Analog Devices' AI Data-Center Business Is Growing More Than 100%

Analog Devices is becoming one of the clearer examples of how artificial-intelligence spending is spreading beyond GPUs. The company's fiscal third-quarter results showed record revenue of $4.02 billion, up 40% from a year earlier, with Data Center and Industrial leading growth.

The more revealing number sits inside Communications. That segment represented 16% of quarterly revenue and grew 84% year over year. Analog Devices said data center now accounts for roughly 80% of Communications revenue, while both optical and power products serving data centers grew more than 100% from a year earlier. That makes ADI's AI exposure increasingly measurable rather than merely thematic.

AI clusters need power conversion and optical control as much as compute

The economics of AI infrastructure are moving outward from the accelerator. As rack density rises, power has to be converted and delivered more efficiently, while higher-bandwidth networks require increasingly sophisticated optical signal chains. Those constraints create semiconductor content around the GPU rather than inside it.

ADI is exposed to both sides. Its portfolio spans power management, signal processing and optical-control components, giving it a way to monetise higher AI-system complexity without competing directly for the accelerator socket. That distinction matters because it makes the company a picks-and-shovels beneficiary of several possible compute architectures rather than a bet on one accelerator winner.

The financial evidence is now strong enough to matter

Many companies have attached AI language to investor presentations. ADI's case is becoming harder to dismiss because the growth is visible in reported segment numbers. Communications grew 18% sequentially as well as 84% year over year, and management explicitly identified data-center optical and power as triple-digit growers.

The wider quarter was also healthy. Industrial represented 49% of revenue and grew 53% year over year, while Automotive grew 16%. That diversification is important: ADI does not require data centers to carry the entire company, but AI infrastructure can raise the growth rate of a business that already has large industrial and automotive franchises.

Acquisitions show where management thinks the bottlenecks are moving

ADI has been adding capabilities around the physical infrastructure of AI. Its Empower Semiconductor acquisition expanded high-density power technology, while the announced $1.35 billion Alif Semiconductor acquisition adds AI-native processors designed to bring local intelligence into physical systems.

The combination suggests a broader strategy. ADI is positioning around systems that sense, process, connect and manage power, from data centers to the edge. Investors should still separate acquisition narratives from realised returns, but the direction is consistent with the organic data-center growth already appearing in the accounts.

What investors should watch next

The first test is whether triple-digit data-center growth survives tougher comparisons. The second is mix: power and optical demand need to translate into attractive incremental margins rather than simply higher revenue. The third is customer concentration, because hyperscale infrastructure spending can produce large orders but also greater dependence on a small number of buyers.

Fiscal fourth-quarter guidance calls for roughly $4.3 billion of revenue, plus or minus $100 million. Continued strength in Communications would reinforce the argument that AI infrastructure is becoming a durable growth leg rather than a one-quarter surge.

GMR view: ADI looks underappreciated in the AI infrastructure map

Our view is that investors still tend to draw the AI semiconductor map too narrowly. Accelerators deserve attention, but the physical constraints around those accelerators are becoming increasingly valuable. ADI has credible exposure to two of the hardest constraints, power density and optical connectivity, and its latest results show that customers are already spending against them.

That does not make ADI a substitute for Nvidia or Broadcom. It makes it a different kind of AI investment case. We think that distinction is attractive: the company can benefit from escalating system complexity without needing to win the headline compute architecture. If data-center growth remains near current levels while the rest of the portfolio stays healthy, ADI deserves to be treated as a core AI-infrastructure supplier rather than an analog chip company receiving a temporary AI halo.