← back to rankingTEM · Tempus AI, Inc.
Medical - Healthcare Information Services · mkt cap $8.7B · calls: Q1 FY2026 vs Q4 FY2025
47.0 conviction · conf-adj 47
conf 6/10 partial
enthusiasm:24.0 · trend:0 · quantifies:12 · impact:8 · under_radar:0 · credibility:0 · confirmation:3
Enthusiasm latest 8 / prev 9 (flat)
The AI thesis is data/model driven: customers license Tempus' proprietary clinical and molecular data and increasingly build foundation or smaller models on it. Credibility is supported by revenue, bookings, TCV, NRR, strategic pharma deals, and algorithm attach-rate figures, but bottomline AI impact is not separately quantified. Latest enthusiasm remains high, though Q4 FY2025 was more expansive about AI strategy and foundation models.
GROUNDED NEXT-FY IMPACT vs CONSENSUS
Grounded on actual base — revenue $1.3B · net income $-0.2B · net margin -19.3% · diluted EPS -1.41
These are next-fiscal-year annual uplift estimates, not next-quarter numbers.
Aggregate next-FY est. rev uplift: 12.5807% · next-FY EPS uplift: % · vs analysts: inline · priced in: high · confidence: 6/10
| Claim | Figure | Arithmetic | Next-FY Rev % | Next-FY EPS % |
|---|
Data & Apps / data revenue scale and projected growth
revenue | ~$400M FY25 data revenue; ~40% FY26 growth; Q1 FY26 $87M, +40.5% YoY | Use the stronger disclosed-base arithmetic: 4 * $100M = ~$400M FY25 data revenue; 40% growth = $160M incremental FY26 revenue; $160M / $1.271789B current revenue = 12.5807%. Q1 actual cross-check: $87M / 1.405 = $61.922M prior Q1, +$25.078M, or ~$100.313M annualized before assumed H2 ramp. EPS% null because current/consensus earnings are negative. | 12.5807 | |
Insights/licensing/modeling +44%
revenue | >44% | Sub-component/driver of the Data & Apps growth already counted above. Using it separately would double-count the same AI-driven data revenue stream; no additive uplift assigned. | | |
Insights +69% in Q4 FY2025 including AstraZeneca warrant
revenue | +69% | Historical and partly inflated by one-time AstraZeneca warrant impact. The forward ex-warrant guide is the ~40% growth used in the aggregate; not additive. | | |
Bookings north of $100M, third straight quarter
revenue | >$100M/qtr | Bookings are not next-FY revenue. They support the same Data & Apps growth and TCV conversion already reflected in the $160M growth estimate; no separate additive uplift assigned. | | |
Strategic data agreements $100M+ each, almost half-dozen, multi-year
revenue | $100M+ each; almost half-dozen; multi-year | Multi-year TCV recognized ratably and embedded in bookings/TCV plus the guided data-revenue growth. Treating 5 * $100M as separate next-FY revenue would double-count; no additive uplift assigned. | | |
Total contract value greater than $1.1B
revenue | >$1.1B | TCV is contracted/backlog-like, not current next-FY revenue. It supports durability of the guided 40% data growth but is not additive without a disclosed conversion schedule. | | |
Net revenue retention 126%
engagement | 126% | Existing accounts expand by 26%, but this is a driver of the same Data & Apps revenue growth. Using $400M * 26% = $104M would overlap with the $160M aggregate growth; no additive uplift assigned. | | |
Algorithm attach rate on Solid Tumor assays
engagement · soft | ~40% | Attach rate is disclosed, but solid-tumor assay revenue, volume, algorithm ASP, and incremental reimbursement are not disclosed. No anchored dollar impact can be calculated. | | |
Tumor Origin addressable frequency
engagement · soft | ~5% of cancer patients | Addressable patient frequency lacks disclosed patient count, test volume, ASP, reimbursement, or conversion rate. No anchored dollar impact can be calculated. | | |
Alzheimer's multimodal model project
revenue · soft | multi-million-dollar project; finishes mid-FY26 | Multi-million is not a specific dollar amount and no contract value or revenue schedule is disclosed. No anchored next-FY revenue impact can be calculated. | | |
Database/data set scale
other · soft | over 450PB / in excess of 500PB | Operational moat/capability metric with no disclosed revenue conversion, price, customer conversion, or cost saving. No anchored dollar impact can be calculated. | | |
Provider distribution
engagement · soft | >5,500 hospitals; >8,500 regularly ordering oncologists | Distribution footprint is disclosed, but no incremental utilization, ASP, reimbursement, or conversion rate is disclosed. No anchored revenue uplift can be calculated. | | |
Foundation-model compute
other | a little over 1,000 H200s; more than 500 GB200 | Compute capacity supports internal oncology foundation-model development and is more likely near-term capex/opex headwind than saving. No disclosed dollar cost or revenue conversion to size. | | |
Assumptions: Next FY = FY2026. AI impact is primarily Tempus's Data & Applications / Insights growth. Use disclosed ~$400M FY25 data-revenue base and ~40% projected growth, yielding $160M incremental FY26 revenue. Overlapping sub-claims, bookings, strategic agreements, TCV, and NRR are treated as support for the same growth, not additive. EPS uplift is null because current and consensus earnings bases are negative.
