AI drug discovery sees another round of intensified efforts from industry giants.

Multinational pharmaceutical companies continue to increase their investment in the AI drug development track. Recently, Swiss pharmaceutical giant Roche disclosed at an investor event that it will promote the construction of an in-house artificial intelligence laboratory to accelerate new drug discovery.

Aviv Regev, Head of Research and Early Development at Genentech, a wholly owned subsidiary of Roche, said, the company aims to achieve "artificial intelligence independence" in the field of drug R&D, and the AI automated laboratory has already started construction.

It is reported that the laboratory adopts the "Lab in the Loop" cycle mode: the artificial intelligence model puts forward hypotheses, the robot system completes the wet experiments, and the experimental data flows back to feed the model iteration, forming a closed loop of dry and wet experiments, and translating experimental results into intellectual property rights.

In terms of capital, Roche plans to reallocate about 2 billion Swiss francs (approximately 2.41 billion U.S. dollars) of remaining R&D funds to invest in R&D projects and efficiency improvement. The company aims to launch up to 20 new molecular entities (NMEs) to the market by 2030.

The company further stated that its artificial intelligence tool Target Nexus is expected to be involved in 80% of portfolio decisions by the end of 2026. Roche recently opened an innovation center in Boston, where part of the focus is the application of artificial intelligence and machine learning in drug R&D.

Empowered by AI, the success rate of Roche's Phase III clinical trials has jumped from 65% in 2025 to over 80%. Regev said that from the fourth quarter of 2025 to the second quarter of 2026, 40% of pipeline decisions have been supported by traceable artificial intelligence.

From an industry perspective, focusing on the life science sector to build AI automated laboratories has become a competitive direction for global industry giants.

On September 18, AI company Anthropic was revealed to be working on the construction of wet laboratories. It already has AI agent control and communication standards for physical equipment, which can provide unified interfaces for devices such as microscopes, automatic pipetting workstations, and robotic arms, leaving only one step to the realization of the "dry-wet closed loop".

On the pharmaceutical enterprise side, Novo Nordisk announced the application of the Claude model and the Claude Science platform to build dedicated solutions for drug discovery challenges identified by scientists; both Twist Bioscience and GenScript Biotech announced that they have reached cooperation with Lilly TuneLab™, the AI/machine learning drug discovery collaboration platform of the world's leading pharmaceutical company Eli Lilly, to provide wet experiment services to participating enterprises and help them generate experimental data quickly.

Huaxin Securities believes that the layout of multinational pharmaceutical companies in AI new drug R&D has gradually shifted from introducing AI new drugs through cooperation to reshaping the drug development infrastructure. The development of AI accelerates new drug R&D. With the increasing demand for wet experiment verification, relevant orders are expected to increase quarter by quarter.

From the perspective of investment logic, Southwest Securities pointed out that AI drug development is at a critical period of commercial value verification, and it is recommended to grasp investment opportunities from three dimensions: R&D, commercialization and BD:

First, on the R&D side: Pay attention to core technology breakthroughs and clinical data disclosure nodes, especially phase II/III druggability verification, as well as data catalysis from industry academic conferences. Enterprises with dry-wet closed loop capabilities have more prominent long-term barriers;

Second, on the industrial chain side: Focus on upstream suppliers that benefit from rising wet experiment demand, who are the core beneficiaries in the AI drug development verification period;

Third, on the BD side: BD transactions between global pharmaceutical companies and AI platforms remain active, and large-value cooperation and License-out are important stock price catalysts.

This article is from the WeChat Official Account "Sci-Tech Daily", author: Zhang Zhen, published with authorization from 36Kr.