Key Takeaways
- Insilico Medicine announced a collaboration with Suzhou Ribo Life Science on May 17, 2026, to develop AI-driven oligonucleotide therapies, extending Insilico’s platform beyond its existing small-molecule pipeline.
- Isomorphic Labs secured $2.1 billion in funding on May 13, 2026, led by Thrive Capital, based on the commercial potential of its IsoDDE drug design platform and AlphaFold 3’s molecular prediction capabilities.
- Insilico Medicine’s Rentosertib reached Phase IIa in approximately 18 months at a cost of around $6 million, according to the company, compared to industry norms of 6-8 years and $100-200 million for the same milestone. Two events this fortnight put hard numbers behind AI drug discovery’s commercial momentum. On May 13, 2026, Isomorphic Labs closed a $2.1 billion funding round, one of the largest ever for a biotech company. Four days later, Insilico Medicine announced a partnership with Suzhou Ribo Life Science to extend AI-driven design into oligonucleotide therapies, a class of drugs that until recently sat well outside what AI platforms were built to handle.
Insilico Medicine: From AI-Designed Drug to Phase IIa Success
Rentosertib (ISM018_055) is the clearest proof point AI drug discovery has produced so far. Insilico’s lead candidate for idiopathic pulmonary fibrosis completed Phase IIa trials with results published in Nature Medicine in April 2026, showing a dose-dependent improvement in lung function: patients on a 60 mg dose gained 98.4 mL in forced vital capacity over 12 weeks, compared to a 20.3 mL decline on placebo. What makes that clinically interesting is the path to get there. Insilico says its Pharma.AI platform brought Rentosertib from discovery to Phase IIa in approximately 18 months at a cost of around $6 million, against an industry norm of 6-8 years and $100-200 million for the same milestone. Those figures come from the company and have not been independently verified, but even discounted they describe a material compression in timelines.
Insilico’s pipeline now spans oncology, fibrosis, immunity and age-related diseases, with wet lab facilities integrated into the platform to allow rapid experimental validation. Earlier in Q1 2026, the company announced a global R&D collaboration with Eli Lillypotentially worth up to $2.75 billion. This week’s Ribo partnership extends that ambition into oligonucleotide therapies, which target disease at the RNA level and require different chemistry than Insilico’s small-molecule work, a meaningful platform stretch, not just a deal announcement.
BenevolentAI: Repurposing Success and Deepening Pipelines
The clearest demonstration of what BenevolentAI‘s platform can do came early in the COVID-19 pandemic. In early 2020, its AI system flagged baricitinib, a rheumatoid arthritis drug, as a candidate that could reduce both viral entry and inflammatory response in COVID-19 patients. The computation took around 90 minutes. Human expert review followed, clinical trials were initiated, and the FDA eventually granted Emergency Use Authorization, putting the drug in front of patients well ahead of a conventional repurposing timeline.
BenevolentAI is now advancing internal programs in neurology, inflammation and oncology, including candidates for ulcerative colitis and ALS, with several programs in or moving toward clinical stages.
Atomwise: Accelerating Hit Discovery with AtomNet
Structure-based drug discovery has a well-known bottleneck: traditional high-throughput screening of physical compound libraries is slow, expensive and produces hit rates typically between 0.01% and 0.1%. Atomwise‘s AtomNet platform attacks that bottleneck computationally, using convolutional neural networks to screen virtual libraries against disease targets at a scale no physical lab can match.
A multi-year study published in Nature Scientific Reports put some numbers on AtomNet’s performance: across 318 diverse therapeutic targets, the platform achieved a hit rate of 5.3% to 7.6% in prospective screening, and identified novel compounds in around three-quarters of cases. The chemical search space available to AtomNet includes more than 15 quadrillion synthesisable compounds, which matters most for so-called “undruggable” targets where conventional libraries run thin. Atomwise has applied this to infectious disease, oncology and neurology, including a well-publicised collaboration with a university team that identified two existing drugs as Ebola candidates within days. The company continues to scale its computational infrastructure while advancing internal candidates across those same therapeutic areas.
Exscientia: AI-Designed Drugs Reaching Clinical Stages
In 2020, Exscientia’s collaboration with Sumitomo Dainippon Pharma produced DSP-1181, one of the first AI-designed drugs to enter human clinical trials. The compound, designed for obsessive-compulsive disorder, reached Phase I in 12 months. That was roughly four and a half years faster than the industry average for the same journey. DSP-1181 was later discontinued after Phase I, which is a routine outcome in drug development, but the milestone it marked was real: an AI platform had designed a novel molecule and navigated the full preclinical and regulatory path to human studies.
Exscientia was acquired by Recursion Pharmaceuticals in late 2024 in a $688 million deal. The merger combined Exscientia’s generative AI and precision chemistry capabilities with Recursion’s high-throughput phenomic screening platform. Several programs from the combined entity are now in Phase II, including candidates for familial adenomatous polyposis and ovarian malignancies.
Recursion Pharmaceuticals: Scaling Phenomic Screening with AI
Recursion’s approach starts with biology rather than chemistry. Its Recursion OS platform processes large-scale cellular imaging data, using deep learning to map how cells respond to disease states and potential drug candidates. The goal is to surface novel therapeutic hypotheses from biological patterns that conventional assays would miss, particularly in rare diseases, oncology and fibrosis.
Key near-term readouts include REC-1245, an RBM39 degrader currently in Phase 1/2 clinical trials, with Phase 1 safety and pharmacokinetic data expected in the coming months and additional dose escalation data anticipated in the second half of 2026.
Isomorphic Labs: Betting Billions on Next-Gen AI Drug Design
The $2.1 billion round Isomorphic Labs closed in May is a striking number even by recent biotech standards. The company was founded in 2021 as an Alphabet spinout, built explicitly to commercialise the structural biology capabilities that AlphaFold demonstrated. AlphaFold 3, released in May 2024, extended protein structure prediction to small molecules, peptides and antibodies, the full range of building blocks relevant to drug design.
Isomorphic Labs has developed its own platform, the IsoDDE (Isomorphic Labs Drug Design Engine), to translate those predictions into drug candidates. The company has said it is working on programs in oncology and immunology, with the intention of bringing AI-designed candidates into clinical trials, though it has not publicly disclosed lead therapies or timelines. Demis Hassabis, who leads both Isomorphic Labs and Google DeepMind, shared the 2024 Nobel Prize in Chemistry for the AlphaFold work, which gives the company’s scientific foundation unusual public credibility, separate from any commercial track record. Whether the $2.1 billion reflects genuine confidence in near-term clinical outcomes, or is a longer-duration bet on platform value, is harder to assess from outside. For a broader look at how AI is reshaping the relationship between research and clinical development, the recent coverage of AI agents compressing investigation timelines in financial services offers a useful parallel, the pattern of AI reducing months-long processes to days is appearing across regulated industries, not just pharma. For more coverage of AI research and breakthroughs, visit our AI Research section.
Originally published at https://autonainews.com/six-ai-drug-discovery-platforms-delivering-clinical-results-and-billions-in/
Top comments (0)