Top line: AI initiatives show up mainly in Data & Applications / Insights. Disclosed base ~$400M growing ~40% implies ~$160M incremental FY26 revenue, equal to 12.5807% of current revenue. Q1 FY26 $87M at +40.5% supports the trajectory, while TCV >$1.1B, bookings >$100M, strategic agreements, and 126% NRR support durability but are not additive.
Bottom line: EPS uplift is not meaningful to quantify as a percent because the company is loss-making and consensus earnings remain negative. Incremental data revenue may be high gross-margin, but foundation-model compute investment and ongoing losses make near-term EPS flow-through unanchored; est_eps_uplift_pct is null.
[EPS uplift n/m (loss-making base)] Consensus FY26 revenue $1,593.3M vs FY25 $1,271.8M = +$321.6M (+25.3%). The AI/data piece (~$160M, +12.6%) is the largest single driver and was explicitly guided ('projecting ~40%'), so it sits squarely inside consensus — priced in. The genuinely un-modeled upside ('data business to multi-billion dollars', 'foundation models accelerate that dramatically') is forward and unanchored, so it can't be scored as 'ahead.' Bottom line: anchored impact is real but inline; the asymmetric upside is soft.
MODEL CONSENSUS (impact)
partial
Consensus keeps the hard disclosed-base FY26 Data & Apps growth estimate and removes overlapping or assumption-heavy additive uplifts.
Conflicts reconciled
- Data revenue uplift: X=12.6 vs Y=7.8876/12.5807 -> used 12.5807 because disclosed ~$400M base plus ~40% guide is stronger than annualizing Q1 only.
- Bookings/TCV/strategic agreements: X=not additive vs Y=sized with ratable assumptions -> used not additive because bookings/TCV are not revenue and conversion periods were not disclosed.
- Insights +44%/+69%/NRR: X=not additive vs Y=sized separately -> used not additive because they overlap with the same Data & Apps growth aggregate.
- Database/provider metrics: X sometimes soft=true and Y soft=true -> used soft=true because no dollar conversion is disclosed.
