# Tempus gets up to $9.5M to test Olivia as autonomous heart-care AI > Source: > Published: 2026-09-10 19:00:25+00:00 # Tempus gets up to $9.5M to test Olivia as autonomous heart-care AI **Eric Lefkofsky's Tempus will develop and validate a heart-failure agent under an FDA-engaged federal program, though its authorization path and treatment authority remain unsettled.** By [RuntimeWire Staff](/author/runtimewire-staff) ยท Published Primary source: [Tempus AI on X](https://x.com/TempusAI/status/2098023384055484841) ## Why it matters Tempus now has federal funding to develop and prospectively test an autonomous heart-failure agent. The study could produce evidence for clinical and regulatory review, but no authorization, treatment authority or patient benefit has been established. [Eric Lefkofsky (@lefkofsky)](https://x.com/lefkofsky?ref=runtimewire) has secured up to $9.5 million in federal support for Tempus to develop an AI agent intended to manage parts of heart-failure care, pushing his decade-old precision-medicine business beyond diagnostics and alerts toward clinical action. The funding comes from the Advanced Research Projects Agency for Health, or ARPA-H. The award, [announced on September 9](https://tempus.gcs-web.com/news-releases/news-release-details/tempus-secures-95-million-arpa-h-advocate-funding-deploy-first?ref=runtimewire), will support product development and a prospective, multicenter clinical study. The $9.5 million is a ceiling; public materials do not disclose the payment schedule, contract conditions or amount disbursed so far. The project extends Olivia, a patient-facing app Lefkofsky introduced in 2024 to collect medical records, track symptoms and help patients prepare for appointments. The product grew from the experiences that led him to found Chicago-based Tempus in August 2015, after his wife was diagnosed with breast cancer. Lefkofsky later [named Olivia](https://www.tempus.com/content/article/meet-olivia/?ref=runtimewire) for a 19-year-old friend of his daughter who died from cancer in 2021. Lefkofsky had already built or co-founded Groupon, InnerWorkings, Echo Global Logistics, Mediaocean and Lightbank before entering healthcare. He holds bachelor's and law degrees from the University of Michigan. Tempus became his attempt to apply the data systems he knew from commerce, advertising and logistics to the fragmented records and treatment decisions he encountered in cancer care. Investors put more than $1.3 billion into Tempus before its public listing, with backers including Google, New Enterprise Associates, Revolution, SoftBank and Baillie Gifford, [according to Forbes](https://www.forbes.com/sites/phoebeliu/2024/06/14/billionaire-groupon-founder-eric-lefkofskys-health-tech-company-goes-public-his-fourth-ipo/?ref=runtimewire). Tempus completed its Nasdaq IPO in June 2024, raising approximately $410.7 million in gross proceeds. The ARPA-H project carries that founding idea into a higher-risk setting. Olivia currently helps patients organize and understand information. The proposed heart-failure version is expected to combine electronic-health-record data with home heart-rate monitoring, detect meaningful changes and support management of symptoms, medications, virtual rehabilitation and scheduling, [according to Tempus's announcement](https://www.tempus.com/news/pr/tempus-secures-up-to-9-5-million-in-arpa-h-advocate-funding-to-deploy-first-autonomous-ai-in-cardiology/?ref=runtimewire). ### From detecting care gaps to acting on them Tempus has spent years building cardiology software that identifies patients who may have undiagnosed or undertreated disease. Its [cardiology portfolio](https://www.tempus.com/cardiology/?ref=runtimewire) includes more than 40 care-gap algorithms across 15 cardiovascular conditions, according to the company. The company also says its ECG-Low EF algorithm is FDA-cleared to detect signals associated with reduced ejection fraction. Those products largely screen, predict and notify. The ARPA-H project is intended to test whether an agent can take defined clinical actions. The project is intended to make Olivia an autonomous clinical agent within the care team. That is an ambitious target. Heart-failure treatment can require repeated adjustments to medication, diet, exercise and follow-up care, with decisions shaped by symptoms, vital signs, laboratory results and other medical conditions. Tempus remains at