Insilico Medicine’s Strong Start to the Year Bodes Well for AI Drug Approval First

Less than a month into 2025 Insilico Medicine has been peppering the AI drug discovery space with much needed good news. The Boston-based group has in the last three weeks announced positive data from two phase I studies, a licensing agreement, the start of a phase I study and a $100m fundraise.

Insilico’s most recent new year gift for its investors was yesterday’s news of the start of a phase I trial for ISM6331, a novel pan-TEAD inhibitor, for the treatment of mesothelioma.

ISM6331, which has been granted orphan drug designation from the FDA, is Insilico’s most advanced in-house oncology programme. TEADs play a key role in tumour growth and tumour resistance, so inhibiting this process could produce compounds that combat drug resistance, making them more potent cancer treatments.

The group also announced last week a second collaboration with Menarini Group, that included an upfront payment of $20m. The licensing deal for an AI-discovered cancer drug comes less than 12 months after Menarini’s subsidiary, Stemline Therapeutics, in licensed preclinical KAT6A inhibitor, MEN2312, for metastatic ER+/HER2- breast cancer for $12m upfront.

The speed by which Menarini was able to take MEN2312 into the phase 1 trials – 10 months after licensing– is likely a factor in the Italian group deciding to return for a second deal.

This time round, including the $20m upfront, Insilico could receive up to $550m and tiered royalties if all milestones are met and the unnamed preclinical product gains clinical approval.

Leading the AI Late-Stage Pack

The licensing news follows Insilico’s success with its inhouse pipeline. Earlier this month the group announced positive results from two phase I studies of ISM5411, a novel gut-restricted and PHD specific Inhibitor, for inflammatory bowel disease (IBD).

The 124-patient trial conducted in Australia and China showed ISM5411 was generally safe and well tolerated, with no reports of serious adverse events. Insilico now plans to take the drug into phase IIa trials in the second half of 2025.

Insilico has one of the most advanced AI drugs in development. Last year the group’s proprietary idiopathic pulmonary fibrosis project INS018_055 completed phase IIa trials. INS018_055 is a novel Traf2- and Nck-interacting kinase (TNIK) inhibitor discovered by Insilico’s target identification engine, PandaOmics.

The group then used Chemistry42, its generative AI chemistry engine, to design the molecular structure of the small molecule. Phase III trials are expected to begin this year.

Alongside INS018_055, Insilico has nine other projects with investigational new drug application (IND) status. Three of these have been out licensed to pharma groups Exelixis, Fosun Pharma and Menarini. In December Insilico revealed that it had received its first clinical milestone payment from Exelixis of $10m for XL309 a treatment for solid tumours.

Funding the Future

Insilico’s development plans for the rest of its pipeline have been given a boost by the $100m series E venture round the group secured in the first week of January. Shanghai’s Pudong Venture Capital and Hong Kong’s Value Partners Group jointly led the round, existing US investor Warburg Pincus also participated.

The money will be used to enhance Insilico’s artificial intelligence models, establishing and upgrading automated laboratories, and advancing its pipeline. Since inception Insilico has raised over $450m in venture financing.

The Lucky One

Insilico’s recent good fortune is in stark contrast to other first-generation AI drug discovery companies. Benevolent AI suffered the mid-stage failure of its dermatitis drug in 2023, torpedoing plans to out license its topical pan-Trk inhibitor. Similarly, Exscientia, another first-generation AI drug company, announced in October 2023 it would be winding down its phase I/II study of cancer drug EXS-2154.

This lack of clinical success has seen these AI drug discovery companies either consolidating their operations or undergoing major restructuring programmes. Over the last two years Benevolent AI has announced a series of cost-saving initiatives, including pulling down the shutters on its US business and implementing swingeing job cuts.

The recent 180-person headcount reduction in December came just two months after founder Ken Mulvany returned to the London-based group as part of a boardroom shake up. Mr Mulvany has set out plans to delist the company from Euronext Amsterdam and return to the group’s “Techbio mission” as part of his turnaround strategy.

Exscientia took the step of merging with fellow AI drug discovery group Recursion in August last year after numerous difficulties including; Bayer ending a $243m partnership, the scaling back of its pipeline in in 2023 and the departure of its CEO last February. Ahead of the merger, which was widely seen as a rescue of Exscientia, the UK-based group axed a quarter of its staff to save costs.

Second generation AI companies have also suffered setbacks. In December Valo Health’s ROCK inhibitor for diabetic retinopathy failed in a phase II trial, as did BioAge Lab’s obesity drug.

Failure to Launch

But Insilico has not been without its own troubles, the group has twice failed to list on the Hong Kong Stock Exchange, with its latest prospectus expiring in September.

It has also been almost 11 years since Insilico’s launch in 2014 and its most advanced product is only now entering phase III. The next two to three years could prove pivotal if the group wants to claim the prize for the first AI generated product to reach the market.

Pulling off a successful IPO would help, but with phase III trial patient recruitment expected this year, a fresh round of funding, and ongoing pharma partnerships Insilico, is better placed than most to get across the winning line.

However, if it fails in this endeavour the questions about hype versus reality in AI drug discovery will only grow stronger.