TeRAM raises $37m to develop 3D SRAM for AI chips
The company is developing custom stacked memory designed to increase bandwidth and capacity for AI computing while reducing power use.
Putting memory closer to the processor
TeRAM is designing custom 3D SRAM that stacks directly onto AI compute chips. The packaging approach is intended to move data more quickly between memory and processors, addressing capacity and bandwidth constraints as AI workloads grow while limiting power consumption.
The technology remains in development: TeRAM is validating its design and targets initial customer production in 2029 for data-center deployments. The company is building customized chips for those first customers rather than selling a commercially deployed memory product today.
Use of funds
Validate technology
Scale the team developing customized 3D SRAM memory chips for initial data-center customers