Sanofi Goes All-In On AI

On Tuesday, June 13, 2023, a very important press release hit the wire – a major pharmaceutical company announced its plans to go all-in on AI. Considering how grand and impactful the news was, I expected a massive wave of top-tier media coverage but after a week, only a few specialized pharmaceutical media outlets picked up the news.

So here it is – Sanofi, one of the world’s leading global pharmaceutical companies, issued a press release explaining its AI strategy and announcing the launch of its company-wide AI system called “plai” with a very bold press release titled “Sanofi “all in” on artificial intelligence and data science to speed breakthroughs for patients.” The title is very unusual for the pharmaceutical industry, which usually opts for more conservative titles for press releases that require months of legal reviews, so this really stood out as a big deal. Another reason why this announcement is such a big deal is the lengthy and detailed quote from the company’s young superstar CEO, Paul Hudson, with no other executives or partners quoted.

From Zero to Leader in AI in Two Years

“Our ambition is to become the first pharma company powered by artificial intelligence at scale, giving our people tools and technologies that focus on insights and allow them to make better everyday decisions. The use of artificial intelligence and data science already support our teams’ efforts in areas such as accelerating drug discovery, enhanced clinical trial design, and improving manufacturing and supply of medicines and vaccines. We have just scratched the surface as to how we embrace these disruptive technologies to achieve our ambition of transforming the practice of medicine”, Paul Hudson, CEO of Sanofi.

In January 2023, during the industry’s largest annual event, the JP Morgan Healthcare Conference in San Francisco, I was invited to a lunch Sanofi hosted for its AI partners. During this lunch, Paul Hudson, his trusted R&D general, Dr. Frank Nestle, and other Sanofi executives, sat at different tables and discussed their plans to go big on AI. What impressed everyone most was how hands-on Paul Hudson was with AI. Literally, hands-on, as he was already actively using the plai app and explaining the integration of generative AI across the organization. He also explained that the plai app is designed and developed in collaboration with a Munich start up – AILY LABS. I hope to cover the AILY LABS story in the future articles as its fearless female co-founders, Bianca Anghelina, CEO and Sara Bisbe López, Chief of AI have very interesting backgrounds.

The lunch included all the companies mentioned in the press release, including Dr. Tomas Clozel, a dear friend for many years, the CEO of Owkin, a global but proudly French AI company specializing in clinical trial analysis that Sanofi did the biggest partnership and investment deal with in 2021. This deal included a $180 million investment into Owkin which instantly gave the company a unicorn valuation per Owkin’s announcement. Among other guests were Dr. Abraham Heifets, CEO of Atomwise, the AI company using deep learning for molecular screening and drug discovery, and your humble servant, Alex Zhavoronkov, the founder of Insilico Medicine, a generative AI company specializing in end-to-end drug discovery. Both Atomwise and Insilico partnered with Sanofi in 2022. The partnership with Insilico, in addition to the drug discovery deal, included complete access to Insilico’s generative AI suite, which we have used internally to nominate thirteen preclinical candidates and advance three therapeutic programs into human clinical trials. This means that Sanofi now has the entire end-to-end generative AI platform, which in addition to Owkin’s capabilities, can rapidly solve diseases and potentially set new records in drug discovery. Sanofi also partnered and extended its collaboration with Exscientia, which made nice progress with its internal pipeline.

Even in January 2023, it was clear that Sanofi made a giant leap in AI compared to its big pharma competitors. Even the partner choice was mostly spot on. While the technology sector, in general, experienced a major boom due to the consumerization of generative AI, the so-called AI-powered drug discovery industry experienced rapid consolidation. The “AI biotechnology” companies that either did not have strong technology, were created by the investors to access financial markets, or jumped into the public markets prematurely, have contracted, went out of business, or were acquired. BenevolentAI failed in Phase II, cut the lion’s share of its staff, and it became apparent that the much-popularized target discovery deal with AstraZeneca announced in April 2019 only progressed in chemistry and lead-optimization in April 2023 according to its pipeline. Cyclica and Valence were acquired by Recursion, and LabVantage picked up BioMax. However, most of Sanofi’s new partners made solid progress, and only Atomwise cut staff to focus on its pipeline.

