Quinas Completes Innovate UK Project Advancing ULTRARAM for AI and Neuromorphic Computing

Why It Matters

By delivering a memory that is fast, non‑volatile and ultra‑efficient, ULTRARAM could reshape AI hardware, lowering power costs and enabling new neuromorphic architectures across data‑center and edge markets.

Key Takeaways

• ULTRARAM combines DRAM speed with flash non‑volatility
• Project funded by Innovate UK, £1.1 m (~$1.4 m)
• Enables compute‑in‑memory for neuromorphic AI chips
• Collaboration includes Lancaster University, IQE plc, and Cardiff partners
• Positions Quinas for next‑gen secure, low‑energy AI hardware

Pulse Analysis

Memory bandwidth and power consumption have become the primary bottlenecks for modern artificial‑intelligence workloads. Quinas Technology’s ULTRARAM tackles this challenge by marrying the nanosecond access speed of DRAM with the persistent storage of flash, all built on compound‑semiconductor quantum resonant tunnelling. The device’s ultra‑low‑energy operation—reported at fractions of a picojoule per bit—offers a sustainable alternative to conventional volatile memory. Completing the Innovate UK‑funded project, which received roughly $1.4 million, validates the technology’s readiness for larger‑scale integration and signals a shift toward re‑imagined memory architectures.

Neuromorphic computing, inspired by the brain’s synaptic structure, relies on performing calculations directly inside memory to eliminate costly data movement. ULTRARAM’s compute‑in‑memory (CIM) capability provides the necessary endurance and non‑volatility to support spiking neural networks and other brain‑like models. Early system‑level simulations show up to 40 % reduction in energy per inference compared with traditional GPU‑based pipelines, while maintaining comparable latency. This efficiency gain is especially attractive for edge devices and data‑center accelerators that must balance performance with strict power budgets, positioning ULTRARAM as a strategic enabler for next‑generation AI hardware.

The successful project also cements a collaborative ecosystem that includes Lancaster University, IQE plc, and research groups in Cardiff, creating a pipeline from device physics to chiplet‑level integration. With accolades from WIPO, Taiwan’s ICTGC, and Silicon Valley’s Future of Memory Conference, Quinas is attracting investor interest and potential OEM partnerships. As the UK government continues to back semiconductor innovation, the company’s roadmap—targeting cross‑bar arrays and commercial production within the next two years—could accelerate the adoption of secure, energy‑efficient memory across AI, autonomous systems, and high‑performance computing markets.