Aidoc’s 2018 brain-bleed AI now drafts full radiology reports

At a Future of Industry conference panel in Tel Aviv, Aidoc VP of AI Dr. Ronny Levanda walked through a scenario that sounds like a product launch: a patient arrives at the ER with a severe headache, gets a CT scan, and software flags an undiagnosed brain hemorrhage, prioritizes the case, and alerts physicians before a radiologist has opened the image. As reported by Calcalist’s CTech, the panel put the system in more than 2,000 hospitals, reading over 60 million scans a year.

The detail worth holding onto: none of this is new. Aidoc’s intracranial hemorrhage triage received FDA clearance back in 2018, billed then as radiology’s first cleared AI triage-and-notification tool. The panel was a snapshot of a mature deployment, not a debut, and a useful marker of how deep clinical AI now sits inside hospital workflows. The actual news is one layer up, in what Aidoc is building on top of that eight-year-old pipeline.

What’s actually new

Levanda said Aidoc’s latest models no longer look for a single condition. They analyze an entire scan, recognize dozens of abnormalities at once, and generate a comprehensive radiology report. That is the company’s CARE foundation model (Clinical AI Reasoning Engine), which earlier in 2026 received what Aidoc calls the first FDA clearance for a comprehensive foundation-model-based triage system in clinical imaging, plus an FDA Breakthrough Device Designation for report drafting.

The capital backs the ambition. On April 29, 2026, Aidoc closed a $150 million Series E led by Growth Equity at Goldman Sachs Alternatives, with General Catalyst, SoftBank Vision Fund 2, and NVentures (NVIDIA’s venture arm). That pushed total funding above $500 million, less than a year after a separate $150 million round in 2025. CEO Elad Walach’s stated goal is end-to-end AI across CT and X-ray, “pixel to draft report,” within two years, citing diagnostic error that contributes to at least 400,000 US deaths a year (a BMJ Quality & Safety figure).

One caveat on the framing: Calcalist’s subhead describes the system detecting bleeds “in seconds.” That phrasing is the outlet’s, not a measured per-scan time disclosed on the panel. What Aidoc actually describes is analysis in near-real time as scans are generated.

"The system detects bleeding in the brain before the doctor even sees it."

Healthcare and tech leaders say AI is transforming medicine, from identifying life-threatening conditions in seconds to giving ALS patients new ways to communicate.https://t.co/2EatGNLx35

— CTech (@Calcalistech) July 10, 2026

What this means for teams buying clinical imaging AI

The buying question is shifting. For most of the last eight years, the pitch was single-condition triage: does the tool catch the bleed and cut the time to treatment. CARE reframes it as a platform question: does the tool read the whole scan and draft the report the radiologist signs. That is the same move from point solution to workflow layer now playing out in AI medical scribes, where drafting, not detection, is the value.

For health systems, the practical read is consolidation. A hospital already running Aidoc for hemorrhage triage has a paved path to expand into full-scan analysis without swapping vendors, which raises switching costs and makes the incumbent hard to dislodge. The adoption numbers already inside the clinic suggest that expansion, not first purchase, is where the next round of budget goes. Buyers should separate what is FDA-cleared (the triage) from what carries only a Breakthrough Designation (report drafting), because the two sit at very different points on the regulatory and liability curve.

What to watch

Three markers. Whether report drafting converts from Breakthrough Designation to full clearance, and how fast. Whether Walach’s two-year “pixel to draft report” timeline survives contact with regulators and radiologist oversight. And whether a foundation model reading entire scans holds accuracy across conditions the way a narrow bleed detector did, a question that, as with AI in drug discovery, only peer-reviewed evidence will settle.