In under two years, Sanofi went from a technology follower in AI to one of the clear leaders in generative AI, and the press release that went mostly unnoticed by the top-tier media provided a clear summary of its recent achievements and a glimpse into its overall strategy.

AI Initiatives at Sanofi:

plai, a company-wide AI-powered “everything app”: plai collects internal company data across various functions to deliver real-time data interactions, insights, and enable scenario planning for decision-makers in the company.

AI in Drug Discovery and Clinical Research: AI and data science are being used to accelerate drug discovery, enhance clinical trial design, and improve the manufacturing and supply of medicines and vaccines. Sanofi has multiple AI programs in clinical research to cut the time, improve outcomes, and automate time-consuming activities. AI has also improved potential target identification in therapeutic areas like immunology, oncology, or neurology by 20% to 30%.

AI in mRNA Research: Sanofi uses AI to select lipid nanoparticles essential for the delivery of mRNA vaccines, reducing the time required for this process from months to days.

AI in Clinical Operations: Digital tools and insights from plai are used to improve the design and execution of clinical trials, including the selection of trial sites that can increase the representation of underrepresented communities in clinical research.

AI in Manufacturing and Supply: Sanofi has digitized its quality assessment processes and implemented AI-enabled yield optimization solutions that help optimize the usage of raw materials and support cost efficiency. Adopting plai has enabled the prediction of 80% of low inventory positions.

Acquisition of Amunix Pharmaceuticals: Acquired in 2022, Amunix uses AI to tailor-deliver medicines that become active only in tumor tissues, sparing normal ones.

Partnership with Exscientia, Atomwise, and Insilico Medicine: These partnerships utilize AI-driven platforms to accelerate the development of new medicines.

Partnership with Owkin: Owkin’s AI-driven platform uses patient data to build models and predict patient treatment responses.

Partnership with Hillo: Sanofi partnered with French startup Hillo to adapt its digital twin AI solution for Sanofi’s connected insulin pens.

Launch of the Open Innovation Portal: Sanofi’s Consumer Healthcare unit launched this portal to invite external entities to respond to identified challenge areas in consumer health, facilitating the creation of innovative self-care solutions.

In AI, Timing is Everything

Regarding investments in AI, the best time was 2014, when the deep learning revolution started. The next best time is today. I would say the same about autonomous robotics and quantum computers. However, when it comes to promoting these technologies publicly and implementing organization-wide changes, timing is everything, and only in 2022 did we see the consumerization of generative AI, which demonstrated irrefutable evidence of its potential. The only company that manages the timing of technology trends perfectly is NVIDIA. Over the years, its charismatic leader Jensen Huang demonstrated an impeccable ability to time technology trends for investment and promotion, turning NVIDIA into a trillion-dollar company.

However, premature promotion of technology may backfire. Industry veterans clearly remember how IBM promoted its Watson AI through Jeopardy! Games and massive investments in PR and marketing. And while it is not apparent from the carefully-curated Wikipedia page, STAT was among the first to report the challenges with the platform in 2018, and in 2021-2022 top-tier media, including the New York Times and Slate, highlighted that the system did not live up to the hype. And at the end, IBM decided to sell Watson Health to Francisco Partners. IBM seems to have learned its lesson, and is now much more careful with promotion of its quantum computing initiatives. Despite making massive progress with quantum computers and releasing a working quantum computer for commercial use by its partners, IBM did not engage in excessive hyping of the quantum technology.

Can Sanofi Succeed Where Others Stalled?

