A robot capable of drawing blood could become the most significant AI health development of 2026.

      One minute and 49 seconds; that was the average time a machine needed to scan a patient's arm, select a vein, insert a needle, fill the tubes, and retract, during the Dutch clinical trial that advanced it toward regulatory approval. A phlebotomist, working meticulously, typically takes about five minutes, according to the company’s estimates. The moment itself was not dramatic; it was merely a patient rolling their sleeve back down and leaving.

      The machine is named Aletta, developed by Vitestro, a company established in Utrecht in 2017 that has focused its entire existence on a problem that the broader medical field considers already resolved. TNW first reported on the company in 2024 when it secured €20 million in funding and appeared, frankly, as a novelty.

      On August 19, 2026, the US Food and Drug Administration approved Aletta through the De Novo pathway, making it the first independent robotic device authorized to draw blood without a human hand on the needle. It is permitted for adults in outpatient environments, under supervision.

      The most significant piece of medical AI to emerge this decade may turn out to be the one that performs the least remarkable task. For three years, the industry has promised that large models would discover new drugs, interpret scans more effectively than radiologists, and reshape diagnoses entirely, yet it has mostly delivered chatbots, a disparity we have extensively covered.

      Meanwhile, a small team from the Netherlands directed a robot at the most mundane procedure in medicine, conducted trials for years, and surpassed a standard that few in consumer AI have ever been challenged to meet.

      The term "ordinary" is key here, carrying substantial weight. “This authorization reflects the FDA’s commitment to advancing innovative medical devices that address a critical public health need while ensuring the safety and efficacy that patients deserve,” stated Michelle Tarver, who heads the FDA’s Center for Devices and Radiological Health, during the announcement of the authorization.

      Blood draws are among the most frequently performed medical procedures in the United States, but patients may encounter delays due to an increasing shortage of skilled phlebotomists.

      There is no cutting-edge lab in San Francisco focused on venipuncture, no established benchmark, no leaderboard, and no viral demos. What exists is a vast number of arms and a dwindling number of trained individuals who can locate veins.

      Vitestro did not present a demo to the regulators. In the A.D.O.P.T. trial conducted across Dutch hospitals, the device achieved a 95% success rate for first-stick attempts compared to a manual benchmark of 93 to 97%, a haemolysis rate of 0.6% against a best-practice limit of 2%, and mild adverse events in 1.3% of participants with no serious or moderate incidents recorded.

      A crucial detail buried in the FDA’s special controls deserves more attention. Manufacturers must now demonstrate performance that is comparable to or exceeding that of trained human phlebotomists across a diverse range of patients, including those with difficult vein access and various skin tones.

      Failed venipunctures are not evenly distributed. They disproportionately affect individuals with darker skin, the severely ill, chemotherapy patients whose veins may have been depleted, as well as the dehydrated, elderly, and obese. A fingertip and an untrained eye are inadequate tools for that task.

      Near-infrared light and Doppler ultrasound interpret pigments differently than human vision, and whether this will bridge the gap in standard practice is now an empirical question that regulators mandate the industry to address.

      In fact, the regulator did something more compelling than approving a product; it created a framework. The De Novo decision set forth special controls regarding labeling, performance testing, and clinical testing, allowing future developers of similar devices to follow the standard 510(k) process instead of starting from scratch.

      Regulators are often criticized for arriving long after a technology has matured, arriving breathless with a clipboard. In this instance, the groundwork was laid before the introduction of the technology. Given that much of health AI currently operates without any oversight—a concern the WHO highlighted in its warning about Europe’s healthcare AI governance gap—this is significant.

      The case for Aletta fundamentally revolves around staffing issues, and such cases have a history. According to US Bureau of Labor Statistics data, there were 139,700 employed phlebotomists in 2024, expected to increase by 6% by 2034, with about 18,400 annual job openings and a median wage around $43,660. Most openings arise because of attrition.

      When a board member from a hospital in Nieuwegein explained why they ordered two of these machines, he did not mention accuracy. He talked about work pressure. This honesty also highlights the reality that makes a phlebotomist reconsider their news.

      Professional skepticism is also worth serious consideration. Educators in phlebotomy have questioned whether comparing a 2026 robot against manual success rates from older