AI Summary
→ WHAT IT COVERS Harry Stebbings, Jason Lemkin, Rory O'Driscoll, and MongoDB CEO Dave Itzchak analyze OpenAI's surge to a $70B run rate, the Cognition-Factory adviser controversy involving Vinod Khosla's public criticism of a portfolio company, Salesforce's $2B Listen Labs acquisition, ElevenLabs doubling its valuation to $22B, and agents reshaping vendor selection. → KEY INSIGHTS - **AI Model Loyalty:** Developer switching costs between LLMs are low and accelerating. OpenAI grew GAAP revenue 70% quarter-on-quarter in Q3, reversing a period where Anthropic dominated coding workloads. Enterprises that signed procurement contracts and trained thousands of employees on one model face higher switching friction than developers, making enterprise embedding the real competitive moat rather than raw model performance benchmarks. - **Agent-Driven Vendor Selection:** Agents are replacing human-driven vendor discovery. Vercel grew to $600M ARR with 50% of new business from agents, up from 3% at the start of the year. When an agent selects a vendor based on stack compatibility and documentation quality, it creates durable demand. Startups must optimize their APIs, documentation, and MCP presence specifically for agent consumption, not just human search. - **Adviser Conflict Management:** The Cognition-Factory situation reveals that implicit confidentiality expectations between founders and board observers are insufficient. Founders should establish explicit written agreements with advisers who access product roadmaps, competitive strategy, and financial plans, specifying restrictions on joining direct competitors. The absence of formal fiduciary obligations for non-board advisers creates legal and reputational grey zones that damage both parties. - **VC Public Criticism Risk:** Vinod Khosla's public description of Factory as a "struggling second-tier competitor" — despite Khosla Ventures leading the company's Series C — handed competitors a permanent recruiting and deal-closing weapon. The panel consensus: any VC publicly criticizing a portfolio company, regardless of frustration, provides rival firms a durable narrative about partner behavior during adversity that compounds over years. - **Open-Source US Model Opportunity:** Enterprises express strong reluctance to deploy Chinese-sourced open-weight models, even under cost pressure, particularly in regulated industries. A US-based open-weight model performing within six to twelve months of frontier quality could capture significant token volume. Cost pressure is intensifying: if OpenAI and Anthropic reach $70B combined ARR, that represents a material share of US corporate profits, forcing CFOs to mandate token budget optimization. - **Acqui-hire Legal Exposure:** Former Groq engineers are suing in Delaware over NVIDIA's $11B license plus $3B stock deal, arguing common shareholders were left with a hollowed company while departing employees received disproportionate consideration. Delaware corporate law requires equal treatment of all common shareholders. If courts classify these structures as de facto acquisitions, the legal and double-taxation risks could effectively eliminate acqui-hire licensing deals as a regulatory workaround. → NOTABLE MOMENT The panel noted that Vercel — founded without any anticipation of agentic workloads — now derives half its new business from agents autonomously selecting it as infrastructure. One panelist demonstrated live that an agent recommended an unfamiliar hosting provider over Vercel for a specific use case, illustrating how vendor selection has shifted entirely outside human decision-making. 💼 SPONSORS [{"name": "Base44", "url": "https://base44.com/20vc"}, {"name": "AlphaSense", "url": "https://alphasense.com/20vc"}, {"name": "Intercom Fin", "url": "https://finn.ai/20vc"}] 🏷️ OpenAI vs Anthropic, Agentic Infrastructure, Venture Capital Ethics, AI Model Switching Costs, Open-Source LLMs, Acqui-hire Legal Risk
