Coherent's AI Datacenter Revenue Surges 59% as Q4 Sales Hit $2.05 Billion

Coherent reported record fourth-quarter fiscal 2026 revenue of $2.05 billion, with demand from AI datacenters driving a sharp increase in its Datacenter and Communications business.

Revenue rose 42% year over year on a pro forma basis, marking the company's first quarter above $2 billion in sales. Non-GAAP gross margin increased 215 basis points to 40.2%, while non-GAAP earnings per share rose 74% to $1.74, according to the company's fiscal 2026 financial results.

The results continued a growth trend that had accelerated through fiscal 2026. Coherent's Datacenter and Communications segment generated $1.62 billion in fourth-quarter revenue, up 59% from a year earlier, while Industrial revenue was $431 million, up 1%.

That shift means Datacenter and Communications now accounts for 79% of Coherent's revenue. The company makes optical transceivers, lasers and other photonic components used to move data across datacenter and communications networks.

AI Demand Is Driving Optical Capacity Expansion

Coherent is increasing manufacturing capacity as AI systems require faster connections between processors, memory and networking equipment. The company said it is on track to double internal Indium Phosphide (InP) output by the end of 2026 and more than double it again in 2027.

The expansion builds on a broader push to increase US production of optical networking components. In June, Coherent signed a letter of intent for up to $50 million in direct funding under the CHIPS and Science Act to expand its 6-inch InP manufacturing facility in Sherman, Texas. The company said the project will double manufacturing space and quadruple wafer production capacity.

The capacity push comes as optical components become a larger part of the AI infrastructure stack. High-speed networking is increasingly important as AI workloads require processors to exchange larger volumes of data with lower latency.

Coherent is also developing technologies aimed at higher-bandwidth AI networks, including 400G-per-lane optical links, silicon photonics, co-packaged optics and multi-rail optical transport. The company demonstrated several of these technologies at its 2026 technology briefing, including a 400G-per-lane optical link and a multi-technology co-packaged optics architecture.

The trend is broader than Coherent. Optical components have already emerged as a supply constraint in AI datacenter networking, adding pressure on suppliers to expand production alongside semiconductor and networking capacity.

Coherent has also deepened its relationship with NVIDIA as AI infrastructure scales. In March, the companies announced a strategic partnership focused on optics technology for next-generation datacenter architectures. Coherent's Sherman expansion was subsequently announced alongside NVIDIA, which has worked with the company for more than two decades.

Coherent Sees Growth Continuing Into Fiscal 2027

Coherent expects the current demand cycle to continue into its next fiscal year. For the first quarter of fiscal 2027, the company guided for revenue of $2.2 billion to $2.4 billion and non-GAAP EPS of $1.85 to $2.05.

Management also expects to reach its first quarter above $3 billion in revenue by the end of fiscal 2027. Customer orders, according to the company's earnings presentation, now extend into calendar 2028.

The company is investing heavily to support that outlook. Capital expenditures reached $556 million in the fourth quarter, compared with $290 million in the third quarter, as Coherent increased spending on manufacturing capacity.

The investment reflects a larger change in the AI infrastructure market. AI infrastructure is increasingly being built across multiple layers of the stack, from processors and memory to networking and optical connectivity.

For Coherent, that shift is already visible in its financial results. The company's latest quarter shows how demand for AI datacenters is moving beyond compute hardware and into the optical systems required to connect increasingly dense AI infrastructure.