2026 technology review
HBM4: turning a bandwidth headline into useful computation
A wider memory interface changes the performance ceiling, but locality, arithmetic intensity, capacity and thermal limits determine what an application actually gains.
Microelectronics.
A research direction at the intersection of materials, electronic components and the systems built around them.
Small structures. System-level questions.
Our long-term interests include how electronic components are characterised, integrated and used in practical systems. The emphasis is on connecting device-level behaviour with requirements at board, module and product level. Explore our technical reviews for sourced developments and the system-level questions they raise.
Research interests
Materials & devices
The relationship between material properties and electronic behaviour.
Integration & packaging
How components connect to boards, modules and the wider system.
Characterisation
Measurement, repeatability and the evidence needed to compare designs.
Open to a focused
research conversation.
We are interested in collaboration around a clear technical question, complementary expertise and a defined contribution.
Contact our teamTechnical reviews
Evidence. Analysis. Perspective.
November 2025-October 2026 review
CXL 4.0: a faster link does not make all memory local
CXL extends the memory-system design space. Link speed, topology, locality and tail latency must be considered together before treating expanded memory as a substitute for local DRAM.
2025-2026 technology review
Advanced packaging: measurement is part of the architecture
As more functions move into a package, material interfaces and measurement access become system-level decisions. Reliability must be designed in rather than inferred from a working demonstration.