Key takeaways from the AI+ conference

AI+ Materials

In Dr. Zhou Min’s keynote ‘AI for Materials’, she highlighted the challenges of AI in materials R&D, including:

• Complexity of modeling, from microscopic to macroscopic levels
• High-throughput experiments require high standards for equipment and significant investment
• The biggest challenge remains the lack of data, especially experimental data, which holds the highest value

Materials Genome Engineering in China, by Dr Yi Jin
Dr. Yi Jin of 2Prime Ventures introduced the progress of China’s Materials Genome Initiative from recent years. The US and China launched their respective Materials Genome Initiatives in 2011 and 2015, both aiming to shorten the time for materials R&D while reducing costs. Currently, much of the work focuses on the field of materials data, including:

• Using dynamic container technology and multi-source heterogeneous material data sources for retrieval, discovery, analysis, and mining
• Automatic data mining from materials-related literature
• Establishing a cloud infrastructure and big data platform bases for data sharing and circulation

From a company perspective, AI+ Material Companies often face a choice between becoming a platform company or a product company. Due to different industrial environments and ecosystems, different companies have made different choices:

• Citrine Informatics (USA), which began exploring AI with materials in 2013, attempted their own product development. They chose to become an SaaS-based company, empowering other materials companies in their R&D and achieving significant commercial progress.
• Bluepha (蓝晶微生物), a synthetic biology startup, designed and built a synthetic biology R&D platform called ‘Synbio OS’. The platform’s functions include cloud design software, autonomous strain construction, and a high-throughput automated small-scale bioreactor array. Utilizing the capabilities of this R&D platform shortens the R&D cycle and significantly raises R&D standards.

• ShineHigh Innovation (杉海创新) and InnoAero (云锦特导) have firmly taken the productization route, launching highly-recognized products in the fields of supramolecular biology and high-performance copper alloys, respectively.

 
Our POV on AI+Materials can be summarized as:

• AI’s impact on the materials industry is going to accelerate and potentially huge. New materials technologies used to be “technology push” (new materials invented, looking for applications), with AI, it could become “market pull” (market needs identified, AI helps quickly come up with new materials). CM’s first investment in the area, a Chinese company called Innoaero, is a case-in-point.
• We see west’s advantage in the area as loads of data residing in large companies, like BASF. A good AI platform/tool start-up has good potential in helping large companies to utilizing that data, with SAAS model.
• We see China’s advantage in the area as quick/low-cost experimentation and rapid re-iteration to test what AI model comes up with – thus quickly identifying new materials if a clear market opportunity exists.
• In short, we are looking at platform/SAAS investments in the west, and product companies in China.