Pegatron Texas plant to build Nvidia AI servers by March 2026

Pegatron’s Texas build‑out is another data point in the quiet arms race over who can supply the infrastructure for large‑scale AI. By committing a full manufacturing footprint in the U.S. for AI servers built around Nvidia chips, Pegatron is betting that hyperscalers and enterprise buyers will increasingly demand domestic assembly for high‑end compute—both for supply‑chain resilience and for political optics. The company is also talking about triple‑digit growth in its AI‑server revenue and expanding into inference and ASIC‑based systems, not just GPU rigs. ([bez-kabli.pl](https://www.bez-kabli.pl/pegatrons-texas-factory-deadline-apple-supplier-targets-end-march-finish-as-ai-server-push-heats-up/))

For the race to AGI, this matters because model progress is increasingly constrained by hardware and logistics rather than just algorithms. The more diversified and geographically distributed the AI‑server supply chain becomes, the easier it is for U.S.‑based labs and cloud providers to keep scaling compute in spite of export controls on China and tight chip supplies. Pegatron’s move also reflects a broader trend: original design manufacturers are repositioning themselves as strategic partners in AI infrastructure, not just low‑margin assemblers. That could squeeze margins for incumbent server OEMs while giving frontier labs more flexibility in how and where they deploy clusters.