TE Connectivity at OCP Summit 2026 The Interview
GPUs get the headlines, but connectors, cables and power racks are where AI data center limits show up, and TE Connectivity is cashing in.
1 min read
GPUs get the headlines, but connectors, cables and power racks are where AI data center limits show up, and TE Connectivity is cashing in.
Key points
- At OCP Summit 2026, TE Connectivity showed three linked system-level racks: optical interconnect, a sidecar power rack and copper IT.
- TE’s Digital Data Networks unit says it has grown massively over the past three years on the AI infrastructure buildout.
- The new VP of product management, Brian Payne, joined after 26 years at Dell, a sign that component makers are hiring for system-level thinking.
- TE argues that as AI clusters grow, high-speed connectivity and power delivery have to be designed together at the rack level.
- TE has been part of OCP since 2017, which ties it into open hardware standards for hyperscale data centers.
Why it matters
In AI data centers, the cable is turning into the chip. Whoever moves electrons and photons between racks the fastest ends up with leverage that GPU vendors can’t ignore.
Sources
Sources
hardware industry
Frequently Asked Questions
What did TE Connectivity show at OCP Summit 2026?
TE Connectivity built its booth around three interconnected system-level racks: an optical interconnect rack, a sidecar power rack and a copper IT rack. The demo was meant to show high-speed connectivity and power delivery for scaling AI infrastructure.
Why do connectors and power matter for AI infrastructure?
As AI clusters scale, moving data between accelerators and delivering enough power to dense racks become system-level constraints. TE argues that interconnect and power have to be engineered together at the rack level, not as separate parts.
How has AI affected TE Connectivity's business?
Brian Payne, VP of product management for TE's Digital Data Networks business, said the division has grown massively over the last three years by building out AI infrastructure. TE has participated in the Open Compute Project since 2017.