AI Exposure Leaders

Revenue tripled to $1.3B selling the active copper and optical cables connecting GPUs inside hyperscale AI clusters, with five of six hyperscalers as customers and guidance for over 80% growth in FY2027.

Credo provides the high-speed connectivity building blocks for AI clusters. Its ZeroFlap Active Electrical Cables (AECs) are plug-and-play copper cables that connect GPUs to switches within racks, delivering up to 1,000 times greater reliability than older optical cables at roughly half the power. For longer distances, the company supplies optical DSP chips to transceiver makers and builds its own full optical transceivers with embedded telemetry software that prevents network link flapping—a critical capability as AI clusters scale to hundreds of thousands of GPUs. The recent acquisition of Dust Photonics added in-house silicon photonics integrated circuits, which simplify optical design and reduce laser count.

Operates over 1 GW of GPU cloud infrastructure generating $2.1B in quarterly revenue, backed by $99.4B in contracted backlog including named commitments from Meta, Microsoft, OpenAI, and Anthropic.

CoreWeave operates a specialized AI‑cloud platform that provides on‑demand access to NVIDIA GPU compute, high‑speed networking, and integrated software tools for AI development. Its stack spans GPU instances across multiple generations, AI‑optimized storage, and orchestration software including the SUNK cluster manager, Mission Control observability suite, and the Weights & Biases MLOps platform. This environment supports the full AI lifecycle — from training and fine‑tuning to large‑scale inference and agentic workflows — and is used by top AI labs, financial firms, and physical‑AI startups.

AI revenue grew 841% year-over-year to $390M quarterly from its vertically integrated GPU cloud, with NVIDIA's $2B equity investment and $20-25B in guided 2026 capex.

Nebius Group is a full‑stack AI cloud platform providing GPU‑accelerated compute, bare‑metal clusters, multi‑tenant cloud, storage, and managed inference tools including Token Factory and agentic search. The platform serves the entire AI lifecycle — from training large models to deploying inference — and is designed for both hyperscale clients and AI‑native enterprises. Two early‑stage units, Avride (autonomous vehicles) and TripleTen (edtech), sit alongside the core business; management intends to deconsolidate them over time.

Data Center revenue reached $75B in the quarter at 91.5% of total with networking revenue tripling. Management describes demand that has 'gone parabolic.'

NVIDIA designs the accelerators, networking, and software that form the backbone of modern AI infrastructure. Its GPUs and systems power the training and inference of frontier models, and its CUDA software ecosystem is the default platform across every major cloud and AI lab. The company has recently expanded into standalone CPUs (Vera) and integrated CPU-GPU platforms (Vera Rubin), transforming it from a GPU company into a full-stack data center platform provider.

Quarterly revenue of $308M from the PCIe retimers and AI fabric switches inside hyperscaler GPU racks, anchored by a $6.5B Amazon purchase commitment.

Astera designs semiconductor-based connectivity products for rack-scale AI infrastructure. Its platform spans PCIe/CXL retimers, Ethernet smart cable modules, memory controllers, and a growing family of smart fabric switches and optical engines — all integrated with the COSMOS software suite. These components connect GPUs, CPUs, memory, and storage inside hyperscaler data centers, making them critical for scaling AI training and inference clusters.

Cloud segment delivered $3.0B (89% of revenue) in nearline HDDs storing AI training data, with long-term agreements extending to 2029 and a greater-than-25% exabyte growth CAGR.

Western Digital is a pure-play hard disk drive manufacturer, producing high-capacity nearline HDDs for hyperscale data centers, client drives for PCs, and consumer external storage. Its current-generation ePMR drives deliver up to 32 terabytes with UltraSMR technology, while a 40TB ePMR and a 44TB HAMR drive are in qualification and ramp phases, respectively. The company also offers data storage platforms (JBODs) for tier-2 cloud providers and high-bandwidth drives designed for AI workloads.

Management disclosed over 80% of $10.2B quarterly revenue comes from AI GPU-related platforms, shipping 6,000+ racks per month with proprietary liquid cooling.

Super Micro Computer designs, manufactures, and deploys high‑performance server and storage systems, with an increasing focus on total data center solutions for artificial intelligence, cloud, enterprise, and edge computing. It integrates GPUs from NVIDIA, AMD, Intel, and Arm into servers and offers a broad portfolio of rack‑scale solutions, liquid cooling, power subsystems, networking switches, and management software. Over 80% of revenue comes from AI GPU‑related platforms, making it an AI infrastructure pure‑play.

