Bayer’s Iambic AI Drug Discovery Wager Targets the 90% Failure Rate

Bayer and Iambic Therapeutics have agreed to collaborate on AI-driven drug discovery, the German pharmaceutical group said on June 22, 2026. The deal gives Bayer access to Iambic’s two flagship models, Enchant and NeuralPLexer, for work on small molecules aimed at hard-to-drug targets. The agreement is the second major AI collaboration Bayer has signed this year.

The partnership is structured as an upfront payment plus milestone and royalty payments, terms the two companies did not disclose publicly. Bayer framed it as a way to translate cutting-edge technology into patient value through Iambic’s AI models.

What Bayer Just Agreed To

The collaboration was announced from Berlin and San Diego on June 22 (Bayer and Iambic’s June 22 deal announcement). Bayer will use Iambic’s AI-driven drug discovery platform, including its flagship models Enchant and NeuralPLexer, to find new drug entry points against targets that have resisted traditional chemistry. The deal focuses on small molecule discovery and aims to strengthen Bayer’s early-stage portfolio.

Juergen Eckhardt, M.D., head of business development and licensing at Bayer Pharmaceuticals, framed the deal as the kind of partnership Bayer wants more of as it modernizes its pipeline. “This collaboration exemplifies our shared ambition to harness AI as a strategic driver of innovation in drug discovery and our focus on building high-impact collaborations that translate cutting-edge technology into patient value,” Eckhardt said in the announcement. He also pointed to “combining complementary capabilities to accelerate scientific insights and improve decision-making across the entire R&D value chain” as the test of the deal. Eckhardt did not name a target, lead candidate, or financial term in the release. Tom Miller, PhD, co-founder and CEO of Iambic, gave the matching words from the other side.

Iambic demonstrates that better technology leads to better medicines, both in our own hands and with visionary partners like Bayer. Through this collaboration, Bayer secures access to industry-leading technology, including Iambic’s frontier AI models Enchant and NeuralPLexer, and together we expand the universe of potential life-saving medicines.

The 10-to-15-Year, $2.6 Billion Problem

The release leans on a familiar set of numbers about the pharmaceutical industry. Traditional drug discovery runs 10 to 15 years from idea to approved medicine and costs around USD 2.6 billion, yet more than 90 percent of candidates fail in clinical trials. That baseline is the structural backdrop that almost every modern AI drug-discovery pitch now opens with.

The Bayer-Iambic release argued that AI-based molecular optimization can identify differentiated hit molecules, shorten optimization timelines, and strengthen Bayer’s pipeline for the decade ahead. Iambic’s release highlighted that its platform “was validated by the discovery of a novel drug candidate advanced to clinic in about one-third industry-standard time.” A Drug Target Review analysis written for 2026 reads the year as “the first large-scale test of whether AI improves clinical success rates beyond the pharmaceutical industry’s persistent ~90 percent failure rate.” That same analysis adds that “additional clinical failures remain statistically likely given historical attrition rates.”

• 10-15 years: the typical time from program launch to first approval.
• ~USD 2.6 billion: the typical capitalized cost of bringing a drug to market.
• ~90%: the share of clinical candidates that never make it to approval.

Two industries, conventional pharma and AI drug discovery, are making the same bet from opposite directions. Conventional pharma believes its 90% failure rate is structural and so spreads risk across hundreds of programs at once. AI drug-discovery companies believe their tools will compress that rate by predicting clinical behavior earlier, before any patient is enrolled. The Bayer-Iambic agreement puts capital behind the second view, with the test still ahead in clinical readouts rather than in any release-day announcement.

Why Iambic Was Worth the Wager

Iambic Therapeutics was founded in San Diego in 2020 and has described itself as a clinical-stage life science and technology company. Iambic was named to CNBC’s 2025 Disruptor 50 list for what the company calls its track record of delivering AI-designed candidates into the clinic faster than industry norms.

The platform is built around two specific models. Enchant is a multimodal transformer that predicts clinical and preclinical endpoints from molecular structure, designed to flag failure modes before a candidate is dosed. NeuralPLexer predicts three-dimensional protein-ligand complex structures, and Iambic claims it is faster and more accurate than AlphaFold3. The two models are the assets Bayer is buying access to, and the specific performance claims are part of the negotiation Bayer signed off on.

Iambic’s most advanced drug candidate, IAM1363, is in a multi-center Phase 1b study for HER2-driven cancers, including HER2-mutated tumors. IAM1363 went from program initiation to investigational new drug application in two years, the company says, versus an industry average of six years. A second candidate, IAM-K1, a KIF18A inhibitor for triple-negative breast cancer and other solid tumors, is being advanced toward clinical candidacy.

