TetraMem Advances Analog In-Memory Computing with Tapeout of New 12nm AI SoC

Leveraging Synopsys cloud-enabled EDA solutions to scale multi-level RRAM-based AI chip development to advanced FinFET nodes

TetraMem, a Silicon Valley-based pioneer in analog in-memory computing (A-IMC), announced the successful tapeout of MLX300, the world’s first 12nm A-IMC system-on-chip (SoC) based on multi-level RRAM technology. The SoC was developed using industry-leading EDA solutions enabled through Synopsys Cloud FlexEDA, and this milestone represents a major step in bringing this architecture toward advanced-node, production-ready computing.

“TetraMem’s architecture represents a fundamentally new approach to AI computing by converging memory and compute together,” said Glenn Ge, CEO and Co-founder of TetraMem. “Leveraging Synopsys EDA solutions through Synopsys Cloud have been instrumental in enabling efficient global collaboration and accelerating our path from research innovation to manufacturable silicon.”

The MLX300 tapeout follows TetraMem’s recently announced MLX200/MLX200M multi-level RRAM A-IMC SoC implemented on a commercial 22nm process, marking two successful full-mask tapeouts achieved in 2026. Both designs leveraged Synopsys EDA solutions delivered on the Synopsys Cloud FlexEDA, enabling highly efficient distributed development across TetraMem’s engineering teams in the United States and Singapore while significantly accelerating design productivity and iteration cycles. For example, TetraMem reduced their overall design environment bring up time from weeks to minutes, and achieved 3x faster Register Transfer Level (RTL) verification simulation relative to on-premises methods.

“TetraMem’s work in analog in-memory computing reflects the kind of bold innovation needed to unlock the next wave of AI performance and efficiency,” said Sanjay Bali, Senior Vice President of EDA Product Management and Strategy at Synopsys. “At Synopsys, we are pleased to support TetraMem with our cloud-enabled design solutions, advanced EDA technologies, and proven IP to help accelerate the development of next-generation AI hardware.”

TetraMem’s unique architecture enables scalable development of next-generation AI accelerators targeting dramatically improved performance-per-watt and reduced data movement bottlenecks. The MLX300 builds on that architecture as a versatile single-chip accelerator for audio, vision, and small language models, supporting up to 64 million INT4 parameters on-chip without external DRAM for model-weight storage. The SoC is targeting energy-efficiency of more than 50–80 TOPS/W, demonstrating the potential of RRAM-based analog computing to deliver significant efficiency gains for next-generation AI inference.

About TetraMem

TetraMem is a Silicon Valley AI semiconductor company developing analog in-memory computing technologies based on multi-level RRAM (memristor) to dramatically improve AI energy efficiency and scalability.

View source version on businesswire.com: https://www.businesswire.com/news/home/20261027687318/en/