Lila's $350M Series A funds AI-run science labs

Lila Sciences, a startup with the audacious goal of building “scientific superintelligence,” has closed a massive $350 million Series A, bringing its total capital raised to an eye-watering $550 million. The round, which the company announced today, includes a strategic investment from NVentures, NVIDIA’s venture capital arm, signaling a deep connection between Lila’s ambitions and the world of accelerated computing.

The funding isn’t just for another AI model trained on existing scientific papers. Instead, Lila is channeling the capital into building out its “AI Science Factories.” This is the core of its pitch: creating labs where AI agents can design and execute their own physical experiments, generating novel data that doesn’t exist in any textbook or database. It’s a bet that the next major breakthrough in AI will come from giving it a physical lab to play in.

Beyond the dataset

Most of today’s AI is incredibly good at pattern recognition within existing information. It can predict protein structures or summarize literature because it’s been trained on immense datasets of what’s already known. Lila argues that true discovery happens beyond the edge of curated data. The company aims to reinvent the scientific method by creating a closed loop where AI can hypothesize, test its ideas in the real world via automated instruments, analyze the results, and learn—at a scale thousands of times faster than a human team.

Lila believes that, much like large language models, scientific discovery is governed by scaling laws. The more experiments an AI can run, the more likely it is to uncover emergent properties and novel discoveries in areas like materials science, chemistry, and life sciences.

The investor list for the Lila Series A reads like a who’s who of strategic capital. Beyond NVIDIA, the round includes IQT, the intelligence community’s venture arm, pointing to a strong interest in national security applications for materials and energy. Analog Devices also participated, highlighting the critical role of advanced sensors and hardware in building these AI-driven labs.

They join an existing roster that includes Flagship Pioneering (where Lila was created), General Catalyst, and a subsidiary of the Abu Dhabi Investment Authority (ADIA). According to the company, this syndicate represents over $2 trillion in assets under management.

With the new funding, Lila plans to scale up its AI Science Factories and begin opening its platform to its first cohort of commercial partners. It’s a monumental bet that the future of science isn’t just about AI that can read, but AI that can *do*.