| Field | Opus 4.8 | GPT-5.5 |
|---|
| Rev uplift % | 12.6 | 12.5807 |
| EPS uplift % | – | – |
| Priced in | high | high |
| vs analysts | inline | inline |
| Confidence | 6 | 6 |
| Top line | AI shows up almost entirely in the Data & Applications / Insights segment. Disclosed base ~$400M (FY25) growing ~40% -> ~$160M incremental in FY26 = +12.6% of total revenue, corroborated by Q1 FY26 actuals ($87M, +40.5%; ex-warrant Insights up 69% in Q4, guided ~40% forward). NRR of 126% and TCV >$1.1B rising faster than revenue support durability. This 12.6% is HARD (disclosed base + guided growth), but it is the bulk of, and consistent with, consensus's +25.3% total-revenue step-up. | The hard, non-duplicative next-FY AI revenue impact is about $160M, equal to 12.581% of the $1.271789B revenue base. Other hard claims support the same Data and Apps trajectory: Q1 Data and Apps growth implies $100.313M annualized incremental revenue, NRR implies $104M expansion, and lower-bound TCV ratable conversion implies $220M of annual revenue support, but these overlap rather than add. |
| Bottom line | Not computable as an EPS %: company is loss-making (NI -$245M) and consensus stays negative through FY27 (-$3.5M), so any EPS-uplift % off that base is an artifact (guardrail -> null). Direction is actually a near-term drag: the foundation-model buildout (1,000 H200 + 500 GB200 clusters) is adopter capex/opex that pressures margins. High-margin data revenue (corp GM 69.6%, licensing higher) is the path to eventual profitability, but the AI investment cycle keeps net income negative for now. | At the assumed 35% incremental margin, the $160M revenue uplift would imply about $56M incremental net income before considering broader company losses. EPS uplift percentages are not meaningful because both current and consensus earnings bases are negative. |
| Reasoning | Consensus FY26 revenue $1,593.3M vs FY25 $1,271.8M = +$321.6M (+25.3%). The AI/data piece (~$160M, +12.6%) is the largest single driver and was explicitly guided ('projecting ~40%'), so it sits squarely inside consensus — priced in. The genuinely un-modeled upside ('data business to multi-billion dollars', 'foundation models accelerate that dramatically') is forward and unanchored, so it can't be scored as 'ahead.' Bottom line: anchored impact is real but inline; the asymmetric upside is soft. | Consensus 2026 revenue of $1.593345B is $321.556M above the supplied $1.271789B base, or 25.284% growth. The deduplicated AI/Data uplift of $160M is 49.758% of that consensus revenue step-up and fits inside existing estimates rather than clearly exceeding them. Consensus 2027 revenue adds another $352.657M, also leaving room for continuing Data and Apps growth. EPS comparison is not meaningful because consensus EPS remains negative in 2026 and near breakeven in 2027. |
Rows highlighted where the two models disagreed.
QUANTIFICATIONS
Data and applications revenue: $87 million; 40.5% year-over-year growth (Q1 FY2026, topline)
“Our data business, data and applications business did extraordinarily well, $87 million of revenue, representing 40.5% year-over-year growth with particular strength in our data licensing and modeling business insights, which grew over 44%.”
Data licensing and modeling business Insights growth: over 44% (Q1 FY2026, topline)
“Our data business, data and applications business did extraordinarily well, $87 million of revenue, representing 40.5% year-over-year growth with particular strength in our data licensing and modeling business insights, which grew over 44%.”
Bookings: north of $100 million (Q1 FY2026; third straight quarter, topline)
“We had our third straight quarter of bookings north of $100 million, with TCV rising and visibility in the best place it's been for our Data and Apps business in quite some time.”
Strategic data agreements: $100 million-plus agreements (Q1 FY2026; multiple years, topline)
“But we now have almost half a dozen folks at that level where people are signing these very large strategic agreements with more coming.”
Database scale: in excess of 500 petabytes (Q1 FY2026, topline)
“But we have a very large database now in excess of 500 petabytes of data, it's all connected to this analytics and model-building platform.”
Algorithm attach rate: roughly 40% (Q1 FY2026, topline)
“Eric, it was highlighted in the prepared remarks that roughly -- you have a 40% attach rate for your algos, I think it was on your Solid Tumor assays in Oncology.”
Tumor Origin algorithm addressable case frequency: about 5% of cancer patients (Q1 FY2026, topline)
“Some of them, for example, are our Homologous Recombination Deficiency algorithm our Tumor Origin algorithm where in about 5% of cancer patients.”
Alzheimer's multimodal model project: multi-million dollar project (Q1 FY2026; finish up middle of this year, topline)
“It has been nice to watch some recent wins in Neurology and in particular, we were just engaged to begin building a multimodal model in Alzheimer's disease that was a multi-million dollar project that we're in the middle of right now that we'll finish up middle of this year.”
Insights growth: 69% (Q4 FY2025, topline)
“Our licensing business, or Insights, was up 69% in the quarter when you factor in the onetime impact of the AstraZeneca warrant, and we're projecting roughly 40% growth this quarter.”
