MobbleOpen in Mobble ⇢
Technology · Semiconductors · published 2026-09-25 · via Semiconductor Digest

TRUMPF Laser Process Brings Integrated Cooling to AI Chips

TRUMPF is presenting an ultrashort-pulse laser process at Semicon that allows cooling structures to be produced directly in AI chip packages on an industrial scale. The company says heat removal is becoming a limiting factor as advanced packaging stacks more chips together. The laser method can work with hard materials such as silicon carbide and diamond, which are difficult to structure with conventional etching.

Expanded Detail

TRUMPF says its ultrashort-pulse laser method can create cooling microstructures directly within AI chip packages at industrial scale. The approach targets advanced packaging, where stacked or tightly linked chips concentrate heat that conventional rack- or data-center-level cooling struggles to remove.

The lasers can structure hard materials including silicon carbide and diamond, which are difficult to etch. TRUMPF claims at least five times faster processing than etching while maintaining surface quality and precise geometry. It is presenting the application at Semicon.

Context

If integrated cooling becomes practical, AI chip designers, data center operators, and ultimately users of AI services may benefit from more powerful, efficient hardware. Better heat removal could lower cooling energy needs, though manufacturing costs and access to advanced lasers may shape which firms benefit. The technology may also extend chip lifetimes and enable denser systems, but its societal effects will depend on deployment, affordability, and broader energy choices.

Expanded detail and Context are AI-generated analysis; the linked article remains the authoritative source.
Read the full article at Semiconductor Digest →
This summary is Al-enhanced to contain extended analysis and broader social context. The original is {NAME); the linked article is the authoritative source. Original headline: “TRUMPF Enables Integrated Chip Cooling for the Next Generation of AI Chips.” Browse more stories.