Machine learning powered protein engineering lab could accelerate biotech innovation

Artificial intelligence is rapidly reshaping scientific research, from drug discovery to materials science. Now, a major new investment aims to push that transformation even further through the creation of the country’s first publicly accessible, AI-powered cloud laboratory dedicated to protein engineering.

Northwestern University has been awarded $20 million over four years to establish the AI-Driven, Rapid, Experimental Automation Machine (DREAM) Cloud Lab, a facility designed to accelerate the design, construction, and testing of engineered proteins using AI, robotics, and laboratory automation.  While the initiative is based in Illinois, its significance extends well beyond the United States. For Canada’s growing biotechnology sector, the project offers a glimpse into how the future of life sciences may operate: cloud-connected, highly automated, AI-driven, and increasingly collaborative.

Why proteins matter

Proteins are the molecular workhorses of life. They carry out countless biological functions, from catalysing chemical reactions and transporting nutrients to defending against disease.

Scientists have long sought to engineer proteins with enhanced or entirely new capabilities. Such proteins can be used to develop therapeutics, create sustainable materials, recycle plastics, improve agricultural productivity, detect environmental pollutants, or support critical mineral extraction.

The challenge is that protein engineering remains difficult, time-consuming, and expensive.

Although breakthroughs such as AlphaFold, developed by DeepMind, have dramatically improved the ability to predict protein structures, researchers still must physically build and test proteins in the laboratory. This experimental stage remains a major bottleneck in biotechnology innovation. The DREAM Cloud Lab is designed specifically to address that problem.

The concept behind DREAM is ambitious. Researchers will access the platform through a cloud-based interface and specify the biological function they want a protein to perform. AI models then generate candidate protein designs. Following biosafety and biosecurity reviews, robotic systems automatically build and test those proteins. The resulting experimental data is fed back into the AI systems, creating a continuous design-build-test-learncycle.

This approach represents one of the clearest examples, yet of what some scientists call the “self-driving laboratory” movement. Instead of scientists manually conducting every experimental step, AI generates hypotheses while robotics performs much of the repetitive laboratory work. Researchers remain central to oversight, interpretation, and decision-making, but automation dramatically increases experimental throughput.

As NSF Assistant Director Erwin Gianchandani explained, programmable cloud laboratories are intended to create a “virtuous cycle” of automated hypothesis generation, autonomous experimentation, and rapid data analysis to accelerate discovery.

A national network of cloud laboratories

DREAM is not a standalone initiative. The laboratory is one of 20 inaugural NSF-funded Programmable Cloud Laboratory (PCL) testbeds, which together form a broader national effort to create AI-enabled, remotely accessible research infrastructure. The NSF plans to invest approximately $380 million across the network.

The goal is to democratize access to sophisticated scientific equipment. Traditionally, researchers must secure substantial funding to purchase advanced robotics, analytical instruments, and computing platforms. Cloud laboratories could allow scientists to remotely access these capabilities without building expensive facilities themselves.  This model has particular relevance for start-ups, emerging companies, and smaller academic institutions.

Canada has developed a vibrant biotechnology ecosystem spanning Toronto, Montreal, Vancouver, Edmonton, Saskatoon, and several emerging innovation hubs. Canadian researchers already play significant roles in synthetic biology, protein engineering, personalized medicine, and AI-driven healthcare. Organizations such as the Vector Institute in Toronto, Mila in Montreal, and the University of Toronto’s Acceleration Consortium have pioneered new approaches linking AI and scientific discovery.

The Acceleration Consortium, for example, has attracted global attention for its autonomous materials discovery platforms, which similarly combine AI, robotics, and machine learning to accelerate experimental science. Projects like DREAM suggest that biology may be the next major domain to embrace this model at scale.

For Canadian biotech companies, access to large open datasets, AI models, and collaborative cloud-laboratory frameworks could shorten development timelines and reduce costs associated with protein discovery programs. Over the four-year funding period, researchers expect the facility to synthesize and characterize more than 300,000 proteins, generating approximately 30 million experimental data points. The resulting dataset is expected to become one of the largest publicly available resources for AI-guided protein engineering.

Large, high-quality datasets remain one of the biggest constraints in applying machine learning to biology. By generating millions of standardized protein measurements, DREAM could help train more accurate and capable AI systems.

Responsible innovation remains essential

As impressive as the technology appears, the project’s architects emphasize responsible governance. According to Northwestern, all proposed protein designs will undergo biosafety and biosecurity review before experimental testing. The project will also incorporate oversight from an Ethics and Responsible Innovation Board to help ensure safe and appropriate use of the platform.  Such safeguards are increasingly important as AI systems gain greater influence over scientific discovery.

Canada has similarly emphasized responsible AI development through federal policy discussions and research initiatives focused on transparency, safety, and ethical innovation.