Liquid AI is the company to watch for teams that need AI running on-device rather than through a cloud API: think automotive assistants, offline mobile apps, or privacy-sensitive enterprise tools. With roughly 130 employees, it competes with Google's Gemma and Meta's Llama edge models rather than cloud-first labs like OpenAI or Anthropic.
Liquid AI, founded in 2023 by four MIT CSAIL researchers, builds Liquid Foundation Models (LFM), device-native AI that runs on phones, vehicles, and edge hardware instead of the cloud. Its models have been downloaded more than 38 million times on Hugging Face and now ship inside Mercedes-Benz's in-car assistant.
Founded: 2023 · HQ: Cambridge, MA, USA · Team: 100-150 · CEO: Ramin Hasani · Funding: $297M total raised across 2 rounds (lead: AMD Ventures, also Andreessen Horowitz, Capgemini Ventures, Breyer Capital); latest Series A $250M closed Dec 13, 2024 · Valuation: $2.35B (Series A, Dec 2024)
About Liquid AI
Liquid AI was spun out of the Massachusetts Institute of Technology's Computer Science and Artificial Intelligence Laboratory (MIT CSAIL) in 2023 by four researchers: Ramin Hasani (CEO), Mathias Lechner (CTO), Alexander Amini (Chief Science Officer), and Daniela Rus, the longtime director of MIT CSAIL. The founders' academic work on liquid time-constant neural networks, continuous-time architectures that adapt their computation to the input rather than using a fixed number of layers, became the technical basis for the company's Liquid Foundation Models (LFM). Liquid AI is headquartered at 314 Main Street in Cambridge, Massachusetts, inside MIT's Kendall Square research corridor. The founding thesis was that general-purpose AI does not have to live only in the cloud: a model built for efficiency from the architecture up can run directly on a phone, car, or industrial sensor without a network round trip. The company's product line is the LFM (Liquid Foundation Model) family, which by mid-2026 spanned more than 50 shipped models and over 3,100 variants, from 230-million-parameter encoders up to a 24-billion-parameter mixture-of-experts model, all distributed primarily through Hugging Face. The current generation, LFM2.5, covers text models (LFM2.5-350M, LFM2.5-1.2B-Instruct, LFM2.5-2.6B, LFM2.5-8B-A1B), vision-language models (LFM2.5-VL-450M, LFM2.5-VL-1.6B, LFM2.5-VL-3B), an audio model (LFM2.5-Audio-1.5B), and retrieval-focused nano models (LFM2.5-Embedding-350M, LFM2.5-ColBERT-350M). Alongside the models, Liquid AI ships LEAP, a software development kit for fine-tuning and deploying LFMs across llama.cpp, MLX, ONNX, CoreML, SGLang, and vLLM runtimes, and Apollo, its hosted access layer for teams that want the same models without managing on-device runtimes themselves. Liquid AI has shipped a new LFM2.5 model roughly every month since March 2026: LFM2.5-350M in March, the vision-language LFM2.5-VL-450M in April, the mixture-of-experts LFM2.5-8B-A1B in May, the flagship LFM2.5-2.6B agent model on August 4, and the vision-language LFM2.5-VL-3B on August 12, the same week this profile was written. On the partnership side, Mercedes-Benz announced in April 2026 that it would embed LFMs directly into third- and fourth-generation MBUX infotainment systems for North American vehicles starting in the second half of 2026, and Alef Education said in January 2026 that it was deploying Liquid AI's on-device models across a platform used by more than 1.5 million students in 14,000 K-12 schools. Liquid AI has raised $297M across two rounds from 22 disclosed investors. The most recent, a $250M Series A that closed December 13, 2024, was led by AMD Ventures with participation from Andreessen Horowitz, Capgemini Ventures, and Breyer Capital, and valued the company at roughly $2.35B, making it a unicorn a little over a year after founding. As of August 2026, Liquid AI has not disclosed a newer round. Revenue comes from the commercial side of the LFM Open License (LFM1.0): the models themselves, built on an Apache 2.0 base, are free to download, fine-tune, and deploy commercially for any company under $10M in annual revenue, and Liquid AI charges companies above that threshold for a commercial license, plus custom architecture work, OEM integration (as with Mercedes-Benz), and on-premises deployment support with SLAs. Liquid AI has not published official revenue figures; third-party estimator Getlatka put 2025 ARR at approximately $13.2M, a figure Liquid AI has not confirmed. CEO Ramin Hasani and CTO Mathias Lechner both came out of Daniela Rus's lab at MIT CSAIL, where Hasani's doctoral research on liquid time-constant networks originated; Alexander Amini, the company's Chief Science Officer, was a CSAIL research scientist on the same liquid-networks work. Rus remains a co-founder and director while continuing to lead CSAIL. Third-party trackers put Liquid AI's headcount at roughly 100 to 150 people as of mid-2026 (Tracxn cites 130 as of June 30, 2026), spread across its Cambridge headquarters and a smaller Boston-area presence. Liquid AI describes its mission as building efficient, general-purpose AI at every scale, with products that are compute- and cost-optimized enough to bring intelligence to any device. The company frames its research philosophy as white-box explainability, favoring architectures whose behavior can be inspected over black-box