Jensen Huang once asserted that life sciences represent the most profound application scenario for AI. With NVIDIA and Eli Lilly reaching a collaboration agreement for an AI joint laboratory valued at up to $1 billion, consensus among top-tier capital has been firmly established. NVIDIA's BIO26 biomedical special report surveys the life sciences ecosystem companies officially announced this year, revealing a clear trend of heavy investment in life sciences, alongside an evolving market perception of valuation for this sector: Anthropic's valuation approaches the trillion-dollar mark following its funding round, leading to the deployment of the Claude Science R&D platform; Chai Discovery secured $400 million in financing based on its biomolecular models and de novo antibody design capabilities, subsequently partnering with pharmaceutical giants including Eli Lilly, Novartis, and Bristol-Myers Squibb; Lila Sciences has seen its valuation pushed to $8.5 billion, driven by its "AI hypothesis plus robotic execution" science factory concept; and XTALPI (02228) has closed the loop between digital reasoning and physical verification, with Agentic AI directly dispatching robot clusters to complete the design-synthesis-test-feedback cycle for new molecules and materials in the real world.

This ecosystem map is evolving into a fortified AI4S industry landscape: Anthropic provides general reasoning, Chai strengthens biomolecular models, Lila and XtalPi explore autonomous laboratories, while companies like Dassault connect instruments with industrial scenarios. AI4S is moving beyond mere "model benchmarking" and entering a "productivity race" in the real world. In drug discovery, a single chemical hallucination can cost months of time and millions of dollars. Model parameters are merely an entry ticket; closed-loop experimental results from dry and wet laboratories are the true touchstone determining the market value of industrialization.

Beyond Project Collaborations: Exporting "Core R&D Infrastructure" to Global MNCs

Among the ecosystem partners showcased by NVIDIA, XTALPI is one of the few companies from China. Its distinctiveness lies not in having trained yet another drug model, but in having models, scientific agents, and robotic laboratories working together in unison. Today, XtalPi's "AI plus Robotics" system has transcended the concept validation stage, earning continuous validation and repeat purchases from Eli Lilly, the world's highest-valued pharmaceutical company. The collaboration trajectory between the two parties demonstrates strong strategic penetration: from the initial small-molecule drug discovery deal potentially worth $250 million, it expanded to a bispecific antibody partnership valued at $345 million; more recently, it completed the delivery of a multi-million compound management system for Eli Lilly's Shanghai R&D Center, along with the acceptance of its HTE (High-Throughput Experimentation) platform. Eli Lilly's continuously expanding repeat purchases—first acquiring small-molecule discovery capabilities, then leveraging the AI antibody platform, and finally directly procuring automation infrastructure—marks XtalPi's transition from single project collaborations to becoming a core R&D process infrastructure supplier endorsed by MNC repeat purchases.

This logic has been substantiated at the financial level. In the first half of 2026, excluding the impact of the high base from prior-year upfront payments, XtalPi's revenue grew 73.8% year-over-year. Among this, revenue from AI4S intelligent solutions reached RMB 193.5 million (a surge of 136.4% year-over-year), constituting half of the total. Quantified efficiency is XtalPi's core moat: Agentic HTE compresses the traditional 3-4 week experimentation cycle down to approximately 6 days; the SureRoute retrosynthesis system reduces the chemical hallucination rate to 4.6%, with a Top-1 recommendation accuracy of 74.3% (several times that of existing general-purpose large models). Concurrently, XtalPi has systematically disclosed its asset base: nearly 40 self-discovered and co-incubated pipelines, including over 10 pipelines at the IND or IND-ready stage, covering small molecules, antibodies, peptides, small interfering RNA (siRNA), and molecular glues. At this point, XtalPi's dual commercial engine has fully taken shape: externally, it exports "equipment plus platform," generating stable infrastructure revenue; internally, it mass-produces pipeline assets, capturing milestone payments and substantial upside from potential asset breakthroughs.

Insights from the "Next-Generation Blockbuster": Valuation Leap by Targeting Undruggable Targets

In August of this year, Revolution Medicines' molecular glue drug RASONQUE (targeting RAS) received FDA approval, with Phase III survival nearly doubling and death risk reduced by 60%, earning it the title of a true "next-generation blockbuster." The core takeaway is that the value of new modalities lies not in competing for market share within traditional red oceans, but in reshaping the boundaries of "what can become a drug." Once historically "undruggable" targets are conquered, the asset ceiling far exceeds traditional R&D outsourcing service fees. Greg Verdine, an early pioneer of the RASONQUE approach, later founded DoveTree and entered into a collaboration with XtalPi potentially worth up to $5.99 billion. XtalPi has already secured a total of $70 million in upfront payments, with the first oncology program advancing to the IND-enabling stage. This not only confirms the high druggability translation rate of the underlying technology but also clearly signals to the market the exponential BD potential of XtalPi's platform for future commercial expansion. This serves as both an endorsement of XtalPi's molecular discovery capabilities by top-tier experts and a validation of its underlying mechanisms. In certain projects, XtalPi's XGlue platform optimized target protein degradation activity to picomolar (pM) levels within a single quarter. In the context of innovative drugs, the ability to rapidly identify molecules that remain highly effective at extremely low concentrations significantly enhances the probability of clinical success in later stages.

