Autoscience, a San Mateo-based developer of autonomous AI systems for machine learning research, has raised $14M in seed funding led by General Catalyst. The company builds agents like Mira and Carl that handle ideation, experimentation, and even peer-reviewed paper production. The capital will fund expansion of its automated AI research lab, including managed services deploying hundreds of AI scientists for enterprise model improvements.
Agentic AI Funding Hits $220B
The round arrives amid record AI startup funding of $220B in January-February 2026 alone per eeNews Europe. Sakana AI raised $135M Series B in November 2025 at a $2.65B valuation, while Poolside AI secured $500M Series C in October 2025. Autoscience differentiates through enterprise-focused deployments and real-world proofs like Kaggle wins.
ML Papers Overwhelm Human Teams
Researchers face over 2,000 new machine learning papers weekly, creating an overload that outpaces human capacity for experimentation and validation according to BusinessWire. Human intuition struggles to keep up with the pace of discovery, bottlenecking applied teams in finance, manufacturing, and fraud detection. Current solutions rely on manual tuning or ensembling, limiting invention of novel architectures.
Autonomous Agents Invent New Models
Autoscience's Mira autonomously discovers new ML architectures, climbing to a Silver Medal in the $50k Kaggle Santa 2025 competition against 3,300 teams without human input per company blog. Carl analyzes papers, runs experiments, and generates full research papers—the first AI system to produce a peer-reviewed paper accepted to an ICLR 2025 workshop via RDWorldOnline. These agents enable machine-scale R&D, compressing years into months.
As Eliot Cowan, Co-Founder and CEO, noted:
"We’ve reached a point where human intuition is no longer enough…"
Investors Back Self-Improving AI
General Catalyst led with participation from Toyota Ventures, Perplexity Fund, MaC Ventures, and S32, signaling conviction in AI systems building better AI per investor sites. General Catalyst's AI bets like Mistral AI and Glean align with Autoscience's lab model, while Toyota Ventures adds deep tech expertise from autonomy plays. This mix provides strategic validation for scaling autonomous research in high-stakes enterprise apps.
Yuri Sagalov, Managing Director at General Catalyst, noted:
"…tackling an increasingly important challenge in machine learning: the pace and scalability of experimentation"
AutoML Market Scales to $61B
The automated machine learning market stands at $3.5B in 2024, projected to reach $61.23B by 2033 at 38% CAGR per Grand View Research. Autoscience targets this growth by automating the full research lifecycle, outpacing competitors focused on narrower automation. Structural shifts toward agentic AI, highlighted by Dario Amodei's focus on self-improving systems, drive capital into labs like Autoscience.
Google X Founders Lead Moonshot
Co-founders Eliot Cowan (CEO) and Charles Spirakis (CTO) bring elite pedigrees from Google X, where they built ML models featured in TIME's Best Inventions and used by the US Air Force and Swiss Re per LinkedIn profiles. Cowan led Carl to ICLR papers and Mira's Kaggle medal; Spirakis scaled geospatial ML at Bellwether. Founded in 2024 with six employees, their moonshot experience positions Autoscience to execute on automated discovery at scale.