Yingxi Intelligence announces $2.75 billion partnership with Eli Lilly; CEO responds to acquisition rumors: currently not considering a sale

On March 30, Tigermed Smart (HK03696, share price HK$58.7, market cap HK$33.6 billion) closed up 2.62%. The previous night, the company released its 2025 annual report. Due to a year-over-year decrease in the upfront payment for pipeline development revenue, the company’s annual revenue was $56.24 million. As of the end of 2025, the company had cash and bank balances of $393 million.

At the same time, the company announced a BD (business development) deal with Eli Lilly: an upfront payment of $115 million, with a total amount of up to approximately $2.75 billion. The transaction sets a new record for the highest cooperation amount in China’s AI pharmaceutical industry, drawing outside attention to the stage of development of AI pharmaceuticals.

“When I joined the company in 2021, it was hard to persuade domestic pharmaceutical companies to embrace AI. The real shift began with improvements in AI technical capabilities, which led the industry to develop a deeper understanding of AI’s applications in the pharmaceutical field.” At an earnings call held on the morning of the 30th, the company’s co-Chief Executive Officer and Chief Scientific Officer Ren Feng said that the urgent breakthrough needed for AI pharmaceuticals is “validation”: first, Phase 3 clinical validation; second, validation for AI automated laboratories.

The proportion of “best-in-class” and “first-in-class” projects is 8:2

In 2025, Insilico achieved revenue of approximately $56.2 million, including $25.0 million from drug discovery, $23.9 million from pipeline development, $4.9 million from software solutions, and $2.5 million from other discovery.

Compared with 2024, the company’s revenue structure showed significant changes. Specifically, drug discovery revenue surged 693.6% year over year to $24.952 million, and its share of total revenue jumped from 3.7% to 44.4%. Pipeline development revenue fell from $76.589 million to $23.885 million, with its share dropping from 89.2% to 42.5%. Software solution revenue grew steadily by 23.8% to $4.913 million.

On the expense side, Insilico’s R&D expenses decreased 11.4% year over year to $81.38 million, which the company attributed to lower third-party CRO (pharmaceutical research outsourcing organization) fees.

Besides the company, the industry is also changing. 2025 was a year of divergence in the global AI pharmaceutical industry. Overseas established “AI+SaaS” company Schrodinger fell into a double slump in both performance and share price due to its shift to in-house R&D. Meanwhile, Recursion, a representative company of self-developed pipelines, saw its losses widen and was cleared out by Nvidia. These developments made the outside world realize that AI pharmaceuticals cannot, in the short term, change the industry’s “decade of R&D for a billion-dollar outcome” rule in the pharmaceutical sector.

At the earnings call, Ren Feng also said that new drug R&D from clinical submission to listing in clinical phases I to III falls within a highly regulated category, so AI cannot accelerate it. Therefore, the company focuses its target on the stage from target discovery to preclinical candidate drugs (PCC). This stage can be iterated rapidly using computation data-driven approaches, but once clinical submission research begins, it is constrained by regulation and time.

Ren Feng explained that, to balance risk, the company has reduced the proportion of First-in-Class (first-in-class) projects to 20%, with 80% being Best-in-Class (best-in-class) projects—balancing innovation and practicality. In the future, it will continue to strengthen its BD capabilities, starting with collaboration with Eli Lilly, aiming to achieve revenue balance and even profitability.

In addition, the company is also developing its software licensing business. The annual report shows that the number of subscription customers increased from 153 to 181 in 2025. The Pharma.AI platform’s commercialization potential in the industry is being released step by step. At the same time, the company is expanding platform applications into non-pharmaceutical fields such as advanced materials, agriculture, nutrition products, and veterinary medicine.