Three Frontier Modalities and Valuation Anchors: From Single-Point Breakthroughs to "Underlying Capacity Encompassing All"

In the three frontier modality tracks of antibodies, peptides, and siRNA, breakthroughs in individual technology platforms have all commanded extremely high valuations in the capital markets. This provides the optimal reference frame for reassessing XtalPi's value:

Antibody Network: AI-driven design is transitioning from a technological narrative into the phase of clinical delivery. XtalPi, driven by its dual-core engine of "dry-wet loop closure plus top-tier veterans," is accelerating the realization of high commercial premiums for its proprietary pipelines. Chai, valued at $3.8 billion, and Generate Biomedicines, holding five clinical pipelines, validate the broad prospects of AI-powered antibodies. XtalPi's Ailux platform has not only integrated computational design with wet-lab developability assessment but has also been validated across over one hundred projects. Leveraging the bispecific antibody collaboration with Eli Lilly worth up to $345 million, its three autoimmune pipelines are directly targeting Phase I clinical trials in 2027. At this critical juncture, Ailux has recruited Dr. Maria G. Belvisi, former core management and CEO-3 from AstraZeneca, as its Chief Scientific Officer. As one of the most prominent cases of senior international pharmaceutical executives joining a Chinese startup, her top-tier industry operational vision will strongly propel pipeline advancement, truly enabling XtalPi to transition from "platform enablement" to "proprietary asset value."

Peptide Engine: The hundred-billion-dollar metabolic market has ignited a merger and acquisition frenzy for peptide assets. XtalPi addresses the core R&D bottleneck with automated synthesis and is launching a dimensionality reduction attack on the vast consumer functional molecule market. Eli Lilly's blockbuster peptide portfolio (projected to generate approximately $36.5 billion in 2025) and Roche's $5.3 billion heavy investment in Zealand Pharma confirm the robust cash-generation capabilities of peptide assets. The primary bottleneck in peptide development has always been the exceptionally high barrier of synthesis and screening. XtalPi's PepiX platform connects AI directly with automated synthesis, allowing certain oral cyclic peptide projects to lock in hit compounds in just two months, directly positioning at the pinnacle of upstream production efficiency for peptide assets while also expanding into consumer goods. Currently, XtalPi has one anti-hair-loss peptide and one food-grade peptide that inhibits carbohydrate absorption, both having passed INCI and FDA filings respectively and been approved for market entry, reshaping the consumer functional molecule market with underlying AI-driven pharmaceutical capabilities.

siRNA: Validation of precise targeting technologies has triggered MNC M&A deals exceeding ten billion dollars. XtalPi, with its high-hit-rate generalization model, confirms the cross-modality scalability of its underlying infrastructure. Novartis's $12 billion acquisition of Avidity and Alnylam's market value exceeding $33 billion establish extremely high valuation anchors for the siRNA field. Benchmarking against global giants, XtalPi has efficiently positioned six siRNA pipelines, with more than half having completed in vivo efficacy evaluations. Its siDiff model demonstrates a hit rate improvement of over 30% in tests on unseen target genes. The IgA nephropathy program obtained superior Non-Human Primate (NHP) efficacy data validation in just 9 months, a process that traditionally takes 12 to 18 months. This represents not just a breakthrough and significant efficiency gain for a single pipeline, but also proof to the market that its "AI plus experimentation" system can deliver overwhelming R&D efficiency when confronted with entirely new modalities.

AI cannot completely eliminate clinical risks, but it can move druggability assessment forward, allowing clearly flawed molecules to be filtered out early. As mature companies validate the commercial ceilings of new modalities, what XtalPi is competing for is the underlying general-purpose mass production capability for these all-modality assets. XtalPi's true scarcity lies in the fact that antibodies, peptides, siRNA, and molecular glues are all sharing the same underlying "AI large model plus robotics" foundational capability and infrastructure.

Reshaping the Valuation Model: From "Project Waves" to an "All-Modality Asset Factory"

It is increasingly difficult to value XtalPi using a single comparable company. For such a composite infrastructure enterprise, the market requires a new "SOTP (Sum-of-the-Parts)" valuation perspective: its AI4S automation infrastructure can be benchmarked against Lila Sciences, valued at $8.5 billion; its antibody platform can reference the premiums seen with Chai or Generate; and its molecular glue, peptide, and siRNA pipelines can be risk-adjusted valued based on development stage, partnership rights, and market potential. Combined with equipment sales, R&D services, high-value milestone payments, and a secure cash buffer of nearly RMB 8.7 billion, this constructs a moat of high safety margin and anti-cyclical resilience. Re-valuation is not about pricing all early-stage pipelines at their successful terminal values, but rather acknowledging the underlying shift in its value creation engine: moving from past project completions to looking at the compound growth of experimental equipment, data assets, and multi-modality pipelines.

The evolutionary logic of the AI4S industry is now unequivocally clear: the first phase competes on model parameters; the second phase competes on closed-loop experimentation; and the ultimate deciding factor is who can solidify experimental capabilities into an industrial-grade system capable of "continuously mass-producing high-value assets." NVIDIA's computing ecosystem is not short of stars, but companies that simultaneously possess "continuous MNC repeat purchases at the top level," "scalable automated laboratories," and an "all-modality pipeline reservoir" are extremely rare. The success of a single pipeline can only define the price of one transaction, but an AI industrial foundation capable of continuously producing pipelines will define the platform value of an entire era.