AI Biotechs Secure Over $500M in Deals Within a Day: Is the Spring of AI Drug Discovery Finally Here?

On December 5, Sanofi and AbbVie announced new collaborations with AI pharmaceutical companies on the same day.

Sanofi Signs Multi-Year Research Collaboration Agreement with French Pharmaceutical Company Aqemia to Develop Small Molecule Candidates for Several Undisclosed Therapeutic Areas. Aqemia will leverage its artificial intelligence (AI) platform to design drug molecules and use physics-based computations to generate data, accelerating drug discovery. Sanofi will be responsible for wet lab research, development, and commercialization activities. Under the terms of the agreement, Aqemia is eligible to receive up to $140 million in upfront and milestone payments across all programs.

AbbVie and BigHat Announce Collaborative Research to Discover and Develop Next-Generation Therapeutic Antibodies in Oncology and Neuroscience. BigHat will leverage its Milliner^TMPlatform: This is a machine learning technology integrated with high-speed wet labs, designed to guide the design and selection of high-quality antibodies for multiple therapeutic targets. Under the terms of the agreement, BigHat Biosciences will receive a $30 million upfront payment and may be eligible for up to approximately $325 million in research and development milestones, as well as potential additional commercial milestones and royalties on net sales.

After weathering the storm of controversy, the surviving AI pharmaceutical Biotechs are no longer pursuing the "pipeline sales" path to monetization but are instead moving towards "research and development collaborations." Has their spring arrived?

Looking back at the two collaborations, it is not difficult to find that the two Biotechs involved in this cooperation are still focused on the early-stage areas of drug molecule/antibody development and design. However,Unlike the early "pipeline-selling" approach, collaborations with pharma are now characterized by large-scale, long-term, in-depth partnerships described as "multi-year," "multi-field," and targeting "multiple therapeutic goals."

Taking Sanofi's "AII in AI" as an example, whether it is the multi-year, multi-target research collaboration with Insilico Medicine and Aqemia, or the cooperation with BioMap based on large models to jointly develop drug discovery models, all are long-term, wide-ranging strategic collaborations.

For Biotech, strategic cooperation is no longer a one-off deal but a deep collaboration where no single move determines the outcome, offering opportunities for trial and error as well as improvement. On the one hand, Biotech will gain access to pharma’s resources, collaboration, and potential data-sharing opportunities, updating and iterating its own platform. As seen in the partnership with Aqemia, Sanofi will leverage its global R&D expertise while independently handling wet lab research, development, and commercialization activities. On the other hand, with multiple pipelines advancing simultaneously, once a single drug pipeline succeeds, Biotech can also earn royalties and corresponding milestone payments.

For Pharma, in-depth cooperation is undoubtedly a long-term investment with less risk and lower cost. The key question is, how can AI pharmaceutical Biotech stand out and come into the sight of Pharma?

Unlike AI platforms that require experimental data for training, Aqemia utilizes 12 years of research from Oxford University, Cambridge University, and ENS/CNRS.Unique Quantum Physical Algorithm, generate its own data, and handle drug discovery projects from the initial stages. Aqemia claims that its deep physics algorithms achieve 10,000 times the speed, effectively guiding generative AI to rapidly and scalably discover drug projects, introducing the innovative engine Launchpad. In collaborations with Sanofi, Janssen, and Servier worth millions of euros, multiple drug pipeline research initiatives have been launched based on Launchpad results, with a particular focus on oncology and cancer immunology.

BigHat's Milliner^TMThe platform integrates machine learning technology,High-Speed Wet Lab Integrated with Synthetic BiologyFor a complete antibody discovery and engineering platform. After platform iteration and machine learning model design, BigHat can use synthetic biology technology for multiple tests in its own laboratory. Based on cellular or other functional assays, BigHat replicates in vivo disease processes to measure the biophysical properties of each variant and their impact on disease activity, and updates the model for iterative predictions.

AI-Driven Biotech No Longer Chooses Blind Expansion but Sells Models and Seeks Collaboration; Pharma Builds In-House AI Teams While Actively Partnering with Outstanding, Promising AI-Driven Biotechs. As Track Transactions Return to Calmness, It Represents Both Rational Development After the Bubble Burst and the Start of Destigmatization and Healthy Competition.

Looking back at the mid-to-late 2010s, Google DeepMind successively developed AlphaGo, AlphaZero, and AlphaFold1. Particularly, AlphaFold2 made a significant leap in protein spatial prediction. Artificial intelligence and machine learning solved century-old problems in the pharmaceutical industry, elevating them to near-mythical status. For a time, people believed that AI and ML could solve all existing issues in the pharmaceutical industry, even evolving to the point of drug generation.

The widespread talk, myths, and hype have created a massive bubble in AI-driven drug discovery — the field is booming, capital is flowing in, companies are emerging, and the stories of pipeline discoveries and technological prowess are dazzling. But the pitfalls were there from the start: when it comes to practical applications, how much of an impact can AI-driven drug discovery really have on the pharmaceutical industry? Beneath the myth of cost reduction and efficiency improvement, what are the success rates of drugs reaching Phase II and Phase III clinical trials? In no time, the bubble of AI-driven drug discovery has burst, turning "myth" into "stigma."

The bursting of the bubble is also a process of track selection and iteration. The FDA and EMA have put forward clear definitions of AI/ML, and the current consensus is: AI/ML is only an auxiliary tool that can help the development of the pharmaceutical industry.

Looking back at the companies in the current track: underlying algorithm innovation, AI and genetics innovation jointly driving drug development, AI drug discovery integrated with wet labs, AI applied to clinical trial development, AI chemical route design + digital lab management......After returning to rationality, AI/ML is gradually moving towards more vertical, more professional, and closer to industrial-end application directions.

Previously, Dr. Tao Du, former senior review officer of the FDA and chairman of Shenzhen Egret Pharmaceuticals, mentioned in an exclusive interview with VCBeat that the role of AI/ML in the future pharmaceutical industry will become increasingly significant.Therefore, AI technology in the pharmaceutical industry is not overheated, but rather not mature enough, or it can be said that the coverage of China's AI pharmaceutical industry is still not wide enough.Given the current situation, AI is still in an early stage in the pharmaceutical industry. In the future, through close collaboration between various AI service providers and traditional pharmaceutical companies, the application of AI in the drug development process will become increasingly widespread.

This year, Insilico Medicine and XtalPi have submitted their IPO applications, striving for a listing on the Hong Kong Stock Exchange. AI pharmaceutical companies that are moving towards industrial applications and collaborative win-win partnerships are embracing more possibilities, staying rational, and filled with anticipation.