Neuromorphic computers — hardware built to mimic brain circuitry — can solve partial differential equations, the math behind fluid dynamics, electromagnetic fields and structural mechanics, according to Sandia National Laboratories computational neuroscientists Brad Theilman and Brad Aimone. Their paper appeared in Nature Machine Intelligence and was funded by the Department of Energy's Office of Science through the Advanced Scientific Computing Research and Basic Energy Sciences programs, plus the National Nuclear Security Administration's Advanced Simulation and Computing program.

That funding mix signals the practical stakes. Supercomputers across the nuclear weapons complex consume immense energy simulating weapons physics. If neuromorphic systems can run those same PDE workloads at brain-like power levels, the portfolio implications for national-security simulation are direct.

What does the result change?

For decades, experts assumed neuromorphic computers suited pattern recognition and AI acceleration — not rigorous numerical mathematics. Conventional wisdom held that PDE solving belonged to traditional supercomputers. The Sandia algorithm overturns that assumption.

"You can solve real physics problems with brain-like computation," Aimone said. "That's something you wouldn't expect because people's intuition goes the opposite way. And in fact, that intuition is often wrong."

Theilman frames the energy argument bluntly: "We're just starting to have computational systems that can exhibit intelligent-like behavior. But they look nothing like the brain, and the amount of resources that they require is ridiculous, frankly."

The researchers weren't shocked by their own result. They argue the brain performs comparable computation constantly, outside awareness.

"Pick any sort of motor control task — like hitting a tennis ball or swinging a bat at a baseball," Aimone said. "These are very sophisticated computations. They are exascale-level problems that our brains are capable of doing very cheaply."

How old is the underlying model?

The algorithm rests on a model computational neuroscientists have used for over a decade. Theilman and Aimone showed it maps onto PDEs in a way nobody had noticed before.

"We based our circuit on a relatively well-known model in the computational neuroscience world," Theilman said. "We've shown the model has a natural but non-obvious link to PDEs, and that link hasn't been made until now — 12 years after the model was introduced."

That 12-year gap matters for R&D managers scanning adjacent fields for transferable methods: fundamental connections between neuroscience models and applied mathematics can sit unused for years until someone looks.

The team's circuit retains strong structural and dynamic similarities to cortical networks in the brain, which gives the work a second life beyond simulation. Aimone suggests computation itself may be the right lens on neurological disease.

"Diseases of the brain could be diseases of computation," Aimone said. "But we don't have a solid grasp on how the brain performs computations yet."

If that hypothesis holds, neuromorphic systems could yield clues for understanding and eventually treating conditions such as Alzheimer's and Parkinson's. That remains a projection, not a measured result.

What comes next?

The researchers describe the current work as foundational. Their immediate question is whether more advanced applied-math techniques have neuromorphic formulations.

"If we've already shown that we can import this relatively basic but fundamental applied math algorithm into neuromorphic — is there a corresponding neuromorphic formulation for even more advanced applied math techniques?" Theilman said.

The team wants applied mathematicians, neuroscientists and engineers to join the effort, and it frames the payoff in dual terms: scientific insight plus a working tool. "We have a foot in the door for understanding the scientific questions, but also we have something that solves a real problem," Theilman said.

The published efficiency claims deserve scrutiny the release does not quantify — no benchmark figures, speedup factors or power measurements appear in the announcement. Sandia positions the work as groundwork for what could become the world's first neuromorphic supercomputer, and the researchers say such machines could eventually play a central role in Sandia's security mission.