Boehringer Ingelheim Bets Up to More Than $1 Billion on Envisagenics’ AI Splicing Platform for Solid Tumors

What happened

Boehringer Ingelheim on Tuesday, Sept. 22, announced a partnership with Envisagenics, a New York-based biotech, to develop cancer therapies aimed at tumor-specific RNA splicing alterations. According to BioSpace, Boehringer could pay more than $1 billion in total, including an undisclosed upfront payment, research funding, option fees and development, regulatory and commercial milestones.

Envisagenics is also eligible for royalties on any products that reach the market.

Deal details

• Total potential value: More than $1 billion, including upfront, research funding, option fees and milestones.
• Royalties: Envisagenics may receive royalties on product sales.
• Division of work: The partners will collaborate on target validation studies. Boehringer can exercise options for exclusive licenses to certain targets and will lead modality selection, development and commercialization.
• Possible modalities: Multi-specific antibodies, T cell engagers and antibody-drug conjugates.

The collaboration builds on Envisagenics’ prior work identifying cancer-specific splicing events. BioSpace described the company’s platform as using AI to screen for and detect these RNA alterations in tumors. The partners describe splicing as one of the last largely untapped sources of therapeutic targets.

Envisagenics chief executive Maria Pineda said Boehringer is “focused on aggressive solid tumors with significant unmet need and limited treatment options.”

The science

RNA splicing is the process by which cells cut and rejoin segments of genetic messages before they are translated into proteins. When splicing goes wrong in cancer cells, it can produce abnormal proteins or protein fragments that are not found on healthy cells. In principle, those differences can be exploited as targets that let drugs distinguish tumor cells from normal tissue.

BioSpace noted that splicing dysregulation can drive uncontrolled cell division in some cancers. Other companies are pursuing the biology with different tools. Skyhawk Therapeutics develops small molecules that modify disease-causing spliced RNA and partnered with Merck KGaA in August 2025 on neurological disorders in a deal valued at more than $2 billion. Rgenta Therapeutics is developing small-molecule glues that tag improperly spliced RNA for destruction.

Why it matters for the sector

Many of the most successful targeted cancer drugs depend on a protein that is present on tumor cells but rare on healthy ones. The supply of such targets is limited, and many of the best-known ones, such as HER2 and TROP2, are already the focus of multiple competing antibody-drug conjugates. New target discovery is therefore a priority for large oncology companies that want differentiated programs.

The deal also reflects the continued flow of AI-enabled drug discovery partnerships. Large drugmakers have signed numerous agreements with AI-focused biotechs in 2026, generally with modest upfront payments and large milestone potential. These structures allow pharma companies to test platforms across multiple targets while limiting near-term spending.

Option-based licensing, where the buyer pays to take a target forward only after validation, is common in early discovery collaborations. It gives the platform company funding to generate data while letting the pharma partner focus resources on the most promising candidates.

For Envisagenics, a partnership with a large privately held drugmaker provides validation for its approach and non-dilutive funding to expand its work.

Undisclosed upfront payments are common in discovery-stage deals, and they usually indicate that most of the headline value is contingent. Observers therefore tend to focus on research funding and option fees, which provide the platform company with cash over the first years of a collaboration.

What to watch

• Which tumor types and targets the partners prioritize for validation.
• Boehringer’s option exercises, which would signal progress toward development candidates.
• Whether Envisagenics discloses the upfront amount or signs additional pharma partnerships.
• Further dealmaking in RNA splicing from companies such as Skyhawk and Rgenta.
• Broader momentum in AI-driven oncology target discovery deals.

Sources: BioSpace

Splicing in More Depth

Most human genes are written in pieces. The coding segments, called exons, are separated by non-coding stretches called introns. After a gene is copied into RNA, the cell’s splicing machinery removes the introns and joins the exons together to form the final message that is translated into a protein.

Splicing is not fixed. Through alternative splicing, the same gene can produce different versions of a protein by including or skipping particular exons. This flexibility is normal and important in healthy tissues. In cancer, however, mutations in splicing factors or broader changes in cell regulation can produce splice variants that are rare or absent in normal cells.

Why that creates targets

• New protein surfaces. An unusual exon junction can create a stretch of protein that does not exist elsewhere in the body, potentially recognizable by an antibody.
• Tumor-specific fragments. Abnormal proteins may be broken down and displayed on the cell surface, where they can be recognized by immune-based therapies.
• Selectivity. The more uniquely a target is found on tumor cells, the lower the risk that a potent drug damages healthy tissue.

How AI Fits Into Splicing Target Discovery

Finding cancer-specific splicing events typically involves analyzing large volumes of RNA sequencing data from tumors and normal tissues. The general workflow looks like this:

1. Data assembly. Sequencing data from many tumor samples and healthy reference tissues are gathered.
2. Event detection. Computational tools identify where exons are joined in unusual ways.
3. Prioritization. Models rank events by how often they appear in tumors, how rarely they appear in normal tissue and whether they produce proteins that drugs could reach.
4. Laboratory validation. Candidate targets are confirmed in cell lines and tissue samples, which is the stage the partners describe as collaborative work.

Machine learning can help with the scale of the search, but laboratory validation remains essential before any target is pursued.

The Three Modalities, Briefly

• Multi-specific antibodies bind two or more targets at once, which can improve selectivity or bring different cells together.
• T cell engagers bind a tumor target with one arm and a T cell with another, directing immune cells to attack the tumor.
• Antibody-drug conjugates carry a potent payload to cells that display the target.

Which modality suits a given target depends on how much of it is present on tumor cells, how it behaves after binding and how safely it can be hit.

Questions Readers Ask

What is target validation?

Experimental work confirming that a target is genuinely present where expected and that acting on it could plausibly affect the disease.

What is an option fee?

A payment the partner makes when it chooses to take an exclusive license to a specific target, typically after reviewing validation data.

Why does Boehringer lead modality selection?

As the party funding development, the larger partner usually decides how to turn a validated target into a drug, drawing on its own engineering and manufacturing capabilities.

This article is for general information only and is not investment or medical advice.