In the pharmaceutical industry, Novartis was one of the companies that seems to have made a bet on AI prematurely. In 2018, Novartis appointed Vas Narasimhan, a McKinsey alumnus who was only 41 as a CEO. Almost immediately, Novartis included AI in every investor call, started sponsoring inspirational events focused on AI, and hired hundreds of data scientists. According to ClinicalTrials Arena, a recent report by GlobalData highlighted that out of the 50 biggest employers in the pharmaceutical industry, Novartis was the company which referred to artificial intelligence the most between October 2020 and September 2021. GlobalData identified 39 artificial intelligence-related sentences in the company’s filings – 0.4% of all sentences. A simple LinkedIn search shows over 1,500 people who are or were working for Novartis with AI in their profiles. By 2020, it felt like Novartis had more heads of AI than the largest AI drug discovery companies had AI scientists. I remember coming and presenting there several times but it seemed like some of them were constantly on vacation or working on AI strategy, trying to fill the newly-created head counts, or saying that they don’t know who to partner with because they did not understand the landscape of AI startups and they need to run pilots. Unfortunately, as the number of AI groups increased internally, the internal AI teams prioritized internal platform development to partnering with expert companies specializing in AI platforms and integrating end-to-end AI systems. In 2018, it also appointed Bertrand Bodson, the former Chief Digital and Marketing Officer at the retailer Sainsbury’s, as Chief Digital Officer, who also joined the Executive Committee (EC), the main decision making body of the organization. In December 2021, Bodson departed, and in 2022 Novartis announced substantial global job cuts – 8,000 out of its 108,000 staff, with up to 1,400 in Basel, Switzerland. The number of job applicants to the leading AI companies suggested that many of them were in AI and data science. Sanofi’s Paul Hudson was the CEO of Novartis Pharmaceuticals from 2016 to 2019, and he probably saw these challenges and had a chance to learn the mistakes and best practices.

Elli Lilly also put some substantial funding behind its push into AI. In 2019, Lilly made a deal with Atomwise to deliver drugs for 10 targets with $1 million upfront for each target. The results of this collaboration were not yet disclosed. Just over a year ago, in May 2022, Lilly partnered with a new company founded in 2019 by a recent Stanford graduate, Genesis Therapeutics, with an upfront payment of $20 million on three initial targets with the option to nominate two more. It remained undisclosed if Lilly would acquire the software from the company. Usually, it takes one to two years to nominate a preclinical candidate for a known target, and soon, we should be able to review the progress. And in May 2023, Lilly partnered with China Shenzhen-based XTalpi, in a deal valued at $250 million, to discover medicines for a disease with no available medicines.

While I would estimate the probability of success of Lilly’s partnership choices to be very low, it quietly boosted its internal capabilities. It recently recruited the AI/ML and computational drug discovery superstar, Dr. Jie Shiye, out of UCB, who is intimately familiar with the top-performing generative AI platforms and developed the first antibody discovered using AI all the way into the clinic. He started building the state-of-the-art end-to-end discovery infrastructure.

In contrast to Novartis and Elli Lilly, Sanofi’s push in AI seems to be well-timed, carefully orchestrated, results-oriented and focused on partnerships with industry leaders with proven technology and future-proof generative AI capabilities. Even the press release announcing that it is “all-in” on AI was not promoted to the mainstream media. It looks like they decided to make a date stamp but not popularize it everywhere.

So while several pharmaceutical companies are leveraging AI in their R&D only at Sanofi did I observe a CEO with an AI-powered R&D app. I am very much looking forward to following the progress in this industry.

Is there an Optimal Partnership Strategy with AI Drug Discovery Companies to Increase Pharmaceutical R&D Productivity?

In my opinion, the best partnership strategy for any pharmaceutical company with any AI company is at the preclinical candidate stage when the AI company has already de-risked the target and demonstrated first-in-class or best-in-class chemistry. At this stage, the pharmaceutical company can rapidly design or co-design the clinical trial pathway and combination strategy that is maximally beneficial to its own pipeline. This strategy gives the CSO and the CEO of the pharma company the ability to see the results of their work in late-stage human clinical trials or even on the market in just a few years. However, due to the current decision-making structures and structures of business development and research organizations, such partnerships are very rare. We are hoping to see more of these deals that can substantially advance the AI drug discovery industry.

In summary, only time will tell how much AI will will impact the big pharmaceutical companies and how much it will impact R&D productivity. Over the past 20 years, we witnesses a revolution in sequencing, multiomics, bioinformatics, the Internet, mobile and other fields; however, the R&D productivity of big pharma did not increase. But the fact that Sanofi, one of the world’s leading pharmaceutical companies, announced that it went “all-in” on AI is a major industry milestone to be date-stamped and followed as living history.