A team at the University of Southern California has developed a memory device that continues working at 700 degrees Celsius (1,300 degrees Fahrenheit), shattering a thermal barrier that has limited electronics for decades. The breakthrough could reshape how AI systems process data while using far less energy, and it opens possibilities for computing in extreme environments like Venus, geothermal plants, and spacecraft where traditional chips fail .

What Makes This Memory Device Different?

The device, called a memristor, is a nanoscale component that both stores data and performs calculations simultaneously. Researchers led by Joshua Yang at the Ming Hsieh Department of Electrical and Computer Engineering at USC built it using an unusual combination of materials: tungsten for the top electrode, hafnium oxide ceramic in the middle, and graphene for the bottom layer .

Tungsten has the highest melting point of any element, while graphene, a single-atom-thick sheet of carbon, is renowned for its exceptional strength and heat resistance. The device retained data for more than 50 hours at 700 degrees without needing to be refreshed, endured over one billion switching cycles at that temperature, and operated at just 1.5 volts with speeds measured in tens of nanoseconds .

"You may call it a revolution. It is the best high-temperature memory ever demonstrated," said Joshua Yang, Arthur B. Freeman Chair Professor at the Ming Hsieh Department of Electrical and Computer Engineering at the USC Viterbi School of Engineering.

Joshua Yang, Arthur B. Freeman Chair Professor, USC Viterbi School of Engineering

How Did Researchers Discover This Breakthrough?

The discovery was partly accidental. Yang's team was initially attempting to create a different graphene-based device that did not work as intended. During that failed experiment, they encountered something surprising that led to this breakthrough. The team then investigated why the device performed so well under extreme heat and uncovered a powerful mechanism at the atomic level .

In conventional electronics, heat causes metal atoms in the top electrode to slowly migrate through the ceramic layer. Eventually, they reach the bottom electrode, creating a permanent connection that short-circuits the device. Graphene prevents this failure because its interaction with tungsten is fundamentally incompatible. Tungsten atoms that approach the graphene surface cannot attach to it, so they drift away instead of forming a conductive bridge .

"To be honest, it was by accident, as most discoveries are. If you can predict it, it's usually not surprising, and probably not significant enough," explained Joshua Yang.

Joshua Yang, Arthur B. Freeman Chair Professor, USC Viterbi School of Engineering

Why This Matters for Artificial Intelligence?

Beyond surviving extreme heat, this device offers a major advantage for AI systems. Many AI applications rely heavily on matrix multiplication, a mathematical operation used in image recognition, language processing, and other tasks. Traditional computers perform these calculations step by step, consuming large amounts of energy. Memristors approach the problem differently by using Ohm's Law, where voltage times conductance equals current, to perform calculations directly as electricity flows through the device .

The result is obtained instantly as the measured current, making the process orders of magnitude faster and more energy-efficient than conventional approaches. Yang noted that over 92 percent of the computing in AI systems like ChatGPT is nothing but matrix multiplication, making memristor-based devices particularly valuable for AI workloads .

"Over 92 percent of the computing in AI systems like ChatGPT is nothing but matrix multiplication. This type of device can perform that in the most efficient way, orders of magnitude faster and at lower energy," stated Joshua Yang.

Joshua Yang, Arthur B. Freeman Chair Professor, USC Viterbi School of Engineering

Where Could This Technology Be Used?

The potential applications extend far beyond laboratory settings. Electronics capable of operating above 500 degrees Celsius have long been a goal for space exploration. Venus, for example, has a surface temperature around that level, and every lander sent there has failed in part due to extreme heat. Current silicon-based chips cannot survive such conditions .

• Space Exploration: Venus landers and other spacecraft could process data directly on site without relying on ground-based systems, enabling real-time decision-making in extreme environments.
• Geothermal Energy Systems: Electronics that function deep underground where surrounding rock can glow red-hot would enable better monitoring and control of geothermal resources.
• Nuclear and Fusion Reactors: Equipment exposed to intense heat in nuclear and fusion systems could operate reliably with high-temperature memory devices integrated into their control systems.
• Automotive Electronics: A device rated for 700 degrees would be extremely robust at the roughly 125-degree temperatures often reached inside automotive electronics, improving durability and reliability.

Yang and three co-authors of the study have already co-founded a company called TetraMem to commercialize memristor-based AI chips at room temperature. Their lab is already using working chips from TetraMem for machine learning tasks. The high-temperature version described in this research could extend those capabilities to environments where traditional electronics cannot operate .

Steps to Bringing This Technology to Market

Despite the promising results, Yang emphasizes that practical applications are still some distance away. The path from laboratory prototype to real-world technology involves several key phases:

• Logic Circuit Development: Memory is only one part of a complete computing system, so high-temperature logic circuits will also need to be developed and integrated with the memristor technology.
• Manufacturing at Scale: The current devices were built manually at very small scales in a laboratory setting, so manufacturing at scale will take considerable time and investment.
• Material Supply Chain: Two of the materials used in the device, tungsten and hafnium oxide, are already widely used in semiconductor production, which accelerates adoption potential.
• Graphene Production: Graphene is newer but is actively being developed by major companies such as TSMC and Samsung, and it has already been produced at wafer scale in research environments.

"This is the first step. It's still a long way to go. But logically, you can see: now it makes it possible. The missing component has been made," noted Joshua Yang.

Joshua Yang, Arthur B. Freeman Chair Professor, USC Viterbi School of Engineering

The research was conducted through the CONCRETE Center, short for Center of Neuromorphic Computing under Extreme Environments, a multi-university Center of Excellence led by USC and supported by the Air Force Office of Scientific Research. The study was published on March 26, 2026 in Science, one of the world's most prestigious peer-reviewed journals .

This breakthrough represents a fundamental shift in how engineers think about thermal limitations in electronics. By understanding the atomic-level mechanisms that prevent failure at extreme temperatures, researchers have opened a new frontier for computing in environments previously thought impossible to reach.