AI startup Liquid AI on Friday closed a $250 million early-stage funding round which was led by its strategic partner Advanced Micro Devices (AMD).
According to a Bloomberg article , the round now values Liquid AI at over $2 billion. OSS Capital, Duke Capital Partners, and PagsGroup also participated in the series A round. Liquid AI previously raised about $46.6 million in seed funding in 2023.
Liquid AI wants to develop more efficient models
The development comes as consumer generative AI, which captured global attention following the launch of OpenAI’s ChatGPT in November 2022 to immediate success has spurred developments across the world. The tech is now at the center of the fast-growing AI sector, which has seen investors pour billions of dollars into generative AI startups.
For Liquid, the startup’s new models are known as Liquid Foundation Models (LFMs) which the company says can deliver performance comparable or superior to traditional large language models (LLMs) like those from OpenAI, Google Cloud, and AWS.
The LFMs are lightweight, general-purpose AI models requiring less data and compute power. The models strangely uses mathematical approaches discovered by studying a worm’s brain, as per Crunchbase.
According to SiliconAngle , Liquid’s models are based on the concept of “liquid neural networks.” This classification is different from generative pre-trained transformer-based models or GPTs that support popular chatbots like ChatGPT and Google LLC’s Gemini.
“At Liquid, our mission is to build the most capable and efficient AI systems at every scale.”
Liquid CEO Ramin Hasani.
“Since our inception, we have demonstrated the scalability of our technology by releasing text-based models, unveiling multimodal LFMs, and collaborating with key partners to showcase real-world impact,” added Hasani.
AMD partnership will allow Liquid AI to optimize its LFMs
According to SiliconAngle, the AI startup offers three model sizes, that is LFM-1B, a dense 1.3 billion parameter model for on-device such as mobile phones, LFM-3B, a 3.1 billion parameter model optimized for edge developments, and the LFM-40B Mixture of Experts, designed to tackle complex tasks.
Liquid AI said LFMs have a reduced memory footprint compared to GPT architecture, which is more substantial for lengthy inputs, for instance when processing extremely long documents or video.
The LFM-3.1B model, according to Liquid AI used significantly less memory than other small models when deployed than models such as Google LLC’s Gemma 2 2B-it or Meta’s Llama3.2 3B.
Now, with this funding, Liquid AI has joined a strategic partnership with AMD bolstering its ability to optimize its foundation models with the chipmaker’s graphic, central, and neural processing units.
The funding will also help the Cambridge, Massachusetts-based startup to develop tailored LFMs for industries like consumer electronics, biotechnology, telecoms, financial services, and e-commerce.
Liquid AI was founded in 2023.
From Zero to Web3 Pro: Your 90-Day Career Launch Plan
