Jintai Holdings' latest financial report: While AI is still in testing, the "first AI pharmaceutical stock" is already making money.

Ask AI · Does Jingtai’s Profit Prove a Turning Point Toward Commercialization of AI for Science?

While the AI industry is still debating monetization paths for large models and many AI-listed companies continue to struggle in losses, Jingtai Holdings (2228.HK), founded in drug R&D, has already used its financial reports to demonstrate another possibility for AI commercialization.

In 2025, all of Jingtai Holdings’ business segments achieved rapid growth. Total revenue reached RMB 803 million, up 201.2%; annual profit was RMB 135 million, with adjusted net profit of RMB 258 million. It delivered full-year profitability for the first time, becoming the first profitable company in the HK-listed AI for Science sector.

What does this mean? At the turning point where AI moves from “capability verification” to “result delivery,” Jingtai becomes the first company to hand in the completed answer.

From AI to AI Agents, Physical AI

According to research reports from Frost & Sullivan and East China Securities, the global AI for Science market size is expected to grow from about USD 800 million in 2023 to USD 5.9 billion in 2030, with a compound annual growth rate (CAGR) of 32.1%. Among them, the penetration rate of drug discovery AI is still at an early stage—meaning the track is wide enough, but there are only a limited number of players who can actually run the track.

In June 2024, Jingtai Holdings, as the first technology company to list on the Hong Kong Stock Exchange under Rule 18C, entered the capital markets. It also became China’s “No. 1 AI pharmaceutical company.”

After one and a half years of development, Jingtai’s AI progress remains ahead, and its technology evolution path is clearly visible: from early AI-assisted drug design, to today’s AI Agents for autonomous decision-making, robotic execution, and then to Physical AI’s deep understanding of the material world.

Compared with traditional R&D models: In the traditional drug R&D process, advancing a single experimental project often requires going through dozens of steps such as target screening, molecular design, synthetic route planning, experiment execution, and data analysis, involving cross-functional collaboration among multiple roles including chemists, biologists, and data scientists. The cycle can easily take months or even a full year. Meanwhile, based on industry data, in a traditional laboratory, a chemist can typically complete only 10 to 20 compound synthesis experiments in one week, and the data are scattered and difficult to reuse.

In Jingtai’s lab, this process is being rebuilt by AI and robots.

The company has built an industry-unique R&D system of “AI models + robotic laboratories + Multi-Agent”: over 200 industry AI models act as “expert brains” responsible for key steps such as target解析, molecule generation, and virtual screening; a robotic laboratory operates 7×24 hours for high-throughput experiment execution, accumulating 200k+ high-quality reaction data entries every month and covering more than 80% of common medicinal chemistry reaction types; Multi-Agent autonomously breaks down R&D objectives, schedules resources across the entire process, and forms a self-evolving R&D flywheel.

_ (Jingtai R&D flywheel diagram) _

According to the introduction, this system can independently and autonomously advance tens of thousands of compound synthesis experiments every week, collecting data at 40 times the efficiency of traditional labs. This means that what used to take months can now be delivered on a weekly basis.

It is also worth noting that in Jingtai’s technical architecture, Multi-Agent is not just a simple automation script, but a true “intelligent project manager.” Each module has clear responsibilities: X-buddy analyzes targets; PatSight performs real-time searches of global patent literature; MolAgent handles molecular design; Vast Agent conducts route and raw material analysis. The robotic orchestration core assigns experimental tasks to robots for execution.

When a R&D task is input into the system, these Agents immediately start collaborating—autonomously decomposing the objective, coordinating with the procurement department to stock supplies, directing robotic arms to adjust experimental parameters, analyzing spectral data, and automatically switching to alternative solutions when anomalies occur.

_ (Jingtai multi-agent diagram) _

This “soft-and-hard integrated” system forms Jingtai’s deepest moat. A company that only does algorithms cannot validate its predictions, while a company that only does automation lacks intelligent scheduling capability. Jingtai connects the two.

Today, this capability has been transformed into mature products that Jingtai outputs externally. During the reporting period, Jingtai reached a cooperation with a certain international top big pharma company to deploy an autonomous conditional screening system for it; it jointly built a new-generation AI-driven molecular simulation platform with Pfizer, and independently developed the protein drug generative AI platform XenProT™; it also actively laid out frontier areas such as molecular glues, peptides, nucleic acids, and virtual cells.

It is worth noting that Jingtai’s AI blueprint has already gone beyond the boundaries of biomedicine and spilled over into the broader field of material science.

