Michelle ThompsonA £225m supercomputer, Isambard-AI, is utilizing artificial intelligence (AI) to advance drug and vaccine development.
Set to become the UK’s most powerful supercomputer when fully operational this summer, Isambard-AI was highlighted in Prime Minister Sir Keir Starmer’s recent initiative to “unleash AI” and drive economic growth.
Professor Simon McIntosh-Smith from Bristol University stated that the supercomputer positions the UK as a global competitor in high-performance computing. Speaking on BBC Radio Bristol, he mentioned that the system is already being used to explore treatments for conditions like Alzheimer’s, heart disease, emphysema, various cancers, and dementia.
Researchers are also leveraging the technology to enhance melanoma detection across diverse skin tones. “It handles tasks that exceed human capabilities due to its immense processing power,” added Prof. McIntosh-Smith.
Isambard-AI is already being utilized to address heart disease and emphysema.
How can artificial intelligence contribute to the development of new vaccines and drugs?
According to Prof. McKintosh-Smith, AI models can now simulate how drugs interact with the body at a molecular level, down to individual atoms and molecules.
“Many drugs function by targeting specific proteins in the body, either deactivating them or altering their behavior,” he explained.
Traditionally, scientists relied on their expertise and educated guesses to predict how treatments would interact with these proteins—a process that was time-consuming, costly, and limited in scope.
“You can only make so many educated guesses,” Prof. McKintosh-Smith said. “Physical experimentation takes time, money, and other resources, making it inherently restrictive.”
This is where advanced tools like the Isambard AI supercomputer transform the process. Using AI, millions of potential drugs can be virtually tested in their early stages, bypassing the need for initial lab experiments.
“AI doesn’t try every possible combination. Instead, it explores random possibilities, identifies the most promising ones, and refines the focus on those,” he said. “It allows us to quickly pinpoint hotspots and accelerate the process in innovative ways.”
Prof. McKintosh-Smith concluded by saying, “The potential to save millions of lives through these advancements is incredibly exciting. It’s amazing to be contributing to this right here in Bristol.”
“Possibly Transformative on a Global Scale.”
Earlier this week, the prime minister highlighted AI’s “vast potential” to enhance public services in the UK. He cited examples like using AI to identify road issues, such as potholes, or to accelerate the diagnosis of diseases like cancer.
When asked about the significant investment in Isambard-AI, Prof. McIntosh-Smith described the potential return as “possibly transformative on a global scale.”
He compared AI’s development to revolutionary innovations like the internet or mobile phones. “One key benefit of government-funded projects like Isambard-AI,” he explained, “is that the work remains in the public interest. Conducted by academics, it tends to be far more transparent compared to research carried out by private companies.”
Once fully operational, Isambard-AI will rank among the ten most powerful computers globally.
Once fully operational, Isambard-AI, located at the National Composites Centre in Emersons Green, will be ranked among the world’s “top 10 fastest” supercomputers.
“We’ve accomplished things no one else has,” stated Prof. McIntosh-Smith.
Supercomputers require immense amounts of energy, and the rapid rise of AI has raised concerns about energy consumption.
Though designed for maximum efficiency, Isambard-AI is no exception. However, Prof. McIntosh-Smith shared with the BBC that the excess energy it generates could provide an unexpected advantage, as it is released as hot water.
“We’re exploring ways to use this waste heat to warm nearby homes and businesses,” he said. “So, if you live near Isambard in the future, we might be able to heat your home using our surplus energy.”
Emily Martinez
