Proposed Memory Solution Can Run At Up To 600°C, Fueling Potential Adoption In Servers

Next-Gen Server Memory Might Feature Modified Material Layer, Replacing The Traditional Silicon Carbide & Operating At Extreme Temperatures

According to research published at Nature Electronics, ferroelectric aluminum scandium nitride (AlScN) based flash drives might be the next-gen standard for the markets, not just because of how good the name sounds but also because of the material's temperature sustainability. Researchers Deep Jariwala and Roy Olsson from the University of Pennsylvania have showcased a memory prototype that remains functional even at 600° C temperature, which is said to be double what current market alternatives offer.

Now, the bigger question is why such memory solutions are needed in the first place. Well, the answer is simple. Integrated with hefty components, Gigantic compute systems generate tremendous amounts of thermal energy, which dissipates through the onboard cooling solution, liquid or air. In memory devices, however, due to their volatile nature, out-of-range temperatures can lead to data losses and performance degradation, which is unaffordable for large-scale systems. It becomes vital for the onboard memory to function continuously, even under high temperatures.

Jariwala says that their solution targets massive AI systems since they allow them to run even under harsher conditions. Apart from temperature benefits, AlScN-based memory devices reportedly have an edge over traditional products in bridging the inefficiencies present in data transfer among the central processing unit and memory.

While silicon carbide technology is great, it is nowhere close to the processing power of silicon processors, so advanced processing and data-heavy computing such as AI can't really be done in high-temperature or any harsh environments.

The stability of our memory device could allow integration of memory and processing more closely together, enhancing speed, complexity, and efficiency of computing. We call this 'memory-enhanced compute' and are working with other teams to set the stage for AI in new environments

- Deep Jariwala, University of Pennsylvania

The research does show that future memory standards could evolve rapidly, with the fundamental change done within the core material layers utilized, and in this case, AlScN looks like a no-brainer, but it's still early to say this.