Eyedentity's AI bet: opticians catch uveal melanoma before specialists

Eyedentity builds an AI diagnostics platform that analyzes fundus images taken during routine eye examinations to flag suspicious lesions and advise opticians whether to refer a patient to specialist care; its first application targets uveal melanoma.

That AI on fundus images already captured at routine optician visits can triage rare eye cancer and cut false referrals, with no new equipment or patient visits.

€1.3M (SEK 15M) oversubscribed first round, led by Norrsken Launcher with Karolinska Institutet Innovation and a group of angel investors, announced 2026-08-20.

Eyedentity was founded in 2025 in Stockholm by Gustav Stålhammar, an ocular oncologist at St. Erik Eye Hospital and professor at Karolinska Institutet, and Mats Holmström, a machine learning engineer who previously worked on automating cancer care at RaySearch Laboratories. They saw that uveal melanoma, the most common primary intraocular cancer in adults, can be hard to distinguish from benign lesions at an early stage - while opticians already image many eyes.

The startup joined the Karolinska Institutet Innovation Incubator and pitched at its InvestorDay, where it first connected with Norrsken Launcher. In August 2026 it closed an oversubscribed €1.3M round led by Norrsken Launcher, with Karolinska Institutet Innovation and angels participating, and is testing its platform in a live pilot with Synologen, a network of independent opticians in Sweden.

Live and pre-commercial as of 2026-09-02: Eyedentity is running its Synologen pilot and using the round to pursue regulatory approval for its first product, extend the platform to conditions such as diabetic retinopathy and glaucoma, and prepare for Nordic commercialization before expanding across Europe and the US.

Fundus images are already captured in volume at opticians, so the company avoided the cost and friction of building a separate screening infrastructure.

Uveal melanoma's early-stage resemblance to benign lesions makes triage the bottleneck, and AI is a natural fit for comparing large numbers of images against expert-labelled cases.

Placing the product with opticians rather than hospitals addresses a growing patient flow - people seeking help for vision problems and dry eyes - where suspicious findings appear first.

An oversubscribed first round led by impact-focused Norrsken Launcher, after a pitch at Karolinska's InvestorDay, validated the funnel from academic incubator to commercial capital.

A rare-disease product can still have a mass-market wedge: route screening through the opticians who already photograph eyes, making detection a byproduct of existing visits.

As of 2026-09-02 Eyedentity is a pre-revenue company with a live pilot at Synologen, no regulatory clearance yet for its first diagnostic module, and no announced later round. The near-term milestones are full regulatory approval, integration with electronic medical record systems so image analysis runs automatically, and expansion of the platform beyond uveal melanoma to diabetic retinopathy and glaucoma before moving from the Nordics into Europe and the US.