AI Drug Discovery Firms Pivot En Masse to 'Shovel Sellers' Amid Clinical Failures and Capital Retreat

Just two years ago,AIThe story of the pharmaceutical industry also unfolds around the grand narrative of "disrupting traditional R&D," with enthusiasm from the capital market and a flood of startups. However, the dilemma of new drug development, often described as "nine deaths and one life," has not been alleviated.AIThe arrival has changed.

Once highly anticipatedAIPharmaceutical Star CompanyVerge GenomicsRecently announced that itsAIDrugVRG50635Clinical Use for Treating ALSThe failure of the trial and the termination of R&D have triggered widespread attention.

ExceptVergeOfVRG50635Externally,2025Year8Month, Schrödinger targets acute myeloid leukemia(AML)DrugSGR-2921Development Halted Due to Toxicity Issues.2023Year, byExscientiaDesign, Sumitomo Pharma's Schizophrenia DrugUlotarontAlso ended due to unsatisfactory results.

Continuous clinical failures have acted as a reality check, forcing the entire industry to reassess its positioning. This also means that future evaluation criteria will become more straightforward: whether clinically validated effective drugs can be discovered.

Verge GenomicsHas announced a transformation and will no longer focus on internal drug research and development but will instead become a company that provides services for large pharmaceutical enterprises.AIA platform company for technical services. "This move makesVerge"able to continue investing in its artificial intelligence, product, and engineering capabilities, which will lead to a leaner cost structure and longer cash flow," the founder said.

Recently, not onlyVerge Genomics, includingRecursionIncluding multiple well-knownAIPharmaceutical companies have also strategically contracted or adjusted their internal R&D pipelines.

For example, Verge GenomicsRecursionUpon completion ofExscientiaAfter the acquisition, it announced the suspension or termination of multiple clinical-stage projects, focusing its R&D efforts on oncology and rare diseases.

Behind the shift is real financial pressure.VergeThe last round of financing remained at2021Year, the same year12Month Obtained9800Million USDBRound of financing. Even with collaborations with multinational pharmaceutical companies, the traditional cash-burning self-research model is unsustainable in the event of core asset failure and no new clinical pipeline to follow up.

To address financial pressures,RecursionThe company announced layoffs of approximately20%, the goal is to2025Annual cash burn controlled at4.5Less than $100 million,2026Year decreased to3.9Below US$100 million to ensure existing funds can support operations until2027End of the year.

These events reflect that the entireAIThe pharmaceutical industry is undergoing a profound and collective strategic shift—From high-risk "gold diggers" to more stable "water sellers".The industry bubble is being squeezed out, and a new stage of development driven by technology and services has already begun.

“AIIt's definitely a trend, but in the short term, the entire industry still can't see the direction clearly."The aforementioned investor admitted frankly.

"Despite the investors' confusion, they dare not miss the opportunity. The problem is, if they wait until the direction is clear before entering the market, the first-mover advantage will have already been taken by others."

This investor had previously missed the cryptocurrency cycle due to overconfidence——15Someone told him about blockchain before the year, and he thought he understood parallel computing and networking, so he looked down on this wave. "Missing an opportunity is not terrible, but missing it due to arrogance is truly regrettable."

In such an environment, investors are actually very distressed. The logic of many investors has now regressed to the most primitive step: choosing a team. Find a reliable team, regardless of what they specifically do, just invest first, similar to the early days of the internet — no one could clearly explain how to make money, so they invested in people.

To put it more bluntly, a lot of investments nowadays have turned into "investing in stories." Whoever tells a story with imagination, coherence, and believability, making people think it might be achievable, gets the investment first. Then this story is used to attract a team.What matters now is imagination more than ever."He said with a bitter smile."

But the story eventually has to return to reality. For the companies at the forefront of the “AI"Pharmaceuticals," he poured cold water: "At this stage today, it is almost still a 'false proposition.' A more accurate statement should be 'AI'Auxiliary Pharmaceuticals'.”

He believes that real pharmaceuticals still can't do without chemists.AIJust a tool. Want to completely rely onAI"Generating new drugs is still far from achievable in reality, as the current amount of fundamental data, such as chemical structures and biological characteristics, is insufficient to truly train a model capable of creatively generating new molecules."AI"The core bottleneck in pharmaceuticals is, in essence, still data accumulation." He concluded.

Specifically,AICan improve the efficiency of preclinical research and development, such as designing lead compounds within a few months, but cannot bypass complex biological mechanisms and high clinical risks. This is also the core reason why many molecules that perform well in preclinical studies fail.

During the drug design and conceptual phase,AICan quickly help with screening. But once it comes to the validation stage,AIPharmaceutical companies and all enterprises spend about the same amount of time, and there is no way to reduce the workload unlessAIThe accuracy can reach100%。

AIPharmaceuticals are essentially just a tool and cannot help pharmaceutical companies solve key bottleneck issues.Even thoughAIPharmaceuticals can improve efficiency and accelerate drug screening, butAIThe learning materials are based on experimental data already created by humans. For new target information,AIThere is still no way to create, nor can it help drugs complete the clinical trial phase.

