CAMBRIDGE, Mass., Nov. 13, 2025 (Globe Newswire) — Liquid AI today announced a multifaceted partnership with Shopify to license and deploy Liquid AI’s flagship product, Liquid Foundation Model (LFM), across quality-sensitive workflows on the Shopify platform, including search and other multimodal use cases where quality and latency are critical. The first production deployment is a sub-20ms text model that powers search. The deal follows Shopify’s participation in Liquid AI’s $250 million Series A round in December 2024 and formalizes the in-depth collaboration already underway between the two companies.
As part of the partnership, Shopify and Liquid have jointly developed a generative recommender system powered by the new HSTU architecture. In controlled tests, this model was proven to perform better than previous stacks, resulting in higher conversion rates from recommendations.
Ramin Hassani, Liquid AI CEO:
“Recommendations are the backbone of decision-making in finance, healthcare, and e-commerce. To be useful in the real world, models need to be reliable, efficient, and fast. Shopify is the ideal partner to validate that at scale. We’re excited to bring the Liquid Foundation model to millions of shoppers and merchants and show how efficient ML translates into measurable value in everyday experiences.”
Liquid’s LFM is designed for sub-20ms multimodal quality-preserving inference. For certain tasks like production, LFM with up to 50% fewer parameters outperforms popular open source models such as Qwen3, Gemma3, and Llama 3 while delivering 2-10x faster inference and enabling real-time shopping experiences at platform scale.
Mikhail Parakin, Shopify CTO:
“I’ve seen many models, and no one else offers sub-20ms inference on real-world workloads like this. Liquid’s architecture is efficient without sacrificing quality. Depending on the use case, models with up to 50% fewer parameters can outperform Alibaba Qwen or Google Gemma and still run 2-10x faster. This is what we need to strengthen commerce.”
Mathias Lechner, Liquid AI CTO:
“We are designing liquid foundation models with intertwined objective functions that maximize quality while making the system the fastest on the market on selected hardware. This makes them a natural fit for e-commerce applications such as personalized rankings, search enhancement generation, and session-aware recommendations under tight latency and cost budgets to provide the best user experience.Shopify environment, we have focused on production robustness and low dispersion tail delay up to drift monitoring.”
The partnership includes LFM’s versatile license across low-latency, quality-focused Shopify workloads, continued research and development collaboration, and a shared roadmap. The current deployment is a sub-20ms text model for search, but the companies are evaluating multimodal models for additional products and use cases such as customer profiles, agents, and product classification. Financial terms were not disclosed.
About Liquid AI
Liquid AI builds Liquid Foundation Models (LFMs), multimodal and efficient models designed for applications where real-time reliability is critical. Founded by Liquid Neural Network researchers, Liquid AI focuses on low-latency inference, resource efficiency, and production-grade safety.
press contact
Rachel Gordon
rg@liquid.ai
A photo accompanying this announcement is available at https://www.globenewswire.com/NewsRoom/AttachmentNg/1a6bcd9c-934a-4db5-9dcd-f64d86f82fe6.
