Kazakhstani medtech startup Cerebra AI has signed a five-year agreement with a major US healthcare group. Its partner has a history spanning more than 150 years and annual revenue exceeding $20 billion.
The healthcare group’s name and the financial terms have not been disclosed: the parties are bound by a nondisclosure agreement. What is known is that the US partner will commit resources to developing Cerebra AI’s platform and help extend it beyond stroke diagnosis. Cardiovascular diseases and different types of cancer are expected to be the next areas of focus. Digital Business reported the agreement.
The partner is large, but its name remains unknown
The agreement runs for five years. That is a substantial time frame for a medical startup: a single pilot is not enough to integrate a new technology into a large hospital system.
It is not yet known how much money Cerebra AI will receive, how many clinics will take part in the project or when the first deployments will begin. The figures — 150 years of history and more than $20 billion in revenue — invite speculation about the partner, but there is no reliable basis for naming a particular healthcare group.
For now, it is more accurate to describe this as a strategic agreement than a full commercial launch in US hospitals.
Cerebra wants to move well beyond stroke
Cerebra AI has developed a system that analyzes standard brain CT scans and helps doctors detect signs of ischemic or hemorrhagic stroke more quickly.
In stroke care, speed directly affects a patient’s outcome. According to the American Stroke Association, an untreated stroke can result in the loss of around 1.9 million brain cells every minute. Even a few minutes saved may therefore matter for a patient’s recovery.
On its website, Cerebra AI says more than 700 doctors at 50 hospitals in three countries use its technology, and that the company’s dataset covers more than 150,000 patients. These are the startup’s own figures and have not been independently verified.
The new agreement is intended to help apply the same approach to other diseases. This remains a development effort: the company has not yet presented public results for cancer or cardiovascular disease diagnosis.
Why Cerebra does not want to put ChatGPT next to a CT scanner
Cerebra AI CEO Doszhan Zhusupov believes that medical AI does not necessarily have to take the form of a large language model.
Large models can process text and answer a wide range of questions, but they can make mistakes, invent facts and struggle to explain how they reached a conclusion. That is not enough for a clinical setting.
Cerebra is taking a different route: a small, specialized model performs one specific task at a defined point in a clinical workflow. The doctor should be able to see which part of a scan the algorithm has highlighted and why it deserves attention.
According to Zhusupov, this approach has already demonstrated its usefulness in stroke care. The startup will now try to apply it to other diseases.
The US market is not yet open to Cerebra
A five-year agreement does not mean the product has already been cleared for use in diagnosing patients in the United States.
Cerebra AI states on its official website that its software has not yet been approved in the US. According to the company, it has completed two preparatory stages of the certification process. One more clinical study is needed before it can submit an application. The team plans to seek FDA approval in 2027.
The FDA reviews medical AI systems as software-based medical devices. Different products follow different regulatory pathways, including 510(k), De Novo and premarket approval. The regulator considers more than an algorithm’s accuracy: it also examines patient risks, data quality and how the system performs after deployment.
Cerebra has secured a US partner, but its main regulatory test is still ahead.
The move into the US took several years
Cerebra AI grew out of a laboratory founded by Doszhan Zhusupov in Almaty in 2018. Its first prototype appeared in 2020, followed by trials in hospitals.
Timur Turlov invested $1 million in the project in 2021. In 2023, the startup raised another $4.1 million in a Seed 2 round. Investors included Big Sky Capital, Most Ventures, Aloca Ventures, Uzcard Ventures, UMAY Angels and Ynta Group. Part of that round used a pre-money valuation of $12 million.
In 2024, the company moved its headquarters to Palo Alto and began preparing for US certification. The new agreement is therefore a continuation of a strategy the team has pursued for several years.
RAEM has previously examined Cerebra AI’s journey in detail — from an Almaty laboratory to clinical pilots and its entry into the US market.
The partner’s name is no longer the main question
The US healthcare group’s name will probably become public sooner or later. More consequential for Cerebra AI are the results of its next clinical study, its planned FDA submission in 2027 and its first real deployments in US hospitals.
Those milestones will show whether the five-year agreement becomes a substantial international business — and whether the Kazakhstani technology can move beyond stroke diagnosis.
Readers can follow the company’s progress in its RAEM Startup Radar profile.