the development stage. The award does not establish that Olivia currently changes prescriptions, provides autonomous treatment or improves patient outcomes. [ARPA-H says patient-facing agent teams must submit an FDA authorization package within 24 months of their contract awards](https://arpa-h.gov/news-and-events/arpa-h-launches-worlds-first-bid-build-fda-authorized-clinical-ai-cardiovascular?ref=runtimewire), but that requirement does not mean Tempus has received authorization or secured a completed regulatory pathway. [ARPA-H's ADVOCATE program](https://arpa-h.gov/explore-funding/programs/advocate?ref=runtimewire) seeks clinical agents that can make defined changes to appointments, medication, diet and exercise while escalating cases to human clinicians. The [agency's September 9 announcement](https://arpa-h.gov/news-and-events/arpa-h-launches-worlds-first-bid-build-fda-authorized-clinical-ai-cardiovascular?ref=runtimewire) says it committed up to $33.7 million during the first year of the four-year program, which has up to $62.7 million in total funding. The ADVOCATE program is a four-year effort with up to $62.7 million in total funding. Its [other named patient-facing participants include Atman Health and UpDoc](https://arpa-h.gov/news-and-events/arpa-h-launches-worlds-first-bid-build-fda-authorized-clinical-ai-cardiovascular?ref=runtimewire), whose approaches emphasize adaptive questioning and clinician-built protocol checks, respectively. Atman is developing a voice-first system that changes its questions as it gathers information. UpDoc separates its conversational interface from a clinician-built rules engine intended to check proposed actions against approved protocols. The same [ARPA-H announcement assigns Stanford University](https://arpa-h.gov/news-and-events/arpa-h-launches-worlds-first-bid-build-fda-authorized-clinical-ai-cardiovascular?ref=runtimewire) to develop a supervisory AI that monitors the clinical agents for unsafe or unusual behavior. [Duke University and Kaiser Permanente](https://arpa-h.gov/news-and-events/arpa-h-launches-worlds-first-bid-build-fda-authorized-clinical-ai-cardiovascular?ref=runtimewire) have multi-site validation and enterprise deployment roles, including work involving rural sites. [Johns Hopkins University's Applied Physics Laboratory](https://arpa-h.gov/news-and-events/arpa-h-launches-worlds-first-bid-build-fda-authorized-clinical-ai-cardiovascular?ref=runtimewire) is the external evaluation partner. That structure addresses a central problem with autonomous medical software: a patient conversation is only the visible layer. The system also needs traceable clinical rules, continuous oversight, integration with health records and a defined point at which a human takes control. ### The grant buys a clinical test, with key terms unresolved For Tempus, the award is small relative to the company. Tempus reported $1.27 billion in 2025 revenue and more than 3,800 employees at year-end in its [annual filing](https://www.sec.gov/Archives/edgar/data/1717115/000119312526066961/tem-20251231.htm?ref=runtimewire). A grant capped at $9.5 million will not materially reshape that balance sheet. Tempus has a funded route into clinical validation and FDA engagement, but the public materials do not yet establish how much authority the heart-failure agent will receive or when it could be authorized. ARPA-H intends to engage the FDA throughout ADVOCATE, while the required authorization package remains an application rather than an approval. Success would widen Lefkofsky's original bet. Tempus began by using large clinical and molecular datasets to help doctors personalize cancer treatment. An authorized heart-failure agent could place the company inside the continuous management of a chronic disease, where software may recommend or carry out care changes rather than only organize information. Whether Tempus reaches that point depends on evidence and regulatory decisions that do not yet exist. The planned study will combine EHR data with home heart-rate monitoring and prospectively validate the system across multiple centers. Public materials do not disclose the sites, enrollment target, endpoints, start date or timeline. They also provide no clinical outcomes, safety results, patient-usage figures or product pricing for the proposed agent. FDA authorization status and the final level of autonomous treatment authority remain unestablished.