Prime Highlights
- Former Cohere research leader Sara Hooker has launched Adaption Labs with a focus on building smarter systems that use computing resources more efficiently.
- The startup aims to reduce costs by creating technology that adapts in real time, avoiding heavy retraining and complex setup processes.
Key Facts
- Adaption Labs has secured $50 million in seed funding, led by Emergence Capital Partners, with support from several global investors.
- The San Francisco–based company was co-founded by Sara Hooker and Sudip Roy, both former senior executives at Cohere.
Background:
San Francisco-based startup Adaption Labs has raised $50 million in seed funding to support its plan to change how technology models are built and used. The company was founded by Sara Hooker, former vice president of research at AI firm Cohere and a veteran researcher at Google DeepMind.
The funding round was led by Emergence Capital Partners, with backing from Mozilla Ventures, Fifty Years, Threshold Ventures, Alpha Intelligence Capital, E14 Fund, and Neo. Adaption Labs declined to disclose its valuation following the raise.
Hooker launched Adaption Labs with the goal of developing AI systems that are smaller, more efficient, and capable of adapting to tasks in real time without the heavy computing costs associated with today’s largest language models. She is joined by cofounder Sudip Roy, who previously served as director of inference computing at Cohere.
Adaption Labs is taking a different path from many major AI companies that focus on building ever-larger models using huge amounts of data. Instead, the startup is working on systems that can adjust to tasks as they are being used, cutting down the need for costly retraining, fine-tuning, or complicated prompt setups.
Hooker has been outspoken against the idea that bigger models are always better. She believes that as costs continue to rise, simply increasing model size is no longer delivering strong results. She believes the next phase of progress will come from systems that use computing resources more efficiently, adjusting their effort based on how complex a task is, instead of depending on sheer size.
Adaption Labs is working on three key areas. First, it is building systems that can create and adjust the information they need in real time. Second, it is developing technology that uses only the amount of computing power needed for each task. Third, it is designing systems that improve over time by learning from how users interact with them.
One of the company’s main areas of research looks at ways to change how systems respond while they are being used, without needing expensive retraining. This approach helps systems adjust instantly while keeping operating costs low.
With costs and energy use rising across the industry, Adaption Labs is positioning itself as part of a new wave of companies working on more practical and sustainable technology solutions.
