Amorphous AI, an Oxford-based health tech startup, has closed a $2.2 million pre-seed round led by Specialist VC and Calm/Storm Ventures. The financing, which included participation from PurposeTech, Z Fellows, BADideas.fund, and Outlast Fund, plus a roster of unnamed angel investors across the U.S. and Europe, will fund the company's push to convert messy clinical documentation into structured, searchable hospital records.
The startup, incorporated in the U.K. in January, uses standard medical terminology systems to make sense of the sprawling, handwritten and often cryptic notes that accumulate in patient files. The company declined to share its valuation or disclose whether it had taken prior funding. A LinkedIn post by team member Kialan Pillay in April mentioned a $1.9 million close, suggesting the round expanded by $300,000 before the formal announcement.
Two Tools, One Mission
At its core, Amorphous AI has built two products. The first, Data Engine, translates free-text hospital notes into structured records using a suite of medical classification standards: SNOMED CT, ATC, LOINC, ICD-10 and RxNorm. The second, Striata, functions as an AI-powered query layer that fields operational and clinical questions from hospital staff. The company claims Data Engine achieves 97.8% entity recall when extracting clinical information, though that figure comes from an internal technical whitepaper without a publication date.
SNOMED International, the nonprofit that oversees global clinical terminology standards, featured Striata on its website in August. According to SNOMED, the tool "ingests raw text, de-identifies it, and maps every concept to SNOMED CT in real time." The organization noted that Amorphous has begun early pilot programs, including one with Latvia's National Health Service, though the startup has yet to publish formal case studies or outcome data.
A Lean Team With European Roots
CEO Nikita Trojanskis, a Latvian national and the sole director listed in U.K. Companies House filings, co-founded Amorphous alongside CTO Dylan Sandfelder. The company lists at least four team members on LinkedIn, with total headcount estimated in the single digits. That puts the startup firmly in scrappy territory, a not-uncommon posture for a pre-seed company tackling the notoriously tangled challenge of hospital data infrastructure.
Calm/Storm Ventures, a Europe-focused healthtech fund that recently closed a new vehicle, co-led the investment. PurposeTech, an early-stage firm that backs health and climate startups, lists Amorphous in its portfolio. Outlast Fund, a Riga-based venture outfit, also participated.
A Crowded, Complex Market
Amorphous is entering a space that has seen a flurry of activity. Health Universe secured $6 million from Kleiner Perkins in March to build AI workflow agents for clinical operations, pointing to a Duke case study that claimed a 30-fold acceleration in trial setup times. Dandelion Health pulled in $14 million in May for a multimodal AI platform aimed at life sciences customers. MDClone launched a generative AI data assistant around the same time, while Datamonk, which focuses on imaging data migrations, raised €1.6 million late last year.
The regulatory landscape has shifted in ways that could help challengers like Amorphous gain traction. U.S. rules that took effect on January 1, 2023 now require certified electronic health record systems to offer standardized FHIR APIs, essentially forcing hospitals to open their data pipes to third-party developers. That change, outlined in a recent brief from the Office of the National Coordinator for Health IT, has lowered some of the historical barriers to interoperability.
Amorphous says it holds ISO 27001, ISO 27017 and ISO/IEC 42001 certifications, though the company has not made those certificates publicly available for independent verification. A few months back, it posted a hybrid Forward Deployed Engineer role in Riga, a signal that the team is preparing to scale operations closer to some of its early pilot sites.
Whether Amorphous can move from pilots to commercial contracts remains an open question. Hospital data is notoriously siloed, and the gap between a promising demo and a deployment that works across thousands of patient records is often wider than founders anticipate. For now, the startup has the capital to prove its technology can handle real-world clinical chaos.
