AI has become a powerful tool for designing proteins, predicting biological interactions and identifying new molecules. But in industrial biotechnology, discovering a promising biological pathway is only the beginning. The harder question is whether that pathway can be engineered into a robust production process that works economically at industrial scale. That is the gap Again is targeting with its acquisition of Genomatica (Geno).
The Copenhagen-based company combines AI-driven bioprocess design and scale-up capabilities with Geno’s two decades of experience in computational pathway and strain design, predictive machine learning and industrial bioprocess engineering. Financial terms were not disclosed. Again started with a simple bet: biology, not petrochemistry, should be the primary manufacturing technology of the future. What began with unlocking novel gaseous feedstock, such as CO2, grew into much more. With the acceleration of artificial intelligence, the company aims to expand their upscaling platform with world-class computational biology capabilities.
The stated goal is an end-to-end computational and engineering platform that can move from de novo pathway design to industrial production.
From digital biology to physical production
The deal reflects an important evolution in the way AI is being applied to biotechnology. Much of the current excitement around AI in biotech centers on the digital discovery layer: finding new drug candidates, designing proteins or predicting molecular structures. Industrial biotech adds another dimension. Here, the biological design has to survive contact with a bioreactor, a feedstock, a production organism and ultimately the economics of a manufacturing plant.
Again argues that this is where Geno’s capabilities complement its own platform. Geno has accumulated decades of experimental and process data covering pathway design, strain engineering and scale-up. Again plans to integrate that data and intellectual property into its computational systems.
The underlying premise is straightforward: better data should make biological models more predictive, while real-world production data should in turn improve the next generation of designs. That creates a feedback loop between computation and manufacturing rather than treating them as separate stages.
Closing the scale-up gap
For industrial biotech, this matters because scale-up remains one of the field’s persistent bottlenecks. A pathway that performs well in a laboratory does not automatically translate into an economically viable industrial process.
Again’s model is built around that transition. The company uses biological systems to manufacture chemicals and other industrial products from alternative feedstocks, including captured carbon. Its first commercial-scale U.S. operation, TXS-1 in Texas City, is designed to produce acetic acid from industrial CO2 emissions. Again says the facility is intended to demonstrate that biological production can be integrated into existing industrial infrastructure.
The Geno acquisition adds the upstream computational layer: rather than simply scaling processes that have already been developed, the combined company wants to design new biological production routes computationally and then take them through engineering and scale-up.
The bigger picture: biology becomes an engineering platform
The strategic significance goes beyond the two companies. Industrial biotechnology is increasingly being positioned as an alternative manufacturing infrastructure for chemicals, materials, fuels, food ingredients and other products traditionally made through petrochemical or resource-intensive processes. Again is already working toward this model, using fermentation and biological conversion to turn alternative carbon sources into industrial chemicals.
AI potentially changes the economics of that model. If computational tools can identify productive pathways faster, predict which biological systems are likely to work and learn from manufacturing data, the development cycle could become less dependent on trial-and-error experimentation.
But there is an important distinction from AI drug discovery. The output is not a molecule on a computer screen. It is a process that has to run thousands of times in a physical plant.
That makes the integration of computational biology, process engineering and manufacturing particularly important. The winners in industrial biotech may therefore not be the companies with the best AI models alone, but those able to connect digital design with physical execution.
A geopolitical dimension
The timing also reflects a broader shift in biotechnology policy. The U.S. has increasingly treated biomanufacturing capacity as an issue of economic and national security, alongside sustainability. The National Security Commission on Emerging Biotechnology has argued that the U.S. needs major additional investment in domestic biomanufacturing capacity and has highlighted limited scale-up capability as a structural weakness.
The BIOSECURE Act, enacted in December 2025, adds another layer by restricting certain federal procurement, grants and contracts involving designated “biotechnology companies of concern.” The legislation reflects growing U.S. concern about dependence on foreign biotechnology supply chains.
Against that backdrop, companies that can design and manufacture biological products closer to the point of demand are gaining strategic relevance. Again has explicitly framed its approach around resilient and more geographically distributed supply chains, including its U.S. manufacturing footprint.
What changes for Again?
The acquisition gives Again three potential ways of capturing value: licensing technologies, co-developing critical production pathways with industrial partners and manufacturing products itself.
That could turn the company from a biomanufacturing developer into something closer to a technology platform for biological production – with computational design, process development and physical assets under one roof.
Max Kufner, CEO of Again, put the strategy succinctly: “AI is revolutionizing biology, but algorithms are only as good as their ability to translate into real-world, commercial-scale execution.”
That may ultimately be the most important point about the Geno deal. The next phase of AI-enabled biotechnology will not be judged only by what algorithms can discover. It will be judged by what biology can actually produce – at scale, at cost and in the real world.
Sign up for our newsletter!
