Whale Rock Capital's highest conviction position is Anthropic, which Alex Sacerdote says emerged as a dark horse after the model race narrowed and coding proved the real unlock.
Sacerdote says Whale Rock invested in Anthropic in August 2025 at about a $180bn valuation and calls the position a home run.
Sacerdote says Whale Rock sold most of its software holdings because incumbents' AI products were not yet moving the needle, and he is net short software.
Whale Rock's Sacerdote on riding AI's S curve through Anthropic
Whale Rock invests by matching technology S curves to durable moats and underappreciated earnings power. Applied to AI, that produced a large bet on Anthropic, where compute is scarce, usage is early and products around the API build lock in. The same lens pushes the firm away from software and toward hardware suppliers where share is shifting fastest.
Takeaways
Anthropic became Whale Rock's top conviction. About 60 companies chased foundation models, and Whale Rock watched the field narrow until Anthropic emerged as an enterprise focused dark horse. Coding ran agentically in 2025, and the firm invested in August 2025 at roughly a 180bn dollar valuation.
AI usage is early and compute is scarce. Only about 0.1 percent of knowledge workers use AI deeply. Sacerdote expects usage to reach 15 percent within four years, and Anthropic revenue has gone from about 100m to 1bn and toward 9bn. He expects compute to run short.
Buy at S curve inflection with a moat. Whale Rock buys when an S curve inflects and earnings power is underappreciated, provided the company has a moat. It bought Nvidia and Tesla at low multiples. Standards and intellectual property protect Oracle and Qualcomm, while Nokia and Motorola lost their moats.
Software looks weak while hardware decommoditizes. Whale Rock sold most software because incumbents' AI products were not moving the needle, and it is net short software. Data center components are decommoditizing, with rising unit growth and shortages at suppliers such as Celestica and Corning.
Rate of share change beats level. Sacerdote proposes a rule of 40 for AI, adding AI share of sales to AI market share. Thirty percent of each scores 60. He says the rate of change in market share matters more than the level.
Stripe opened the door to private investing. Whale Rock invested privately in 2020 by buying Stripe in a secondary at a 35bn dollar valuation. It estimated take rates from disclosed volume above 500bn dollars, at 40 to 50 basis points for Stripe versus 25 to 30 for Adyen.
Risks center on sentiment and model progress. Sacerdote names public negativity about AI, slower model improvement, a shift to open source models, and wasted compute if a major player retreats. He sees the application layer as unclear, having watched Bret Taylor's Sierra.
How Whale Rock researches and packages ideas
Follows Philip Fisher's scuttlebutt approach, with thousands of face to face meetings with management teams each year.
A team of 10 experienced people runs the research engine, called the Whale Rock Learning Machine.
Tap a name to go there · tap a line to see why they are linked
On this episode 2
Show 1
Ran a long short fund for its first 15 years, then launched a long only fund in 2020, a private option, and a 2021 hybrid fund.
The new Whale Rock Mega Cap Tech Fund picks from the top 30 global tech companies, after Hamilton College's investment committee asked for mega cap exposure.