The AI sector was jolted on June 8th by a significant policy document from the National Data Administration.
The "Implementation Plan for Advancing the Construction of High-Quality Industry Datasets" marks the first national-level, systematic deployment of data to empower AI development, explicitly identifying real-world interaction datasets and embodied intelligence data as key construction areas to promote the assetization of vertical industry data.
The high-quality datasets for physical interaction and environmental sensing encouraged by the new policy directly target AI firms whose core data originates from real-world experiments.
Under this national guidance, a clear industrial trend is emerging: data is set to become the most critical asset for the future of AI. Companies capable of continuously generating high-quality experimental data from the physical world are no longer mere "service providers"; they are evolving into "infrastructure operators who own data oil fields."
XTALPI (ASX: 02228) stands out as a leading player with deep moats in this arena, noted as the "first AI+ robotics stock." Its decade-long development of a Physical AI experimental infrastructure precisely meets the core demands of the new policy.
Re-evaluating XTALPI through this lens suggests its asset classification may need reassessment. The company has spent over a decade building not merely an AI drug discovery platform, but a "production line for experimental data from the physical world"—AI model prediction, wet-lab robotic execution, data feedback for model iteration, and Multi-Agent intelligent scheduling.
This complete R&D loop achieves three things: making experiments replicable, allowing data to accumulate, and enabling experience to compound. The development of over 200 specialized AI models covering small molecules, peptides, biologics, molecular glues, and oligonucleotides indicates this production line has achieved standardized, multi-category delivery at a factory-scale capacity, not just lab samples.
This loop supports the standardized delivery of the company's two core businesses—drug discovery and future chemistry—enabling the sustainable generation of scarce scientific experimental data for the industry. The continuous accumulation of this data, in turn, iteratively improves AI models and boosts R&D efficiency, creating a self-reinforcing positive flywheel.
What is the commercial essence of this loop? It represents a paradigm shift in scientific R&D from "craftsmanship" to "industrialization." This model of generating high-quality data from the physical world aligns perfectly with the policy's encouragement of real-world interaction and embodied intelligence datasets. Consequently, robotic laboratories are upgraded from mere R&D tools to new infrastructure that continuously produces high-value industry data.
XTALPI's AI+robotics labs have pioneered standardization and scale in the field of scientific experimentation, which has long relied on an "apprenticeship" model. The new national dataset construction policy provides, for the first time, a direction for institutional exploration into the assetization and circulation of vertical industry scientific data, opening a valuation and monetization pathway for vast proprietary experimental data to transition from "internal fuel" to "tradable asset."
For a long time, the massive experimental data accumulated by XTALPI was used only for internal model iteration, with no market valuation for this asset, leading to persistent undervaluation. Models like data trusts and asset securitization explored by the new policy are transforming the question from "who owns high-quality industry data" to "who owns balance sheet assets that can be priced."
This shift in perspective could impact valuation frameworks far more than short-term earnings fluctuations. More importantly, the value of data lies not in static stockpiles but in "living data streams."
XTALPI's flywheel structure—more experiments generate more data, more data strengthens models, stronger models drive more experimental demand—ensures data is not a one-time fossil fuel but a self-regenerating resource. In 2025, the company's chemistry service renewal rate exceeded 75%, and its partner network covers global top-tier pharmaceutical firms, providing a ready ecosystem for data sharing and collaborative innovation.
Repeat customer orders mean each collaboration reinforces the company's competitive moat, rather than being a one-off project delivery. This "stronger with use" scalable benefit structure is extremely rare in the B2B services sector.
Cross-enterprise collaboration based on standardized experimental data gives XTALPI a formidable competitive advantage in the high-certainty field of AI for Science.
Financial data confirms the flywheel has passed the break-even point. In 2025, the company achieved revenue of RMB 803 million, a year-on-year increase of 201.2%, and profit attributable to owners of RMB 135 million, making it the first profitable full-year company in the Hong Kong-listed AI for Science sector.
More crucially, year-end cash reserves exceeded RMB 7 billion, meaning the company does not face a trade-off between "investing in R&D" and "survival," allowing the flywheel to accelerate without funding concerns.
Recently, positive catalysts have materialized in quick succession, further validating the platform's commercial traction. In May, the company received a second $19 million milestone payment for the DoveTree pipeline collaboration, with a total potential deal value of up to $5.89 billion, demonstrating platform delivery capabilities that attracted investment from top pharmaceutical expert Gregory Verdine.
In early June, the company's oral hypoglycemic peptide Tensotide™ received US Self-affirmed GRAS certification, overcoming industry challenges like gastric acid degradation, marking another milestone in the consumer health space for glucose control and weight management.
On June 9th, a new drug collaboration targeting GPCR was signed, with a total potential value exceeding $400 million. Notably, beyond milestones and sales royalties, the partner will cover all R&D costs incurred by XTALPI, allowing it to participate in potentially high-return projects with zero financial risk.
XTALPI is steadily demonstrating its systemic capabilities and platform maturity, with early signs of brand premium and pricing power emerging.
A more noteworthy signal comes from insiders. Following their first lock-up extension in May 2025, the company's three co-founders and executive directors, Dr. Wen Shuhao, Dr. Ma Jian, and Dr. Lai Lipeng, voluntarily extended their share sale restrictions a second time in May 2026 until June 12, 2027, involving a total of 555 million shares, representing approximately 12.9% of the total share capital.
Amid a backdrop of frequent share sales in the Hong Kong biotech sector, the founders' commitment, expressed through the tangible time cost of their capital, conveys strong confidence in XTALPI's future, a signal often more significant than analyst reports.
In recent years, global competition in AI drug discovery has intensified, with the convergence of innovative drugs and AI becoming a strategic imperative and a clear industrial trend.
XTALPI has built a triple moat of "time + capital + data" through its investment in physical wet labs, over a decade of data accumulation, and a proven standardized delivery system, making it difficult for latecomers to replicate such depth quickly, even with significant capital.
From an AI drug discovery tool provider, to a Physical AI infrastructure operator, and potentially a scientific data asset platform—XTALPI's business narrative is undergoing a three-stage leap.
While the market may still debate with outdated frameworks like "Is there an AI drug discovery bubble?", the real question to answer is: in the historical shift of scientific experimentation from craft to industrialization, who holds the truly scarce, critical resources?
Benefiting from the dual tailwinds of national efforts to industrialize data as a factor of production and to empower the real economy with AI, the company, with its mature trinity of technology, data, and commercialization, is poised for a systemic value re-rating.
Comments