Projected Insights growth: roughly 40% (Q1 FY2026 projection given on Q4 FY2025 call, topline)
“Our licensing business, or Insights, was up 69% in the quarter when you factor in the onetime impact of the AstraZeneca warrant, and we're projecting roughly 40% growth this quarter.”
Total contract value: greater than $1.1 billion (Q4 FY2025, topline)
“Total contract value was greater than $1.1 billion and most importantly, has been rising faster than revenue over the past several quarters.”
Net revenue retention: 126% (Q4 FY2025, topline)
“And net revenue retention was 126%, which is super strong, all things considered.”
Data set scale: over 450 petabytes (Q4 FY2025, topline)
“We have over 450 petabytes of connected multimodal data, which flows from our Diagnostic business, which has real-time insights, real-time connection to outcome and response, is able to track patients longitudinally, rich molecular data, rich imaging data.”
Provider distribution: more than 5,500 hospitals; more than 8,500 regularly ordering oncologists (Q4 FY2025, topline)
“And then once we generate insights or some kind of contextualization that we want to put in the hands of a doctor, because we're connected to more than 5,500 hospitals, because we have more than 8,500 regularly ordering oncologists and thousands of other physicians and other areas, we can deliver these insights in real time as part of routine clinical care.”
Data revenue scale versus prior target: 4x $100 million (Q4 FY2025, topline)
“And we're now 4x that and projecting to grow 40% even at this scale.”
Foundation model compute: a little over 1,000 H200s; more than 500 GB200 (Q4 FY2025, topline)
“And so those efforts will go on. We've also -- that was a particular cluster we set up of, about a little over 1,000 H200s dedicated to that Oncology foundation model. We've also procured a second cluster of more than 500 GB200.”
PAST (realized)
- Our data business, data and applications business did extraordinarily well, $87 million of revenue, representing 40.5% year-over-year growth with particular strength in our data licensing and modeling business insights, which grew over 44%.
- So Merck was a very large strategic data and modeling collaboration, we have very large collaborations with people like AstraZeneca, GSK, and BMS.
- Our AI advantages are continuing to take hold.
- So the foundation model had a deliverable in Q1 where we had to hit certain benchmarks that AstraZeneca had established. We think we've hit all those benchmarks.
- And it's because we're demonstrating real -- our data is demonstrating real value where these clients are able to use our data and our technology to be more intelligent about what assets to go forward with, refine their early-stage discovery projects, design more intelligent Phase IIs and Phase IIIs, recruit the right populations at the right time and get their drugs approved and in market faster.
CURRENT (now)
- And we're seeing this migration where people aren't just licensing our data more and more, they're actually building a models with us, whether those are foundation models as is the case with AstraZeneca, or they're building smaller models, leveraging our data.
- That's now connected to not just CPUs, but GPUs, and people are increasingly building proprietary models to get smarter about their own internal R&D programs.
- It has been nice to watch some recent wins in Neurology and in particular, we were just engaged to begin building a multimodal model in Alzheimer's disease that was a multi-million dollar project that we're in the middle of right now that we'll finish up middle of this year.
- And as these algorithms basically get more and more pervasively ordered, they're just another tool in this overall bag of technology-enabled assets that our physicians increasingly rely on.
- We have over 450 petabytes of connected multimodal data, which flows from our Diagnostic business, which has real-time insights, real-time connection to outcome and response, is able to track patients longitudinally, rich molecular data, rich imaging data.
FORWARD (guidance)
- The data business -- especially the data and modeling business in the U.S., getting to multi-billion dollars.
- And if you look at the foundation model efforts that we're engaged in, that's going to do nothing, we think, but accelerate that dramatically.
- And so to the extent today that we're X percent better than somebody else because we can generate Y percent more insights, you should expect that to grow quite a bit over the next several years.
- We think they'll be catalytic both to our diagnostic business as we keep dropping insights into our tests that make our test smarter and better than others.
- And then we also intend to propagate these insights through our data business so that our clients get more value.