In January 2026, Jingtai signed a strategic cooperation agreement with Jingke Energy (688223.SH), a subsidiary of Jinko, a leading integrated photovoltaic company. Together, they will build the world’s first full closed-loop stacked battery manufacturing line of “AI decision-making—robot execution—data feedback.” It is expected to reach an experimental throughput of about 1,000 wafers/day, representing a 100-times-plus efficiency improvement over current experiments. In the consumer health sector, Jingtai developed two innovative topical efficacy ingredients for hair growth and hair-firming needs: Remeanagen™ and the peptide AquaKine™. Their combined formulation product, Groland, ranked No. 1 on Tmall International’s new products list for comparable products. In the chemical sector, BASF has adopted Jingtai’s formulation stability testing intelligent work.

This cross-domain transfer capability demonstrates Jingtai’s platform generality and scalability—when AI Agents can seamlessly switch from managing drug discovery pipelines to other industry domains, the boundaries of Physical AI are being redefined. More importantly, each successful cross-domain implementation accumulates new data assets for Jingtai, further strengthening its technology barriers.

Big pharma orders keep coming, and incubation moves into the harvesting period

Unlike the project-based model of traditional CROs where “stop working and stop funding,” Jingtai’s platform-based services are highly reusable. Once the underlying R&D system is validated by big pharma, marginal costs continue to decline. Repeated training also makes the R&D flywheel run even more smoothly—this is the key to its ability to achieve profitability while maintaining high growth.

Relying on the industrialized validation and continuous iteration of its “AI + robots” platform, Jingtai has secured multiple high-value collaborations across small molecules, large molecules, and frontier therapy tracks. Top global pharma companies such as Eli Lilly, Pfizer, Roche, and DoveTree all appear on Jingtai’s customer roster. In the drug field, the relevant targets cover multiple high-value areas including oncology, autoimmune diseases, and neurological diseases.

Customer stickiness and repeat purchases are the core validation of platform value. Pfizer and Jingtai have partnered for ten years, including participating in the R&D of the blockbuster drug Paxlovid. In 2025, they again announced a new cooperation to continue deepening the construction of the AI-driven molecular simulation platform. Eli Lilly renewed its cooperation in two rounds, with a total cooperation amount of USD 345 million; the payment structure includes profit-sharing from future sales. During the reporting period, Jingtai also reached cooperation with Roche and Haleon, delivering multiple sets of systems and R&D platforms.

Big pharma’s continuous repeat purchases and additional orders have brought stable and substantial revenue to Jingtai Holdings. In 2025, Jingtai’s drug discovery solution business revenue surged significantly from RMB 104 million in 2024 to RMB 538 million in 2025, up 418.9%.

After its strategic upgrade, AI for Science (AI4S) intelligent solutions (formerly the “intelligent robot solutions” business) achieved rapid growth—rising from RMB 163 million in 2024 by 62.6% to RMB 265 million in 2025.

_ (Jingtai AI robot laboratory workstations) _

At the same time, Jingtai’s business model has upgraded from a single project service to a “platform licensing + joint development + milestone revenue” strategy.

For example, its independently developed AI peptide R&D platform PepiX™ has cooperated with Gan & Lee Pharmaceuticals. Jingtai will receive platform licensing fees, an upfront payment, milestone payments for clinical and commercialization, and a share of pipeline re-licensing revenue. This means Jingtai’s revenue curve is shifting from linear growth to nonlinear growth: supported by service fees in the short term, driven by milestones in the mid term, and locked in via profit-sharing in the long term.

Even more noteworthy is that Jingtai’s “incubation universe” is entering the harvesting phase. Unlike pure investing, Jingtai has both pipeline returns and equity holdings in incubated companies. For some projects, it also continues to add investment to maintain its ownership stake. Through R&D collaborations, it protects the pipelines—forming a composite returns model of “pipeline revenue + equity + ongoing investment.”

For example, the incubation company JiTai Technology, which is set to pursue an IPO in Hong Kong, has completed Phase III clinical trial primary endpoints for its MTS-004 targeting dysregulated emotional control in pseudo-bulbar affect (PBA). It is expected to be the first AI formulation new drug in China to complete Phase III, filling a domestic gap. Leeman Bio continues to innovate CAR-T therapies in areas such as relapsed/refractory lymphoma and leukemia. In clinical trials, it achieves 100% complete remission at an extremely low traditional dosage of 1‰, and is expected to bring a RMB 1.2 million “sky-high” anti-cancer drug into an affordable era; it has already obtained FDA clinical approval. SIGEX1194? For HiG? (Ignore) — SIGX1094 from His^? Actually: “希格生科的全球首个弥漫型胃癌靶向药SIGX1094预计2026年第三季度进入Ⅱ期临床，泛TEAD抑制剂SIGX2649即将提交IND申请。” (Keep as-is translated.)
In addition, SIGX1094, the world’s first targeted drug for diffuse gastric cancer by Xige Biosciences, is expected to enter Phase II clinical trials in the third quarter of 2026, while the pan-TEAD inhibitor SIGX2649 is about to submit an IND application.

