- The pact fast-tracks innovative, energy-efficient solutions for next‑gen specialty devices
- Specialty technologies are now essential in AI data centers for power management, bandwidth
Driving the news: Lam Research and CEA-Leti, a research institute pioneering micro- and nano-technologies based in France, announced a new multi-year agreement to advance the development of next-generation specialty technology devices.
The partnership accelerates materials exploration, process pathfinding, and manufacturable solutions for power electronics and silicon photonics, both of which are critical to AI infrastructure.
Why it matters: As AI workloads scale, specialty technologies are no longer optional. They are essential for delivering the energy efficiency, bandwidth, and performance required in next-generation AI data centers.
As the semiconductor industry seeks to ramp production for AI, Lam is strengthening strategic partnerships with institutions like CEA-Leti to unlock breakthroughs faster than ever.
What they’re saying: "By rapidly characterizing how new materials behave in complex device structures, we can pinpoint critical integration challenges and provide Lam with actionable feedback," says CEA-Leti CEO Sébastian Dauvé. "This accelerates the transition from promising process innovation to validated functionality in new generations of specialty technology devices.”
"Specialty technologies today are found in nearly every electronic device, and their performance and power efficiency will continue to be challenged as end applications grow,“ says David Haynes, vice president of Specialty Technologies at Lam. “We look forward to working with CEA-Leti at the leading edge of materials science, process technology, and industry-specific applications."
Specialty Technologies’ Increasingly Important Role in AI
According to Haynes, specialty technologies are a critical enabler in AI data centers, especially where bottlenecks in power or connectivity directly limit performance.
- Power devices: Power is now a first order constraint in AI data centers. High density AI server racks draw enormous currents, making highly efficient, tightly regulated power conversion nonnegotiable. Advanced power devices built on new materials such as SiC and GaN are becoming essential—regulating voltage, reducing losses, and maintaining stable operation under extreme load.
- Connectivity: As systems expand, electrical interconnects become constrained by speed, loss, and power consumption. To keep GPUs and accelerators synchronized across racks, the industry is shifting to optical links. Silicon photonics anchors the technology roadmap for this transition, enabling the bandwidth, latency, and energy efficiency required for large‑scale AI clusters.
Lam and CEA-Leti Agreement Is Latest Ongoing Partnership
The new agreement is the latest in a long history of successful co-development projects between Lam and CEA-Leti, including recent work on pulsed plasma technologies.
Lam’s new joint research with CEA-Leti will focus on exploring a range of new, novel materials and films for lower-power, high-performance specialty technology applications and devices, including:
- MEMS
- 3D imaging and sensors
- Power management devices
- RF solutions
- Photonics and MicroLED displays
- Optical interconnects
Lam's innovative etch and deposition technologies, including the company's breakthrough pulsed laser deposition (PLD) system, Lam Prestis™, will be integral to the research.
Vahid Vahedi, Lam Chief Technology and Sustainability Officer, says “this agreement enables Lam to complement its industry-leading capabilities in etch and deposition with CEA-Leti's deep expertise in device characterization to fast-track the development of new, ground-breaking advancements and future generations of more energy-efficient and higher-performance specialty technology devices."