PRICED-IN (REFINED)
UNKNOWNEst. revisions unknown · Fwd P/E -82.8 · EV/Sales 6.6x
LLM synthesis failed: codex: OpenAI Codex v0.136.0
--------
workdir: /tmp/aes-codex
model: gpt-5.5
provider: openai
approval: never
sandbox: danger-full-access
reasoning effort: medi
COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS
RECENT AI CATALYSTS & NEWS
BUSINESS IMPACT - QUALITATIVE MATERIALITY
-/10 immaterial unclear · soft evidence
LLM materiality judgment failed: codex: OpenAI Codex v0.136.0
--------
workdir: /tmp/aes-codex
model: gpt-5.5
provider: openai
approval: never
sandbox: danger-full-access
reasoning effort: medium
reasoning summari
OPTIONS / MARKET STRUCTURE
option liquidity: good
proxy inputs — dollar-ADV $300M · beta 3.62095 · px $49.62
source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.
CONFIRMATION — INSIDERS · 13F · LANGUAGE
Mixed — insiders selling, institutions adding, management language unknown.
INSIDERS selling 59 open-market sell(s) vs 0 buy(s) — net distribution
INSTITUTIONS (13F) adding as of 2026-03-31: 82 new / 113 closed positions; 248 increased / 110 reduced; institutional ownership +2.71pp; -46 net 13F holders
MGMT LANGUAGE n/a unknown language analysis failed: codex: OpenAI Codex v0.136.0
--------
workdir: /tmp/aes-codex
model: gpt-5.5
provider: openai
approval: never
sandbox: danger-full-access
reasoning effort: medi
VERBATIM AI QUOTES
“Our data business, data and applications business did extraordinarily well, $87 million of revenue, representing 40.5% year-over-year growth with particular strength in our data licensing and modeling business insights, which grew over 44%.”
— Eric Lefkofsky, Q1 FY2026
“And we're seeing this migration where people aren't just licensing our data more and more, they're actually building a models with us, whether those are foundation models as is the case with AstraZeneca, or they're building smaller models, leveraging our data.”
— Eric Lefkofsky, Q1 FY2026
“But we have a very large database now in excess of 500 petabytes of data, it's all connected to this analytics and model-building platform.”
— Eric Lefkofsky, Q1 FY2026
“That's now connected to not just CPUs, but GPUs, and people are increasingly building proprietary models to get smarter about their own internal R&D programs.”
— Eric Lefkofsky, Q1 FY2026
“So Merck was a very large strategic data and modeling collaboration, we have very large collaborations with people like AstraZeneca, GSK, and BMS.”
— Eric Lefkofsky, Q1 FY2026
“It has been nice to watch some recent wins in Neurology and in particular, we were just engaged to begin building a multimodal model in Alzheimer's disease that was a multi-million dollar project that we're in the middle of right now that we'll finish up middle of this year.”
— Eric Lefkofsky, Q1 FY2026
“We have a variety of algorithms that we've built over the years.”
— Eric Lefkofsky, Q1 FY2026
“We -- off of our transcriptomic assay, RNA assay, we can actually predict that with super high fidelity.”
— Eric Lefkofsky, Q1 FY2026
“And as these algorithms basically get more and more pervasively ordered, they're just another tool in this overall bag of technology-enabled assets that our physicians increasingly rely on.”
— Eric Lefkofsky, Q1 FY2026
“Our AI advantages are continuing to take hold.”
— Eric Lefkofsky, Q4 FY2025
“So I mean, I think the most interesting business models, I believe, surrounding AI, in particular, large language or large multimodal models, really center around access to proprietary data to train models and proprietary distribution once you have a model that generates insight.”
— Eric Lefkofsky, Q4 FY2025
“We have over 450 petabytes of connected multimodal data, which flows from our Diagnostic business, which has real-time insights, real-time connection to outcome and response, is able to track patients longitudinally, rich molecular data, rich imaging data.”
— Eric Lefkofsky, Q4 FY2025
“So we have this really unique proprietary data set that you can use to train AI, to train models.”
— Eric Lefkofsky, Q4 FY2025
“And it's because we're demonstrating real -- our data is demonstrating real value where these clients are able to use our data and our technology to be more intelligent about what assets to go forward with, refine their early-stage discovery projects, design more intelligent Phase IIs and Phase IIIs, recruit the right populations at the right time and get their drugs approved and in market faster.”