Revenue surged 90% to $808M on the laser chips and optical switches connecting AI clusters, validated by NVIDIA's $2B equity investment and a multibillion-dollar optical circuit switch agreement.

Lumentum designs and manufactures optical components—primarily indium phosphide (InP) laser chips, optical circuit switches, and transceivers—that are essential to the high-speed fiber connections inside AI data centers. Its EML and CW lasers drive the transceivers that link GPUs and switches; its Optical Circuit Switches replace electrical spine switches with optical fabrics, enabling more efficient scaling. The company is one of the few to own captive InP wafer fabs, giving it control over the supply of a critical bottleneck component.

Data Center revenue accelerating from 46% growth to a guided 55% on custom AI accelerators, PAM4 interconnect DSPs, and switching silicon, with a raised $16.5B FY28 revenue target.

Marvell Technology is a fabless semiconductor company that designs the critical chips powering AI data centers. Its portfolio spans custom AI accelerators (XPUs) built for hyperscale customers, high-speed optical interconnects moving data between GPU clusters, and Ethernet switches forming the network backbone. As large language models scale and cluster sizes grow, the need for faster, more efficient data movement and specialized compute has made Marvell’s products essential to the AI infrastructure buildout.

FY2026 revenue guided to $19B manufacturing 400G/800G/1.6T networking switches and Google TPU systems, with three hyperscalers each over 10% of revenue.

Celestica designs and manufactures high-speed networking switches, AI/ML compute systems, and rack-scale integrated systems for hyperscale data centers. Its Hardware Platform Solutions (HPS) group co-designs next-generation architectures, including 1.6T and co-packaged optics switches, providing critical infrastructure for AI workloads. The company is the preferred manufacturing partner for Google's TPU systems and serves major hyperscale cloud providers including Google, Amazon, and Meta.

Disclosed 71% of revenue from wafer-level burn-in systems screening AI processors before advanced packaging, with a $41M hyperscaler production order and FY2027 guidance implying up to 200% growth.

Aehr Test Systems designs and manufactures burn‑in and test systems used to screen semiconductor devices for early‑life failures. Its equipment stresses chips at high temperatures and power levels to simulate years of operation in days, catching defects before the chips are packaged or shipped. As AI accelerator power and transistor density rise, burn‑in becomes essential for yield and reliability, and Aehr’s systems are among the few capable of testing wafers at thousands of amperes.

AI-related demand surged to 70% of record $1.282B quarterly revenue, up from just 10% compute share in 2023, testing the AI accelerators, HBM, and silicon photonics powering data centers.

Teradyne makes automated test equipment (ATE) that validates semiconductors — from AI accelerators and high‑bandwidth memory to networking silicon and hard‑disk drives — and it sells collaborative robotic arms and autonomous mobile robots under the Universal Robots and MiR brands. Its test systems are the final quality gate before chips ship, making it indispensable in the AI hardware chain. In Q1 2026, semiconductor test contributed 87% of revenue, robotics 7%, and product test 6%.

NVIDIA's $2B strategic investment validates its role supplying the 800G/1.6T optical transceivers and indium phosphide lasers connecting AI clusters, with InP capacity quadrupling by end-2027.

Coherent is a vertically integrated manufacturer of photonic components and systems. Its products — optical transceivers, indium phosphide lasers, optical circuit switches, and co-packaged optics — provide the high-bandwidth, low-latency interconnects essential for scaling AI data centers. The company supplies nearly every major hyperscaler and system OEM, anchored by a strategic partnership with NVIDIA.

Book-to-bill hit 2.9x and backlog doubled to $15B. The CEO stated 'we're still in the early stage of the infrastructure build out for AI.'

Vertiv designs, manufactures, and services critical infrastructure for data centers, communication networks, and commercial sites worldwide. Its products span power management (UPS, switchgear, busway), thermal management (air and liquid cooling, chillers, CDUs), integrated modular solutions, and lifecycle services. In the AI era, where power densities and heat loads are extreme, Vertiv’s equipment is essential to build and operate AI data centers.

Data center work reached 56% of $2.9B quarterly revenue, up from 33% eighteen months ago, with its electrical segment growing 88% same-store and $12.5B in record backlog.

Comfort Systems USA is a national provider of mechanical and electrical contracting services. It designs, installs, and services HVAC, plumbing, piping, controls, electrical systems, and fire protection for commercial, industrial, and institutional buildings. In recent years, a dominant share of its work has shifted to advanced technology projects—specifically, the construction of hyperscale data centers.

