Scaled Cognition, a reliability-focused lab that develops the Agentic Pretrained Transformer model, raised a $100M Series A led by Khosla at a $750M valuation
AI models can be ‘like schizophrenic geniuses,’ says CEO who raised $100 million in round led by Khosla Ventures
Context & Ripple Effects
Khosla Ventures has repeatedly backed AI infrastructure and application-layer companies in the coverage, from Scale AI’s data operations to Factory’s coding agents and Symbolica’s alternative foundation-model approach. Its lead role here extends that pattern to a lab centered on model reliability.
The round also arrives as Khosla is associated in the relationship data with deploying AI into mature operating businesses. Reliability is consequential in that context because agentic systems must perform consistently before they can be used to automate business workflows at scale.
First-order effects
- Scaled Cognition gains $100M to develop and commercialize its Agentic Pretrained Transformer model, with Khosla’s backing validating reliability as a distinct AI-model investment thesis.
- Khosla adds another position across the AI stack, alongside investments connected to data infrastructure, alternative model architectures, and agent-based software.
Second-order effects
- Other model and agent developers face added pressure to demonstrate dependable task execution, not only raw model capability, when seeking enterprise adoption and venture funding.
- Companies building AI automation for operational use cases may gain another potential model supplier, increasing interest in architectures and evaluation methods designed around reliable agent behavior.
Third-order effects
- If reliability-focused model labs can convert funding into production performance, differentiation in AI may shift from access to general-purpose models toward dependable orchestration and execution in specific workflows.
- The pattern supports a more vertically integrated AI investment model: investors back core models and agents while also seeking operating businesses where automation can be deployed, though the corpus does not establish whether these investments will be connected commercially.
The trend: AI investment is broadening from general model scale toward the reliability, routing, and deployment layers needed to make agents useful in real business workflows.