Interview, Podcast, Fireside Chat
Anthropic Raises $45B but Falls Short on Compute & Thoma Bravo Hand Back Medallia Keys to Creditors
- Venture capital returns are projected to rely heavily on a small number of massive exits, with ten "super-big exits" potentially covering total capital requirements, while public market valuations for non-AI SaaS are expected to revert to fair value metrics rather than narrative premiums.
- By late 2026 into 2027, most workflows are predicted to be managed by AI agents that autonomously select models and vendors, shifting market dynamics from human-anchored preferences to choices based on innovation and momentum, potentially ranking Anthropic above OpenAI and OpenAI above Gemini.
- Anthropic is projected to reach a $100 billion run rate within two years if growth continues, requiring significant capital expenditure that risks creating $150 billion in stranded capacity if targets are missed, while open questions remain regarding whether compute currently equals revenue versus future demand.
- Market leaders including OpenAI are expected to leverage their position to "own the agentic layer" by becoming the default API choice, whereas competitors like Canva face risks of being bypassed by agents or relegated to non-default status despite a potential wildly successful IPO.
- Hyperscalers such as Google and Amazon are anticipated to bundle chips with equity investments to capture margins, with Google predicted to benefit from infinite capacity to allocate resources across Gemini, Anthropic, and its own infrastructure, while NVIDIA is expected to remain the primary investment vehicle for the AI sector.
- Investor scrutiny is expected to intensify within a quarter or two, demanding specific metrics on agent usage and headless API calls, leading to the devaluation of contracts based on churn deferral if customers migrate to AI alternatives regardless of multi-year terms.
- Early-stage venture capital is forecast to consolidate around fewer, larger winners requiring billion-dollar revenues and 40% growth rates, with Y Combinator revenue guidelines becoming a standard for seed-stage transparency as traditional exits become limited.
- Mid-to-late stage investors anticipate tough consolidation acts for privately held assets approaching $2 to $3 trillion, with private equity models potentially disrupted if AI renders B2B software non-durable, while 2021 vintage PE deals are predicted to suffer significant losses affecting risk appetite.
- Social dislocation, including revolts and layoffs, is expected to persist throughout the decade, alongside a potential California "billionaire tax" and a shift toward founders transferring control to friends as a new exit route when IPOs are unavailable for sub-critical mass companies.
- Investment strategies include selling positions in companies like Figma and Duolingo if agentic workflow turnarounds are not demonstrated, while Robinhood and AngelList funds aim to provide public access to top-tier private firms like SpaceX and OpenAI despite valuation concerns.