The startup claims its technology can outperform leading AI systems of similar size on various benchmarks and is better suited for running on devices, reducing reliance on costly data centers. Liquid AI's CEO, Ramin Hasani, emphasizes that their technology offers an alternative to transformers, potentially overcoming the limitations faced by current AI developers. The funding will be used to scale their technology, which they believe can achieve more with fewer resources, addressing the growing debate about the cost-effectiveness of advanced AI systems.
Key takeaways:
- Liquid AI is raising $250 million in Series A funding at a $2.3 billion valuation, led by Advanced Micro Devices Inc. and other investors.
- The startup is developing AI systems inspired by the brain structure of the nematode worm Caenorhabditis elegans, offering a more flexible and efficient alternative to traditional neural networks.
- Liquid AI's technology requires less data and computing power than conventional transformer-based models, making it suitable for running on devices rather than relying on costly data centers.
- The company claims its text-based AI system has outperformed some leading alternatives of similar size on various benchmarks.