London-based artificial intelligence company Apoha has officially launched after operating in stealth mode since its founding in 2021, announcing both its commercial debut and the completion of a $36 million fundraising round. The company has developed a technology platform designed to help machines interpret the physical behavior of materials in ways that mimic human sensory capabilities such as touch, taste, and smell.
Founded by former University of Oxford postdoctoral researcher and mechanical engineer Shamit Shrivastava and former Goldman Sachs executive Anshika Srivastava, Apoha has built a proprietary system that combines robotic laboratory automation with machine learning to generate and analyze a new category of structured data describing how matter behaves under different conditions.
The company’s process begins by placing a small sample of material—ranging from proteins and lipids to soil—into a liquid environment. Controlled stresses are then applied to the sample while specialized equipment captures resulting wave pattern data. According to Apoha, its machine learning models can convert this information into more than 1,000 behavioral descriptors within minutes.
These descriptors are subsequently analyzed by a second AI system that evaluates how materials may perform in real-world applications. The approach differs from conventional testing methods, which typically assess individual properties separately.
Describing the concept, Shamit Shrivastava said, “it’s like a new kind of Raman,” referencing Raman spectroscopy, a widely used analytical technique that studies how light scatters when directed at a substance.
The company refers to its platform as Liquid State Intelligence (LSI) and reports that the technology is protected by more than 60 patents covering its hardware, software, datasets, and AI models.
Reflecting on the development effort behind the platform, Shrivastava stated, “The technology took 15 years of science and 5 years of company-building to bring to life.” He added, “There is no shortcut to this data class—it cannot be scraped from the internet, synthesised, or retrofitted from existing assays.”
Apoha is already working with organizations in pharmaceuticals, food and beverage production, and materials science. The company cited collaborations with Boehringer Ingelheim and Somru BioScience among its early industry partnerships.
According to Apoha, Boehringer Ingelheim has used the platform to identify antibody candidates using as little as 8 micrograms of material. The company also highlighted findings from a preprint study coauthored by Apoha researchers and Boehringer principal scientist Maureen Crames, which reported that LSI successfully identified antibody-drug candidates with multiple biophysical liabilities and outperformed 12 conventional pharmaceutical assessment tests used to evaluate candidate behavior under real-world conditions.
Click here for the original news story.
