Data Center and AI Demand Drives Q1 2026 Growth Across Fabless Chip Suppliers
Редакция ЦОД:Главное
Data center and AI-related demand was the principal revenue driver for the ten leading fabless semiconductor companies in the first quarter of 2026, according to an industry note. The growth extended beyond AI accelerators to server CPUs, networking, optical interconnects and data-center power systems.
- Data-center and AI demand was cited as the main revenue driver for leading fabless suppliers in Q1 2026.
- NVIDIA reported US$15 billion in quarterly networking revenue, with year-on-year growth described as threefold.
- Broadcom demand was supported by switches, optical DSPs and other interconnect products.
- AMD raised its server CPU market estimate to US$120 billion, while Marvell’s Data Center revenue share exceeded 75%.
- MPS is extending its data-center power portfolio into the high-voltage segment.
Data-center and AI demand supported revenue growth across the leading fabless semiconductor suppliers in the first quarter of 2026, spanning compute, networking, optical connectivity and power delivery.
NVIDIA benefited from increased GB300 shipments. Its quarterly networking revenue reached US$15 billion, reported as three times the level of the prior year. The result underlines the importance of networking alongside accelerator deployments in AI infrastructure build-outs.
Broadcom was supported by demand for switches, optical DSPs and other interconnect products. The industry note also said that Google and Anthropic are expanding TPU deployments, although it provided no deployment volumes, schedules or infrastructure specifications.
Elsewhere, AMD raised its estimate for the server CPU market to US$120 billion. Marvell’s Data Center business exceeded 75% of company revenue, supported by its data-center interconnect operations. MPS is expanding its data-center power portfolio, including in the high-voltage segment.
Taken together, these indicators suggest that AI infrastructure spending is being reflected across several semiconductor categories, not only accelerators. This could leave interconnect throughput and high-voltage power-system design as important areas of both supply-chain demand and infrastructure constraint for data-center operators.
The reported results point to a broad infrastructure investment cycle rather than demand confined to accelerators. For data-center operators and supply-chain participants, AI build-outs could continue to support multiple layers of the hardware stack, while placing greater emphasis on interconnect performance and high-voltage power architecture.
- TechNews 科技新報 / КитайскийПервоисточник →
Материал подготовлен редакцией ЦОД: Главное на основе указанных первоисточников.