Miller has used that track record to position Iambic against bigger AI-platform competitors. “While many companies are building AI models for drug discovery, Iambic stands alone in terms of successfully and rapidly delivering new medicines to human clinical trials,” Miller said at the time of the CNBC 2025 Disruptor 50 announcement. The clause that does work for Bayer is “delivering new medicines,” not “models.” The valuation argument leans on the clinical pipeline, not the architecture.

How the Bet Fits Bayer’s Wider AI Strategy

Bayer has put an internal R&D productivity target on its own AI program. Sai Jasti, SVP and head of data science and AI at Bayer, has said the company has an internal goal to increase R&D productivity by 40% in 2030, and he has acknowledged “we are still not there, especially on the research side.” The Iambic agreement is one of the steps Bayer is taking toward that target.

It is the second major AI partnership Bayer has signed this year, and the two deals do not overlap. In January 2026, Bayer signed a three-year strategic collaboration with Cradle, an Amsterdam-based AI protein engineering company, focused on antibody design and lab-in-the-loop optimization (Bayer’s January 2026 deal with Cradle for antibodies). The Iambic deal targets small molecule discovery at hard-to-drug targets. Cradle is for biologics; Iambic is for chemistry.

Bayer spent 5.8 billion euros on research and development in fiscal 2025, when the group reported sales of 45.6 billion euros and around 88,000 employees. The R&D budget and the Iambic deal sit on the same balance sheet: Bayer is buying optionality on faster small-molecule discovery, and the Iambic agreement is one specific wager that two models can do what thousands of chemists have not yet done at scale. Bayer is one of several top-tier pharma groups now making such a wager.

Where the Bet Stays Unproven

The release-day framing leans on early-stage optimism, but the harder test still sits ahead. The Drug Target Review analysis for 2026 reads the year’s Phase III readouts as “the first large-scale test of whether AI improves clinical success rates beyond the pharmaceutical industry’s persistent ~90 percent failure rate” (predictions on AI drug discovery clinical tests in 2026). Its author adds that “scientific commentators have questioned whether AI fundamentally improves clinical outcomes” and that AI-discovered compounds so far show “progression rates similar to traditionally discovered molecules.” The Phase III data may end up showing faster timelines without improved efficacy, which the same analysis calls “a commercially valuable but scientifically underwhelming outcome.” Until Phase III data lands from AI-discovered assets, the Iambic deal sits in the early-stage category the rest of the field already occupies.

The deal’s structure also carries that caution. Industry analysis in the same Drug Target Review piece describes a roughly 50:1 ratio between announced “biobucks” in AI drug-discovery deals and actual upfront payments. Bayer has not disclosed the upfront figure for the Iambic agreement.

1. The deal’s value to Bayer depends on whether Iambic’s models generate differentiated candidates, not just faster ones.
2. The high Phase 1 success rates reported for AI-discovered drugs come from a small sample of AI-native biotechs.
3. Phase 3 readouts in 2026 and 2027 will determine whether AI improves success rates beyond the ~90% failure pattern.

The wider 2026 picture offers both bigger bets and smaller ones. In March 2026, Eli Lilly signed an expanded research collaboration with Insilico Medicine that includes $115 million in upfront and could reach approximately $2.75 billion in development, regulatory, and commercial milestones, plus tiered royalties. The Drug Target Review piece also notes that “multiple companies shut down entirely despite substantial backing; others announced 20 percent+ workforce reductions and several pursued delisting.” The Bayer-Iambic agreement is a smaller, milestone-driven wager by comparison, but it is the same kind of bet the wider industry is placing through 2026.

Frequently Asked Questions

What did Bayer and Iambic agree to?

On June 22, 2026, Bayer and Iambic Therapeutics announced a drug discovery collaboration focused on small molecules. Bayer will use Iambic’s AI platform, including the Enchant and NeuralPLexer models, to find candidates for hard-to-drug targets. Iambic will receive an upfront payment plus milestone and royalty payments under terms the companies did not disclose.

What do Enchant and NeuralPLexer do?

Iambic describes Enchant as a multimodal transformer model that predicts clinical and preclinical endpoints from molecular structure. NeuralPLexer predicts three-dimensional protein-ligand complex structures, and Iambic says it is faster and more accurate than AlphaFold3.

When will drugs from this collaboration reach clinical trials?

Neither company has published a development timeline for assets produced under the deal. Iambic’s existing lead candidate, IAM1363, is already in a Phase 1b study for HER2 cancers; new programs generated under the Bayer collaboration are expected to be earlier-stage.

How does this compare to Bayer’s other AI deal with Cradle?

The two deals target different parts of the drug pipeline. The Cradle agreement, signed in January 2026 for three years, focuses on antibody design and lab-in-the-loop protein optimization. The Iambic agreement focuses on small molecule discovery for hard-to-drug targets.

Why is the 90% clinical failure rate the number that matters?

It is the structural backdrop every modern AI drug-discovery pitch opens with. Whether the most advanced AI-discovered drugs can compress that rate is the question Phase 3 readouts through 2027 are now positioned to test.