layer — from high-bandwidth memory to PCBs requiring 40 layers instead of 10, to liquid cooling, to fiber optics. Companies like Celestica (AI server manufacturing), Corning (fiber), and Elite Materials (copper clad laminate) are highlighted as having transformed from commodity businesses into critical infrastructure suppliers with strong competitive positions and rapidly rising ASPs and margins.

Conversely, traditional enterprise software companies are described as being in serious trouble. Whale Rock sold almost all application software and was net short entering the year. CIO budget priorities have shifted toward AI tokens with faster ROI, seat-based software faces pressure from AI reducing headcount, and the threat of AI-native competitors building replacements is real. The one potential positive for incumbent software is if AI agents end up operating within existing systems of record (like CRM or Slack), which could reinforce their positions.

The investor also discusses how they access private markets despite being primarily a public markets firm — through deep due diligence, building relationships over time, and demonstrating long-term ownership commitment (as opposed to VC sellers). He describes the Whale Rock 'learning machine' as 2,500–3,000 face-to-face management meetings per year, compounding knowledge over 20 years, and using a scuttlebutt approach from Philip Fisher. The firm has evolved its product offerings from a long-short fund to include long-only, hybrid public/private, and a new Mega Cap Tech fund targeting the top 30 global market caps.