shortcuts, and publishes a model card for every LFM release on Hugging Face with benchmark tables, intended use, and stated limitations. Liquid AI has not published a formal responsible-scaling policy or named third-party red-teaming partners, in contrast to frontier labs like Anthropic or OpenAI. Liquid AI competes for the on-device and edge tier of the model market against Google's Gemma family, Meta's Llama edge and mobile variants, and Microsoft's Phi series, rather than against cloud-first frontier labs. Its differentiator is architecture: LFMs mix convolution and attention rather than using a pure Transformer stack, which Liquid AI says lets a 3B-parameter vision-language model like LFM2.5-VL-3B run in about 3GB of memory at 228 tokens per second on an Apple M5 Max. The tradeoff is capability ceiling: Liquid AI has no flagship model built to compete with GPT-5, Claude, or Gemini on frontier reasoning benchmarks, and its own LFM2.5-VL-3B model card recommends against using it for long-context or reasoning-heavy tasks. The LFM lineup has passed 38 million downloads on Hugging Face, a proxy for developer adoption relative to other small-model providers. Because LFMs run locally rather than through a hosted API by default, inference does not require sending user data to Liquid AI's servers, which the company markets as a privacy-by-architecture advantage over cloud-only competitors. Liquid AI has not published SOC 2, ISO 27001, or GDPR compliance certifications, and has not disclosed an EU AI Act classification as of August 2026. The Mercedes-Benz MBUX rollout, scheduled for the second half of 2026, is Liquid AI's first large-scale automotive production deployment and the clearest signal of whether its edge-AI thesis can scale beyond developer downloads into shipped consumer hardware. The company's release cadence, a new LFM2.5 model roughly every month through 2026, suggests it intends to keep competing on frontier-efficiency (best capability per gigabyte) rather than frontier-scale.
Mission
Build efficient, general-purpose AI at every scale, so intelligence runs on any device without a network round trip.
Products
- Liquid Foundation Models (LFM) (Foundation model family): https://www.liquid.ai/models
- LFM2.5-VL-3B (Vision-language model): https://huggingface.co/LiquidAI/LFM2.5-VL-3B
- LEAP (Edge fine-tuning and deployment SDK): https://www.liquid.ai/leap
- Apollo (Hosted model access platform): https://www.liquid.ai/apollo
Liquid AI Models on HokAI
Links
Frequently Asked Questions
How much funding has Liquid AI raised?
Liquid AI has raised $297M total across two rounds. Its Series A closed on December 13, 2024, at $250M led by AMD Ventures, with Andreessen Horowitz, Capgemini Ventures, and Breyer Capital also participating, valuing the company at roughly $2.35B. As of August 2026, no newer funding round has been publicly disclosed.
What products does Liquid AI make in 2026?
Liquid AI's product is the LFM (Liquid Foundation Model) family: more than 50 device-native models, most recently the vision-language LFM2.5-VL-3B released August 12, 2026 for on-device document and screen understanding. The LEAP SDK handles fine-tuning and deployment across llama.cpp, MLX, ONNX, and vLLM runtimes, and Apollo provides hosted access to the same models for teams that don't want to manage on-device runtimes themselves.
How does Liquid AI handle data privacy and compliance?
As of August 2026, Liquid AI has no publicly listed SOC 2, ISO 27001, or GDPR certification. Its main privacy argument is architectural: because inference happens on the user's own hardware rather than a hosted API, LFMs never need to transmit prompts or documents back to a Liquid AI server. Enterprise customers needing on-premises or air-gapped deployment can arrange it directly through Liquid AI's commercial licensing team.
Who are Liquid AI's main competitors?
In the on-device tier, Liquid AI's closest rivals are Google (Gemma), Meta (Llama's edge and mobile variants), and Microsoft (Phi); all four ship small, efficient models meant to run outside the cloud. Liquid AI's edge is architecture, a convolution-attention hybrid tuned specifically for latency and memory rather than a scaled-down Transformer, but it loses on raw capability to cloud-first frontier labs like OpenAI, Anthropic, and Google DeepMind for any task that needs large-context reasoning.
How do you start using Liquid AI's models?
The fastest path is Hugging Face: download any LFM model under the LiquidAI organization and run it locally with llama.cpp, MLX, or ONNX, or try the WebGPU browser demo with no install at all. Teams that want fine-tuning or production deployment tooling should install the LEAP SDK; companies over the $10M revenue threshold in the LFM Open License need to contact Liquid AI's sales team for a commercial license before shipping a paid product.
Top Alternatives
- Google DeepMind: Pick Liquid AI if you need a model running locally on a phone or car; pick Google DeepMind's Gemma for tighter integration with Google Cloud and Android.
- Meta AI: Pick Liquid AI for smaller, latency-tuned edge models; pick Meta AI's Llama for a larger open-weights ecosystem and community tooling.
- Microsoft: Pick Liquid AI for on-device deployment across arbitrary hardware; pick Microsoft's Phi for tight Azure and Windows Copilot integration.