Healthleap Raises $38M for an AI That Reads Records and Flags Risk

Healthleap announced on October 7, 2026 that it raised $38 million for a tightly scoped idea: an AI that reads electronic patient records and flags patients at risk of an illness nobody has diagnosed yet. The round combines an $8 million seed co-led by Sequoia Capital and First Round Capital with a $30 million Series A led by Hummingbird Ventures. The valuation was not disclosed.

Quick answer: what does Healthleap do?

The platform analyzes the history already sitting in a patient’s electronic record — labs, visits, medications, notes — and surfaces who deserves a closer look based on patterns consistent with an undiagnosed condition. It is not a diagnostic system: it is a triage system, and the decision stays with the clinician. The $38 million round was announced on October 7, 2026.

Why this framing works

Most AI-in-healthcare promises stumble at the same point: asking the model to diagnose. That is expensive to validate, hard to regulate and risky to get wrong. Healthleap inverts the problem — the model does not give the diagnosis, it orders the queue. It is the same kind of use that already worked in radiology: the machine does not read the scan for the doctor, it says which scan to read first.

The fact underpinning the business is old and well known: the information that would point to the illness is often already in the record, scattered across years of entries nobody has time to cross-reference. The product does not create new knowledge — it recovers what was already written.

The round in market context

Notice the contrast in scale. After a quarter dominated by hundreds of millions flowing into infrastructure, $16 to $38 million checks are reappearing for specific applications. It signals that money is walking back down the stack: from the model layer to the problem layer.

Why this matters

For founders, it is the week’s most useful read: there is room again for vertical products, with smaller checks and a clear thesis, after two years in which nearly all capital went to data centers and frontier models. And Healthleap’s framing is replicable in any sector that accumulates records it cannot cross-reference — legal, accounting, insurance, industrial maintenance.

The constraint is also replicable: the value depends entirely on how clean and how connected the underlying records are. In systems where databases do not talk to each other, the same product delivers a fraction of the result.

Frequently asked questions

How much did Healthleap raise?

$38 million in total: an $8 million seed co-led by Sequoia Capital and First Round Capital, and a $30 million Series A led by Hummingbird Ventures.

Does the AI make the diagnosis?

No. It analyzes the record and flags patients who merit closer evaluation. The clinical decision stays with the physician.

What was the valuation?

It was not disclosed.

Does this kind of product work outside healthcare?

The framing — cross-referencing old records to surface what went unnoticed — applies to any sector with accumulated history, such as legal, insurance and industrial maintenance.

At DigitalRadar, we follow where venture capital is actually betting. Stay on the radar.