In addition,AIModels rely on high-quality, standardized data for training, while biomedical data often face issues of "islanding" and insufficient standardization.

After obtaining the data, it is necessary to train the model based on the data to achieve algorithm iteration, butAIThe accuracy of the algorithm still needs improvement. "At present, purely artificial intelligence cannot create this drug; it still relies on the experience of scientists. It is merely a tool for scientists and cannot replace them."

When the capital market discoversAIThe role is relatively limited, and it cannot overcome the inherent complexity of biology in a short period of time, so this field has begun to be re-examined.

The current market changes are also very obvious: an increasing number ofAIPharmaceutical companies are starting to shift towards providing services because investors are no longer willing to pay for companies that have no cash flow in the long term. "The general bottom line for investors now is: if a company still can't generate cash flow within two to three years, they basically won't invest."

He pointed out,This is also why the overall development of new drugs has slowed down — the path to new drugs is inevitably a long-term investment.AIThe pharmaceutical industry has not fundamentally escaped this logic; it may have only shortened the time somewhat, but whether it can truly shorten it remains unknown."Therefore, now manyAIPharmaceutical companies are starting to actively seek service-oriented businesses to establish cash flow as soon as possible, in order to hedge against future uncertainties. How long this trend will last, no one knows."

The squeeze of capital and reality willAIPharmaceutical companies have been pushed towards the same exit:Provide services and become a "water seller."This collective shift is reshaping the industrial ecosystem of pharmaceutical research and development.

This transformation is essentially a reshaping of specialized division within the industrial chain.Start-upAIThe company focuses on the most cutting-edge and innovation-dense early discovery and design stages, acting as an "advanced R&D outsourcing hub" or "molecular design factory"; whereas large pharmaceutical enterprises leverage their comprehensive advantages to lead late-stage development, clinical trials, and commercialization, while shouldering the main financial risks and ultimate returns.

This transformation trend is fully embodied in Insilico Medicine.1Month5On [date], Insilico Medicine announced that it had reached a deal with Servier worth up to8.88A multi-year, multi-billion-dollar oncology drug development collaboration. It will lead the use of its self-developed artificial intelligence platform.Pharma.AIConduct early drug discovery and development, while Servier will lead the subsequent clinical validation and commercialization. This demonstrates its "AI"Platform Service Provider" Positioning.

Insilico Medicine adopts "self-developed pipeline+"Technology licensing" dual-driven, platform revenue becomes the replenishment for R&D. According to early investor Qiming Venture Partners, Insilico Medicine's "Pharma.AI"The platform has already partnered with the highest revenue-generating companies globally.20Chinese pharmaceutical companies13The cooperation with the JiaYou software platform proves the platform-based transformation.

Not long ago, Insilico Medicine2025Successfully listed on the Hong Kong Stock Exchange at the end of the year, with its Hong Kong public offering receiving approximately1,427Times oversubscription, representing the high valuation given by the capital market to its business model.

In addition, recently Huashen Zhiyao overseas companyEarendil LabsAnnounced a partnership with Sanofi worth up to25.6USD billion strategic cooperation, willEarendilThe discovery platform is applied to multiple autoimmune and inflammatory disease projects. Sanofi will lead the development and global commercialization of bispecific candidate drugs in this collaboration.

Some companies choose to focus on the most advantageous single link to provide in-depth services.For example,Relay TherapeuticsLeveraging its expertise in computational protein dynamics, it provides molecular design services for specific targets to its partners.

While advancing its internal pipeline, XtalPi has expanded its automated laboratories and intelligent R&D capabilities toAI+New materials and other broader fields, achieving diversified income.

AIPharmaceutical companies transition toCRO, driving the industry to shift from labor-intensive to technology-driven. This also poses challenges to traditionalCROChallenges for enterprises.

For traditionCROEnterprises, with their traditional medicinal chemistry and preclinical research services, are being impacted by “AI+The direct impact of the "automated experimentation" model. Some leadingCROHave to start vertical integrationAIAbility, or extend downward to production.

In ChinaCROCompanies, such as WuXi AppTec and ChemPartner, for example, through self-research and development/M&A (e.g., Pharmaron's controlling stake in Haizhi Wisdom) constructionAIPlatform, consolidate the entire process barriers, seize high-value innovative drug orders.

It is worth noting that, whetherAIPharmaceutical companies, orCROEnterprises that fail to enhance their innovation capabilities can easily fall into issues such as converging technological pathways,单一 commercialization models, and target crowding, ultimately leading to low-level price wars.Only those technology platforms that can truly tackle "undruggable" targets and significantly improve the success rate of molecular design possess sustained competitiveness.

It can be predicted that the next few years will beAIThe Shakeup Period for Pharmaceutical Platform Service Providers. Companies with solid technology, accurate positioning, and the ability to seamlessly integrate into pharmaceutical R&D processes will stand out.

As capital shifts from "indiscriminate speculation" to rational regression, and companies collectively pivot from "gold mining" to "selling water," this transformation—though driven by the pain of capital pressure and technological reality—may carve out a more sustainable path for the entire industry.