Boehringer's $15M Bet That the Next PD-1 Is Hiding in Your T Cells
Boehringer Ingelheim is teaming up with AI-biotech Immunai in a $15M deal to hunt for the next PD-1: unknown T-cell targets hiding across cancer and autoimmune disease. It's a small bet that could reshape how we think about immune checkpoints.
The biggest drug class in cancer history started with a simple idea: take the brakes off T cells and let them attack tumors. PD-1 inhibitors turned that idea into roughly $42 billion in annual sales. Now Boehringer Ingelheim wants to find the next set of brakes. And it's hiring an AI company to do the searching.
Boehringer just signed a multi-project discovery collaboration with Immunai, an AI-biotech startup that has built what it calls the world's largest single-cell immune database. The deal is worth up to $15 million, runs through 2027, and has one goal: find novel targets on dysfunctional T cells that could lead to new drugs for both cancer and autoimmune disease.
Small dollar figure. Potentially enormous implications.
Why PD-1 Isn't Enough Anymore
Immune checkpoint therapy is one of the great success stories of modern medicine. Block PD-1 or CTLA-4, and some patients' immune systems wake up and destroy their tumors. The problem? Only about 20% of cancer patients respond to these drugs. The rest have T cells that are exhausted, suppressed, or confused by mechanisms we don't fully understand yet.
The industry has tried to fix this. LAG-3 made it to market (relatlimab plus nivolumab), but the gains were incremental, not transformational. TIGIT, once the darling of next-gen checkpoints, has produced mixed results in late-stage trials. TIM-3 is still mostly in early testing. The co-stimulatory approach (hitting the gas pedal instead of releasing the brakes) is promising but complicated by toxicity concerns with targets like 4-1BB.
In short, the field has been playing whack-a-mole with individual targets. Boehringer and Immunai want to take a fundamentally different approach: use massive datasets and AI to find targets nobody has thought to look for.
Immunai's Secret Weapon: A Billion-Cell Atlas
Immumai isn't your typical AI drug discovery shop. While many AI biotechs focus on designing molecules, Immunai focuses on understanding the immune system at an almost absurd level of detail.
Get tomorrow's biotech intelligence before your competitors.
Join thousands of biotech professionals who start their day with our free, daily briefing.
Their core asset is AMICA, a curated immune cell atlas that integrates data from over 300,000 patient samples across more than 500 diseases. The atlas doesn't just catalog which genes are turned on in each cell. It layers multiple types of data on top of each other: gene expression, surface proteins (more than 80 markers per cell), T-cell receptor sequences, and spatial information about where cells sit in tissue.
Think of it like Google Maps for the immune system. Most researchers have been navigating with a paper road map. Immunai built satellite imagery, street view, and real-time traffic data, all at once.
On top of AMICA sits their ImmunoDynamics Engine, an AI layer that links immune cell behavior to clinical outcomes. It can spot patterns across diseases that human researchers would never notice, especially patterns that are shared between cancer and autoimmune conditions.
That cross-disease insight is exactly what Boehringer is paying for.
The Collaboration Playbook
The deal unfolds in two phases. First, Boehringer and Immunai will build a large data foundation by analyzing T cells from thousands of patient samples spanning multiple cancer types and autoimmune diseases. Immunai's AI will sift through this data looking for shared dysfunction patterns: molecular signatures that appear when T cells go haywire, regardless of whether the underlying disease is melanoma or lupus.
Once the AI identifies promising candidates, phase two kicks in. Immunai will run functional experiments in its New York lab to confirm that the targets actually matter biologically. Are they druggable? Do they drive disease? Can you flip them on or off and change outcomes?
Targets that survive this gauntlet get handed to Boehringer's drug discovery machine. The deal includes an option to expand if the science delivers, and negotiations on what that expansion could look like are reportedly already underway.
Boehringer's Bigger Picture
This isn't a one-off experiment for Boehringer. The German pharma giant has been quietly assembling one of the most aggressive immunology portfolios in the industry, with a deliberate strategy of blurring the line between oncology and autoimmune disease.
Consider the recent deal sheet. A €1.1 billion collaboration with OSE Immunotherapeutics on SIRPα (a myeloid checkpoint) that started in cancer and expanded into metabolic liver disease. A roughly $357 million pact with Cue Biopharma for selective B-cell depletion in autoimmune conditions. A partnership with CDR-Life on trispecific T-cell engagers designed to "reset" the immune system. An up-to-€640 million licensing deal with Kyowa Kirin for a first-in-class autoimmune small molecule.
The Immunai deal fits this pattern perfectly. Boehringer is betting that cancer and autoimmune disease are two sides of the same coin: T cells that attack too little in cancer, and T cells that attack too much in autoimmunity. Finding the shared regulators could unlock treatments for both.
Why This Matters Beyond Boehringer
Immumai has quietly become one of the most connected AI platforms in pharma. AstraZeneca signed on first, back in 2022, and has expanded repeatedly; the relationship now spans oncology, inflammatory bowel disease, and a recent $37.5 million extension into 2027. Teva joined in late 2024. Bristol Myers Squibb signed a multi-year deal in early 2026.
Boehringer makes it four major pharma partners, and the trajectory tells a story. These aren't curiosity purchases. AstraZeneca is using Immunai's platform on its phase 3 volrustomig program, a PD-1/CTLA-4 bispecific antibody. BMS is feeding clinical data through the system across multiple programs. The technology is moving from "interesting experiment" to "embedded infrastructure."
That mirrors a broader industry shift. In 2025, AI drug discovery partnerships generated $43.4 billion in total potential deal value across 114 deals. Sanofi inked a $2.56 billion AI partnership with Earendil for autoimmune targets. Eli Lilly signed an exclusive platform deal with Insilico Medicine worth up to $2.75 billion. The numbers keep getting bigger because the early results keep encouraging more investment.
The Bottom Line
At $15 million, the Boehringer-Immunai deal won't make anyone's top-ten list for deal size. But size isn't the point. The point is what it represents: a bet that the next great immune checkpoint isn't going to be discovered by a grad student staring at Western blots. It's going to be found by an AI chewing through hundreds of millions of immune cells, connecting dots across diseases that nobody thought were related.
PD-1 was a revolution. The question now is whether the next revolution is hiding somewhere in that data, waiting for the right algorithm to pull it out. Boehringer is paying $15 million to find out. If it works, that'll look like the bargain of the decade.
4 min read
Lantheus Had the Data. The Factory Broke the Dream.
Lantheus' cancer imaging agent had flawless clinical data and a clean safety profile. The FDA rejected it anyway, and the reason has nothing to do with science. It's a cautionary tale about the hidden risk that keeps torpedoing biotech approvals.
