(Alexander Limbach/Shutterstock)
Artificial intelligence continues to reshape how science and engineering are done from the ground up. The U.S. National Science Foundation (NSF) is putting serious weight behind that shift, funding new efforts that weave AI into discovery, testing, and deployment. These programs work together to speed up the journey from bold ideas to real-world impact and strengthen the nation’s position in the global race for AI-driven innovation.
The first stage focuses on the very building blocks of innovation. Nearly $32 million from the NSF Use-Inspired Acceleration of Protein Design initiative will support teams using AI to create enzymes, proteins, and materials with highly specific properties. NSF says “significant progress” in predicting protein structures has opened the door to designing them with much greater precision.
“NSF is pleased to bring together experts from both industry and academia to confront and overcome barriers to the widespread adoption of AI-enabled protein design,” said Erwin Gianchandani. The funded projects range from developing bio-based chemicals to producing recyclable high-performance plastics, unlocking new possibilities in manufacturing, health, and energy.
(JHVEPhoto/Shutterstock)
The investment comes at a pivotal moment for the field. In recent years, tools like Google DeepMind’s AlphaFold have made it possible to predict protein structures with remarkable accuracy, shifting the focus from mapping these molecules to designing them for specific uses. That change allows AI to help create materials, medicines, and industrial processes that once took years to develop.
NSF’s Ideas Lab process brought together researchers and industry leaders to ensure these designs address real-world needs from the start. Projects include AI-engineered enzymes for producing bio-based acrylates used in paints and Plexiglas, and bacteria that can create recyclable plastics able to withstand high temperatures. These efforts show how AI-driven protein design could ripple far beyond the lab, with applications in manufacturing, healthcare, and energy.
While AI-driven protein design is opening the door to new scientific breakthroughs, the next challenge is making sure those advances can hold up outside the lab. That is the aim of NSF’s AI-Ready Test Beds program, which is putting more than $2 million into planning grants for environments where researchers can push AI systems to their limits. These sites are meant to replicate the messy, unpredictable conditions of the real world so ideas can be proven, refined, and made ready for use.
“The AI-ready test beds program is uniquely NSF,” said Ellen Zegura, acting assistant director for CISE. “It leverages critical, existing test environments to drive AI progress.” By upgrading established research facilities to handle advanced AI evaluation, the program ensures that performance testing also measures how well systems adapt and recover when things don’t go as planned.
The projects cover a wide range of settings. There is a city-scale wireless network in New York, an agricultural systems lab at Cornell University, a disaster resilience hub at the University of Maryland, and autonomous vehicle testing grounds at the University of Michigan. Each location offers its own set of real-world challenges, giving AI systems the chance to prove they can work reliably when the variables are constantly changing.
(metamorworks/Shutterstock)
The next step is making sure researchers have the tools to turn promising ideas into fully tested solutions, no matter where they are. The AI breakthroughs reshaping protein design and proving themselves in real-world trials still need a place where experiments can be run quickly, repeatedly, and without the limits of a single physical lab. NSF’s answer is a new network of programmable cloud laboratories, built to make that kind of access possible.
These facilities will be equipped with AI-enabled systems that can manage every stage of an experiment, from preparing the setup to interpreting the results. Researchers will be able to log in remotely, design workflows, and watch them unfold in real time, with the system adjusting conditions as new data comes in.
Although the first focus areas will be biotechnology and materials science, the reach could be much broader, opening opportunities for universities, startups, and even classrooms to participate in high-level research. By expanding access to cutting-edge experimentation, NSF hopes to accelerate AI-enabled science across a much wider community.
The potential isn’t limited to materials or manufacturing. AI-guided drug discovery is already making headlines, with companies like Insilico Medicine and Recursion moving computer-designed drugs into clinical trials. Programmable cloud laboratories could accelerate that trend, giving researchers the ability to go from model to molecule faster and with far fewer bottlenecks.
(Shutterstock AI Image)
“This is about creating the spaces, tools, and collaborations where AI can make the biggest difference,” said Ellen Zegura, acting assistant director for CISE. “If we can connect research to real-world challenges and give scientists the means to act on their ideas, the results can benefit people and industries across the country.”
Behind each of these programs is a vision of giving researchers the freedom to think bigger, move faster, and tackle problems that once seemed out of reach. By linking advanced AI capabilities with the tools and environments to use them, NSF is helping ensure that ideas born in the lab can make a difference far beyond it.