Tempus AI (NASDAQ:TEM) executives highlighted accelerating growth across both its diagnostics and data businesses during the company’s fourth-quarter and full-year 2025 fiscal results conference call, pointing to strong oncology testing demand, rapid adoption of minimal residual disease (MRD) testing, and expanding data licensing activity.
Management characterizes 2025 as “exceptional,” with core revenue up more than 33%
Founder and CEO Eric Lefkofsky said 2025 “was an exceptional year” for Tempus, with both major businesses “growing rapidly and performing well above expectation.” He said total revenue in the company’s core business increased by more than 33% year over year, adding that growth was higher when including the acquisition of Ambry.
Data business momentum, contract value, and retention highlighted
Lefkofsky said the data segment is growing faster than diagnostics and described it as consisting primarily of licensing and applications. He said the licensing business (Insights) was up 69% in the quarter when factoring in the one-time impact of an AstraZeneca warrant, and he projected “roughly 40% growth” for the current quarter.
Management also emphasized increasing commercial traction and visibility:
- Total contract value (TCV) was described as greater than $1.1 billion, with Lefkofsky noting it has been rising faster than revenue over the past several quarters.
- Net revenue retention was 126%, which he characterized as “super strong.”
During Q&A, Lefkofsky said bookings have been strong enough that the company is starting 2026 with greater visibility into the data revenue build than it has had previously. He also noted it has been “normal” for Tempus to generate about $100 million of revenue within a given year from in-year bookings, describing the current demand environment as “crazy amounts of demand” for its data products.
AI strategy: proprietary data and distribution positioned as core differentiators
Responding to questions about competition from broader AI entrants in healthcare, Lefkofsky argued that durable AI business models depend on proprietary data for training and proprietary distribution for delivering insights. He said Tempus has both “at scale,” citing more than 450 petabytes of connected multimodal data generated through its diagnostic operations and longitudinally connected to outcomes and responses.
He also emphasized distribution through clinical workflows, citing connections to more than 5,500 hospitals and more than 8,500 regularly ordering oncologists, along with thousands of other physicians. Lefkofsky said replicating Tempus’s real-time data pipelines and structured datasets would be “quite hard,” and he framed the company’s client renewals and expansion as evidence of value—particularly through helping biopharma customers refine discovery efforts, design trials, recruit appropriate patient populations, and accelerate drug development timelines.
Product and operational updates: Paige Predict, ASP trajectory, and foundation model progress
Executives discussed several initiatives intended to expand clinical insight and, over time, improve economics.
On Paige Predict, Lefkofsky said the capability exemplifies Tempus’s strategy of “contextualization” by pairing technology with diagnostics. He described Paige Predict as enabling digitization of pathology slides and using them to generate insights that can predict mutations expected to appear in next-generation sequencing. He said this can be helpful when sequencing yields results that cannot be returned, noting that some failures are inherent in sequencing processes. He also said speed—rendering results in hours—can be an additional advantage.
CFO Jim Rogers provided detail on average selling price (ASP) and the path toward higher levels. He said ASPs in Q4 were around $1,640, up about $40 quarter over quarter. Rogers reiterated the company’s view that there is $500 or more of potential upside to ASP based on mix and several drivers, including continued migration of xT CDx from an LDT version to an FDA-approved version. He said Tempus plans that by the end of 2026, the “vast majority” of volume will be on the FDA-approved version, describing it as the largest ASP driver in 2026.
Rogers also noted that Tempus has submitted xF (liquid biopsy) to the FDA, but said it is unlikely to have much of an ASP impact in 2026, with expected contribution beginning in 2027. He added that there is still potential upside from progress with commercial payers.
On AI model development, Lefkofsky said the oncology foundation model had a Q1 deliverable tied to benchmarks established by AstraZeneca and that Tempus believes it hit those benchmarks and submitted the work to AstraZeneca, which is testing the model. He said Tempus established a dedicated cluster of a little over 1,000 H200s for the oncology foundation model and has procured a second cluster of more than 500 GB200, which he said provides greater compute power than the first cluster. Lefkofsky added that Tempus is running additional internal models across other modalities and disease areas, citing radiology, pathology, cardiology, and neuropsych data, and said the company is “doubling down” on those efforts.
MRD gating tied to reimbursement timing; oncology growth seen as durable
Tempus executives described MRD as a major growth vector, while emphasizing that commercialization remains intentionally constrained. Lefkofsky addressed a question about MRD volume and sales coverage, calling a “20 times higher” comment hypothetical but meant to illustrate demand strength. He confirmed that MRD sales efforts are currently constrained to a small portion of the sales force and said the company intends to “eventually unblock” the effort as reimbursement and timing align, seeking to scale in a financially disciplined way.
On MRD reimbursement and assay development, Lefkofsky said the company is in a “back and forth” with MolDX regarding a first-generation colorectal cancer (CRC) assay and said timing is uncertain. He added that the current MRD offering is largely tumor-informed (he estimated roughly 95%) and that the MolDX decision is “not that relevant” to current penetration. He also discussed a pivot to a second-generation tumor-naive assay after concluding performance from the first version was not sufficient, and said the second version is “coming along well.”
Rogers said the company is not seeing oncology volume growth slow meaningfully, noting that the first quarter is off to a “good start.” He said hereditary growth is expected to moderate as Tempus laps share gains, with “lumpiness” possible in 2026; the company referenced high-teens longer-term hereditary growth as achievable, potentially lower in Q1 before improving.
Discussing mix, management said both tissue and liquid comprehensive genomic profiling are growing, with liquid growing “a little faster” than solid, consistent with long-term trends. Executives also said they are not seeing major mix shifts or a single driver (such as repeat testing) dominating growth.
Lefkofsky closed by stating the company’s priorities remain focused on bringing technology and AI to diagnostics and ensuring clinical and research decisions are data-driven, while noting Tempus has made no major changes to its sales organization following a reorganization earlier in 2025.
About Tempus AI (NASDAQ:TEM)
Tempus is a technology-driven healthcare company that applies artificial intelligence and machine learning to clinical and molecular data in order to advance precision medicine. Its primary focus lies in oncology, where the company offers comprehensive genomic profiling, digital pathology services and data-driven insights to inform personalized cancer care. By integrating DNA and RNA sequencing with structured clinical information, Tempus enables clinicians and researchers to identify targeted treatment options for patients based on the genetic characteristics of their tumors.
The company’s core offering centers on a scalable, cloud-based analytics platform that aggregates vast amounts of molecular and clinical data.