Reuniting with Eli Lilly again, but the “AI+Biotech” positioning remains unchanged

According to the introduction, the cooperation model between Insilico and Eli Lilly in this round includes two aspects. For external licensing, it covers globally exclusive licenses for orally administered therapeutic drugs that are in the preclinical stage and have “best-in-class” potential. Joint R&D means that both parties will carry out multiple R&D collaborations around the targets selected by Eli Lilly, combining Insilico’s Pharma.AI platform and Eli Lilly’s industry expertise.

In fact, the earliest trace of the cooperation between Insilico and Eli Lilly goes back to 2023. At that time, both sides initially agreed on a Pharma.AI software cooperation, and Eli Lilly began trialing Insilico’s AI tools. In 2025, the two sides officially started a drug R&D collaboration. Eli Lilly participated in Insilico’s IPO on the Hong Kong Stock Exchange at the end of that year, becoming the first cornerstone investor for a HK IPO.

Ren Feng also disclosed at the earnings call that the two sides plan to explore deeper levels of synergy in the future, including collaboration on the MMAI Gym platform (a training framework used to train foundation model expertise specialized in biomedicine). Eli Lilly has ample Nvidia chip computing power, which may enable further cooperation on compute power and AI platform applications, with expectations to achieve a fourth round of collaboration.

Reporters noted that MMAI Gym is a platform launched by Insilico in January 2026, aiming to convert foundation models with causal reasoning capabilities into high-performance engines capable of handling real-world drug discovery and development tasks. Unlike the Pharma.AI platform, which is aimed at pharmaceutical companies, MMAI Gym’s direct target audience is not pharmaceutical companies but foundation model makers; pharmaceutical companies are the end application users.

Based on this, industry insiders believe Insilico’s business model centered on “out-licensing drug discovery R&D projects” is going to change. However, in an interview on the 30th, Ren Feng clearly told reporters that the company’s main business model will not undergo major changes. The company’s core business model at present is still “AI+Biotech” (an AI biotech company), rather than simply “AI+CRO” (an AI contract research organization).

“We do both ‘sell shovels’ (meaning providing AI platform services and pipeline licensing to customers) and ‘dig our own gold mine’ (meaning independently pushing forward the core pipeline R&D).” In Ren Feng’s view, these two lines of business run in parallel and are mutually reinforcing. The company will always use pipeline licensing and partnerships driven by AI as its core sources of revenue, while continuing to deepen its positioning as a platform-based biotech to provide collaboration services to customers.

The U.S. market accounts for 56% of revenue; addressing rumors of Eli Lilly’s acquisition

Reporters noted that Insilico’s overseas revenue share is extremely high. The U.S. market is its largest revenue source, and last year this portion accounted for about 56.14%. China is the company’s second-largest market, with last year’s revenue share at about 33.06%.

When interviewed by reporters, Ren Feng provided a detailed breakdown of the differences between domestic and overseas pharmaceutical companies in the AI pharmaceutical field. He said that overseas pharmaceutical companies are more willing to embrace advanced technologies. Not only are their willingness to pay and capabilities higher, they are also more inclined to accelerate AI R&D progress through diversified collaborations. Domestic pharmaceutical companies, by contrast, are relatively cautious, with lower willingness to pay and capabilities, and are more inclined to build in-house AI teams to carry out related work.

At present, considering the characteristics of the domestic market, Insilico has already explored adapted collaboration pathways and launched a cooperation model of “co-owning intellectual property.” Specifically, in the early stages, domestic pharmaceutical companies can make upfront payments lower than the level of overseas collaborations, but the company obtains more rights and interests related to projects, balancing short-term gains through the realization of future value. Ren Feng believes this is an effective approach that fits the characteristics of the domestic market at present.

As for market rumors that Eli Lilly would acquire Insilico, the company’s founder and CEO Alex Zhavoronkov made a clear response.

He said that the rumor mainly stems from speculation triggered by the review by the U.S. Federal Trade Commission (FTC). However, if the company’s existing clinical pipelines and molecular generation capabilities are taken into account, a reasonable valuation should be far higher than the current level. Therefore, the company is currently not considering selling.

（Source: Economic Daily News）