The documentary You Can See Everything shows Elizabeth Holmes and her partner Billy Evans working on what Evans calls Theranos 2.0, a new venture named Haemanthus that centers on a diagnostic device using lasers. The film presents the effort as a fresh start, though Evans describes it as getting the band back together, and the official line holds that Holmes is not leading it. The project has drawn attention because Theranos became a prime example of overinflating what technology can do. The Theranos device never worked as advertised, and the company brought in Siemens devices using traditional blood testing methods to fake results.
Haemanthus has no website beyond a sparse LinkedIn page listing two to 10 employees and describing its work as AI-native sensors for health that capture thousands of biomarkers. Most public information comes from the film and a 2025 New York Times article, which reports the company is building a device that uses a laser to scan blood, saliva, or urine from pets and analyze the sample at a molecular level. The startup is apparently aiming at veterinary tests first before perhaps eventually moving to humans, with a long-term goal of a stamp-size, wearable version for humans that the article says will require three years and $70 million.
A patent published by The New York Times indicates Haemanthus uses Raman spectroscopy, a technique that does involve lasers but was first discovered in 1928. The method identifies chemical structures in a molecule and has numerous biological and medical uses, including noninvasive study of tissue samples. The New York Times report describes aims to identify and qualify biomarkers such as glucose and hormones and then use AI to detect cancer and infections. Researchers are exploring the technology for early-stage cancer diagnosis and glucose monitoring, though that does not establish it as a better method than existing ones in every scenario.
Practical obstacles have long surrounded small-sample blood testing. A finger prick carries a higher risk of rupturing blood cells, which can skew results, and the finger is prone to surface contamination. A finger prick sample may also mix with interstitial fluid and be diluted. Different biomarkers can require wholly different testing procedures, and blood alone cannot test for every disease.
Timelines for medical devices are another point of scrutiny. Evans said in the documentary, in 2023, that Haemanthus would have a novel, revolutionary prototype working in about a year. The article notes it is already 2026. The article also points to the fastest recent public health effort, the covid-19 vaccine, which took roughly 10 months to develop against a typical five-to-15-year timeline, building on three decades of mRNA research and tens of billions of dollars, largely from the government.
The article compares the Haemanthus timeline with the wearable market, where the biggest names can only detect, not diagnose, signs of perhaps two or three conditions. Noninvasive continuous glucose monitoring is treated as a holy grail, an endeavor that started in 1975 and that Apple has been rumored to be working on for over 15 years. Glucose monitoring attempts began in the mid-1800s, the first FDA-approved continuous glucose monitor was the Medtronic MiniMed in 1999, the first real-time version arrived in 2004, over-the-counter models reached the market in 2024, and this year Abbott announced a monitor that can measure two biomarkers.
More technology news from TechManNews.
