Communications equipment new orders rise 40% year over year amid AI infrastructure boom

The AI buildout has a hardware problem, and right now, hardware suppliers are very happy about it. New orders for US-manufactured communications equipment surged 43.1% year over year in July 2026, according to US Census Bureau data, pushing the year-to-date increase to 35% through the same month. The category covers the physical plumbing of the internet: high-speed optical networking gear, routers, and fiber optic components that move data at the scale AI workloads demand.

What is actually driving the orders

Industry research from Dell’Oro Group identifies AI-driven data centers as the primary catalyst for telecom equipment revenue growth, now outpacing investment in traditional 5G infrastructure. Global telecom equipment revenues grew 5% year over year in the first half of 2026, with nearly all of that growth attributable to AI data center connectivity.

Core capital goods orders rose 3.3% in March 2026 alone. That was the largest single-month gain in that category since 2020. Computers and electronic products led the charge, reflecting AI infrastructure as the organizing principle behind corporate capital allocation.

Company-level results confirm the macro picture

Keysight Technologies reported a 35% year-over-year increase in communications solutions revenue in a single quarter. Keysight makes the test and measurement equipment used to validate that networking hardware actually performs at rated speeds, which means its order book is a leading indicator for the sector.

HPE reported record networking orders and backlog in its third quarter of 2026. Record backlog signals surging demand but also means customers are waiting longer for delivery, indicating that demand is currently running ahead of the company’s ability to fulfill it.

What this means for the competitive landscape

The strategic logic for hyperscalers is straightforward. The networking layer determines how fast training runs can exchange gradients across thousands of accelerators, which directly affects how quickly a model can be trained and at what cost. That makes networking equipment a non-discretionary line item for any serious AI infrastructure project.

The risk worth watching is concentration. A significant portion of communications equipment demand is flowing from a relatively small number of hyperscale buyers. The 43.1% year-over-year figure for July reflects a moment when everyone is building at once.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.