20VC: NVIDIA Crushes Quarter and Buys Hugging Face | OpenAI Cuts Off Cursor | Instinct Hits $2.5BN Valuation and The Race for AI Assistants | Cognition Raises at $46BN, Linear $2.5BN and Clay $7BN

What It Covers

Harry Stebbings, Rory Driscoll, and Jason Lemkin break down NVIDIA's record $96B quarter and near-$13B Hugging Face acquisition, OpenAI cutting off Cursor, the AI assistant category led by Instinct at a $2.5B valuation, and late-stage rounds for Cognition at $46B, Clay at $7B, and Linear at $2.5B.

Key Questions Answered

• •NVIDIA Demand Signal: NVIDIA guided 70% revenue growth for fiscal year ending January 2028, well above the 44% analysts projected. The only three risks that could derail this are hyperscalers stopping compute purchases, financing structures collapsing, or end-user demand failing to materialize. All three remain stable, making NVIDIA's forward trajectory the clearest signal that AI infrastructure spending continues accelerating for at least 12 more months.
• •Open Source as Compute Strategy: NVIDIA's $12.9B Hugging Face acquisition follows a clear economic logic: open-source models operate at roughly 30% gross margins versus 70% for closed models like OpenAI or Anthropic. Lower margins for model providers mean more revenue flows to compute. For GPU sellers, promoting open weights directly expands the addressable market by keeping more dollars in the infrastructure layer rather than the application layer.
• •AI Cyber Risk is Immediate: The OpenAI-Hugging Face security breach involved 500–1,000 agents running persistently, combining intelligence with continuous operation to find and chain together system vulnerabilities undetected for weeks. Every CISO at a Fortune 500 company should treat this as a transition from bow-and-arrow attacks to missile-level threats. Open-source models are roughly six months behind frontier labs, giving organizations months, not years, to upgrade defenses.
• •Compound Startups Are Now the Baseline: AI coding tools have increased software output roughly 100x, forcing every startup to ship across the full product stack rather than owning a single point solution. Companies like Owner.com are building every feature a restaurant operator needs simultaneously. Startups that raise less capital, particularly European companies with $3M rounds, face structural disadvantage against US counterparts deploying tens of millions to compound product surface area at speed.
• •Agent-Friendly Products Win Procurement: Clay at $7B and Linear at $2.5B share a common trait: AI agents autonomously select them over competitors when executing GTM and engineering workflows. Agents test APIs, evaluate reliability, and default to products with clean integrations. This creates a durable moat because agent selection is merit-based and cannot be influenced by sales relationships, making early agent-layer adoption a compounding distribution advantage that is difficult for competitors to reverse.
• •Outcome-Based Pricing Reshapes Enterprise SaaS: Salesforce's commitment to multi-surface access via Claude MCP and its acquisition of Intercom signals a structural shift in enterprise software from seat-based to outcome-based pricing. Salesforce currently spends $300M annually on Anthropic, roughly 5% of its $6B engineering budget. For investors evaluating legacy SaaS, the key question is whether the system of record can deliver measurable outcomes, or whether agents built on top will capture that value instead.