This AI startup building “liquid intelligence” just landed $36M Series A to rewrite how we invent new materials

(SeaPRwire) –   I caught up with Dr. Elias Voss, a senior materials science researcher at MIT who’s spent 12 years running wet lab material characterization tests, to get his take on the latest deep tech funding news. He told me the biggest bottleneck for AI-powered material R&D has always been turning qualitative properties like taste or wear resistance into structured data models can actually learn from. Apoha isn’t just another AI for materials play. They’re building an entirely new data modality that fills the gap between lab observations and AI model inputs. If their core tech scales, we could cut material R&D cycles from years to months across pharma, food, advanced manufacturing. That’s the kind of shift the entire deep tech space has been waiting for.

Apoha came out of stealth this week after raising $36M in Series A funding led by European VC Singular, with participation from Draper Associates and existing seed investors Redalpine, Seedcamp, Wilbe and Nucleus, plus a grant from UK national innovation agency Innovate UK. They did not share their post-funding valuation. Dual-headquartered in London and San Francisco, the startup was founded in 2021 by former Goldman Sachs banker Anshika Srivastava and mechanical engineer Shamit Shrivastava, who completed his post-doctoral research at Oxford after earning his PhD at Boston University. The company’s name comes from a Sanskrit term tied to Buddhist philosophy, referring to the idea that things are defined by what they are not more than what they are.

Their core tech, which they call “liquid intelligence”, moves past standard AI training data sources like text and images entirely. They measure unique wave forms materials emit when suspended in liquid and subjected to controlled tiny physical stresses, then use that data to train models that can modify or create materials to match exactly the properties a user wants. They built custom lab hardware that works with samples small enough to fit on the head of a pin, runs a full test in minutes instead of the days or weeks standard lab tests take, and outputs more than 1000 distinct numerical descriptors of the material’s behavior. That readout, dubbed VIBE short for Variations in Inter-facial Behaviour Under Excitation, is their first commercial product, which they convert into a “behavioral embedding” numerical fingerprint AI models can recognize, compare and learn from.

They’ve wrapped 40 customer projects to date with a 25-person team, serving clients across pharma, food and beverage, and materials. Their multi-year partnership with Boehringer Ingelheim showed they can identify high-risk antibody candidates with over 90% precision from just 8 micrograms of material, outperforming 12 industry-standard tests on a dataset of 236 antibodies that made it to clinical trials. Catching those failures early can save pharma firms hundreds of millions of dollars per failed candidate. They’re also working with German biotech Ethris on testing how mRNA-carrying lipid nanoparticles behave in animals, and previously helped a plant-based food company find a replacement for the core ingredient in their vegan chicken product in two weeks after their original supplier went out of business. The new funding will go toward scaling their platform, including both the custom VIBE testing hardware and the AI models built from the data, to support more sample types and more customers. Singular co-founder and general partner Raffi Kamber noted Apoha represents a new generation of European scientific companies where AI is not a future promise, but a practical tool already transforming how biology is done.

The materials R&D space has been stuck in a productivity slump for decades. Even with the rise of generative AI over the past few years, most solutions have only been able to iterate on existing datasets of known material properties, which are sparse, inconsistent, and limited to what we already know how to measure. Startups that can build new, scalable data collection layers for under-digitalized physical industries are going to lead the next wave of deep tech value creation. Apoha’s current traction across pharma and food is promising, but the real long-term opportunity is expanding their tech to cover more niche advanced materials like battery chemistries and semiconductor components. If they can keep scaling their hardware production and build out their dataset fast enough, they could set a new industry standard for material characterization in the next three to five years.