Xili Technology, Merda Bio, and others are also making rapid progress. The pipeline META-001-PH for primary hyperoxaluria by the latter has obtained FDA orphan drug designation. The IBD oral small-molecule inhibitor MP-5342 is expected to start clinical trials in the fourth quarter of 2026. The combined global market size for the related indications totals several tens of billions of dollars.

As incubation collaboration pipelines advance further, they are expected to bring Jingtai a steady stream of high-value late-stage returns. This composite model of “pipeline revenue + equity appreciation” allows Jingtai to both share in the growth dividends of ecosystem enterprises and avoid the risks of heavy-asset pipeline investment. It is also highly likely to be replicated at scale.

Valuation upside opens up: the value of an AI platform that’s the first to become profitable

In 2025, the rapid development of artificial intelligence technology drove AI’s upgrade from an “efficiency tool” to an “innovation engine.” Combined with the reconstruction of industrial infrastructure by AI + robots, it has accelerated the expansion of related markets.

At present, valuations of domestic AI large-model-related companies are generally pinned on high expectations. As of March 24, SenseTime’s market cap is about HKD 77.2 billion; Zhipu’s total market cap is HKD 292 billion; MiniMax’s total market cap is HKD 323 billion. By comparison, Jingtai Holdings—whose self-developed large-model and “AI + robots” flywheel effects continue to become more apparent—has a valuation of HKD 39.1 billion, with substantial room for growth.

Looking back at classic cases from the internet era, Tencent achieved profitability in the industry first by monetizing its instant messaging business. It was that injection of real cash that enabled subsequent investments in games, expansion into social platforms, and entry into payments. Ultimately, through a series of precise acquisitions and ecosystem expansion, Tencent grew into a technology giant with a trillion-dollar market cap.

Its core logic was this: in emerging industries, once the first company achieves commercialization and proves its model is sustainable, it will form an insurmountable moat through network effects and data accumulation, ultimately resulting in a “winner-takes-all” landscape.

Now Jingtai Holdings’ sector is experiencing a similar turning point—most AI pharmaceutical companies are still struggling along the break-even line. Compared with global peers, the global AI drug discovery leader Schrödinger generated revenue of USD 256 million in 2025 (+23.3%), with a net loss of USD 103 million. Another company, Recursion, which tried to form a stronger position through mergers with Exscientia, still reported losses exceeding USD 100 million in a single quarter after the 2025 merger and was forced to cut nearly half of its R&D pipelines.

By contrast, Jingtai not only achieved profitability for the full year first, but also avoided the risks of heavy-asset pipeline investment with its platform-based model. As of December 31, 2025, the company’s total cash balance was RMB 200k. In 2026, the net proceeds from newly issued convertible bonds were RMB 7.07B, providing ample support for ongoing R&D spending, technological iteration, and investments in M&A.

This advantage of “making money first” is turning into strong momentum for M&A integration and domain coverage. After acquiring Liverpool enantiomeric chemistry technology company LCC in 2025, Jingtai has already shown its ambition to complete its technology landscape through capital operations. In terms of platform reusability, its “AI + robots” R&D system is essentially a general methodology for solving physical-world problems. The technical capabilities accumulated in the pharmaceutical domain can be migrated to areas such as materials, chemical engineering, and new energy at low marginal costs.

Currently, its combined formulation products in the consumer health sector have been launched. The perovskite stacked battery developed together with a subsidiary of Jinko Energy is expected to move toward large-scale production over the next three years or so, and Jingtai is entitled to share in the joint commercialization benefits. In the materials domain, another state-owned central enterprise has also signed a cooperation with Jingtai for an intelligent R&D platform for catalytic materials worth tens of millions of RMB.

Jingtai Holdings states that the company will continue to deepen technology iteration of its flywheel effect, while also pushing the flywheel effect to accelerate spillover into broader areas such as consumer health, photovoltaic new materials, and advanced chemicals—realizing cross-domain empowerment from biomedicine to physical sciences, and achieving sustainable, high-quality value growth.

While the market is still pricing the “imagination space” of general large models, Jingtai has already proved the “monetization capability” of AI for Science with real money. And as policy dividends from the “14th Five-Year Plan for the AI + robotics” release, Jingtai—one of the first to run commercialization—may replicate Tencent’s path of “profitability—M&A—ecosystem expansion,” becoming a platform-level giant in the AI for Science sector.

_ This article is for reference only and does not constitute any investment advice. _