— Eric Lefkofsky, Q4 FY2025
“The benefit of having a system where you're sequencing tons of patients and following mutations and you're digitizing pathology slides and tracking those is you can begin to correlate these things.”
— Eric Lefkofsky, Q4 FY2025
“And if you look at the foundation model efforts that we're engaged in, that's going to do nothing, we think, but accelerate that dramatically.”
— Eric Lefkofsky, Q4 FY2025
“So the foundation model had a deliverable in Q1 where we had to hit certain benchmarks that AstraZeneca had established. We think we've hit all those benchmarks.”
— Eric Lefkofsky, Q4 FY2025
ANALYST QUESTIONS ON AI
Q (Q1 FY2026, Kallum Titchmarsh): Can you maybe just talk about how discussions with Large Pharma customers have been trending so far this year, particularly as interest in AI appears to be evolving. And I'm curious where de-identified data is sitting in the hierarchy of needs.
A: And we're seeing this migration where people aren't just licensing our data more and more, they're actually building a models with us, whether those are foundation models as is the case with AstraZeneca, or they're building smaller models, leveraging our data.
Q (Q1 FY2026, Matthew Shea): Is there anything in terms of either the recent Gilead or Merck deal that you would call out in terms of size or scope that's maybe different from some of your other strategic collaborations, or just anything notable to call out with those 2 wins in particular?
A: Merck was a very large strategic data and modeling collaboration, we have very large collaborations with people like AstraZeneca, GSK, and BMS. That's another very large collaboration of that magnitude.
Q (Q1 FY2026, Subhalaxmi Nambi): Could you break down for us what percentage of your data licensing comes primarily from Oncology? And what has come from other areas like rare disease, cardiovascular, longer term, where do you see that mix shaking out?
A: It has been nice to watch some recent wins in Neurology and in particular, we were just engaged to begin building a multimodal model in Alzheimer's disease that was a multi-million dollar project that we're in the middle of right now that we'll finish up middle of this year.
Q (Q1 FY2026, Mark Schappel): Eric, it was highlighted in the prepared remarks that roughly -- you have a 40% attach rate for your algos, I think it was on your Solid Tumor assays in Oncology. I was wondering if you could just break down a little bit further which products are driving the higher attach rates there?
A: We have a variety of algorithms that we've built over the years. Some of them, for example, are our Homologous Recombination Deficiency algorithm our Tumor Origin algorithm where in about 5% of cancer patients. We don't know the site of primary diagnosis.
Q (Q4 FY2025, Kallum Titchmarsh): The markets are a bit anxious around AI and how value is getting distributed within that ecosystem. And we're now obviously seeing kind of traditional AI players push into the healthcare sphere. So I'm curious how you feel your position is protected on the data side?
A: Tempus is uniquely positioned in that. We have both of those at scale. We have over 450 petabytes of connected multimodal data, which flows from our Diagnostic business, which has real-time insights, real-time connection to outcome and response, is able to track patients longitudinally, rich molecular data, rich imaging data.
Q (Q4 FY2025, Subhalaxmi Nambi): Earlier in 1Q, you launched Paige Predict. Given you previously discussed that you don't expect this to contribute meaningfully to revenue this year, can you discuss the strategic value of the added capability when samples are QNS?
A: Being able to digitize the pathology slide and render insights even when sequencing fails just makes our tests a little better than somebody else. The fact that we can also render those results in hours, makes us a little faster to deliver those insights.
Q (Q4 FY2025, Ryan MacDonald): Can you just give us an update on where things stand on the development of that foundation model? I think you mentioned last quarter, you're hoping to have the first version of the model ready in the first quarter of '26 here.
A: So the foundation model had a deliverable in Q1 where we had to hit certain benchmarks that AstraZeneca had established. We think we've hit all those benchmarks.
Q (Q4 FY2025, Unknown Analyst): Could you just kind of walk through what each of the main growth drivers will be for xT, xR, xF, xH and xE in 2026 and beyond?
A: Our core technology advantage drives the growth of all 5 of those assays.