Data center revenue is 56% of quarterly sales and supply-constrained, with HBM TAM pulled forward two years to $100B by 2028.

Micron Technology designs and manufactures DRAM, NAND flash, and NOR memory chips, along with solid-state drives and multi-chip packages, under the Micron and Crucial brands. Its memory products are foundational to every tier of the AI stack, from high-bandwidth memory (HBM) for GPU accelerators to high-capacity DDR5 for cloud servers and LPDDR for AI-enabled edge devices.

Revenue guidance raised twice in six months to 40%+ growth, manufacturing nearly every major AI accelerator on N3/N2 nodes while spending $60-64B in capex.

TSMC operates as a pure‑play foundry, fabricating integrated circuits on silicon wafers for customers who design but do not manufacture chips. Its process portfolio spans mature specialty nodes to the most advanced gate‑all‑around N2 technology, and it also provides design enablement, mask‑making, advanced packaging (3DFabric), and testing. The company is the sole‑source manufacturer for nearly every high‑performance AI chip, from NVIDIA’s GPUs to custom ASICs from Broadcom, AMD, Amazon, Meta, and Google.

Basics brand revenue surged 72% selling AI-centric free cooling chillers and liquid cooling CDUs to data centers, with book-to-bill over 2.0 and backlog up 160%.

AAON designs and manufactures highly configurable HVAC and liquid‑cooling systems for commercial, industrial, and data‑center environments. Its BASX segment specializes in thermal management for hyperscale data centers, offering airside cooling units, liquid Coolant Distribution Units, and AI‑centric free cooling chillers. The company’s engineered‑to‑order approach and broad product suite position it across the data‑center cooling spectrum, from air to liquid.

2026 revenue guidance raised EUR 5-7B driven by AI demand, as the sole EUV lithography supplier required to manufacture every advanced AI chip.

ASML designs and manufactures the lithography systems that print circuit patterns on silicon wafers. Its extreme ultraviolet (EUV) machines are essential for producing the most advanced logic and memory chips, including AI accelerators and high-bandwidth memory. The company also provides deep ultraviolet (DUV) immersion and dry lithography, together with metrology, inspection, and a growing installed‑base services and upgrades business.

AI semiconductor revenue hit $10.8B (49% of total, +143% YoY), designing custom accelerators for six hyperscalers while guiding to $56B in FY2026 AI revenue.

Broadcom designs and supplies a vast portfolio of semiconductor and infrastructure software products. Its semiconductor segment spans custom AI accelerators (XPUs/TPUs), Ethernet switches, SerDes, optical components, RF filters, and storage connectivity — chips that underpin hyperscale data centers, enterprise networking, and broadband.

Data Center revenue grew 57% to $5.8B quarterly as Meta committed to 6 GW of Instinct GPUs and the AI server CPU TAM was revised upward to over 35% CAGR through 2030.

AMD designs high-performance server CPUs (EPYC), AI accelerators (Instinct GPUs), and system-level platforms (Helios) that are central to the AI infrastructure buildout. Its server CPUs are used for orchestration and data management in AI clusters, while its GPUs train and run inference on large AI models. The company is expanding from selling individual processors to providing integrated rack-scale systems that combine CPU, GPU, and networking, making it a deeper partner to hyperscale customers.

Disclosed AI fabric revenue target of $3.5B (30% of sales, more than doubling YoY), providing the 800G EtherLink switches forming the Ethernet backbone of AI clusters.

Arista Networks develops high‑speed networking platforms built on its Extensible Operating System (EOS), a single software binary that runs across its entire switch and router portfolio. This unified stack enables consistent automation, telemetry, and security, making it especially valuable for hyperscale and AI operators that must manage massive, distributed fabrics. The company’s Ethernet solutions span data center leaf‑spine fabrics, inter‑data‑center routing, campus switching, and WAN connectivity.

Data center orders surged 240% YoY with AI now 50% of new orders, deploying a grid-to-chip franchise spanning Boyd liquid cooling, solid-state transformers, and a $1.5B capacity ramp.

Eaton is an intelligent power management company whose grid‑to‑chip portfolio spans power distribution, power quality, and liquid cooling for AI data centres, along with aerospace and electrical systems for commercial, industrial, and utility markets. Its products—from medium‑voltage switchgear to cold plates and integrated modular platforms—form the infrastructure layer between the power grid and the high‑density computing that AI demands.

Optical Communications revenue grew 36% with three multi-billion-dollar hyperscaler supply agreements, providing the high-density fiber and connectivity backbone of AI data centers.

Corning is a materials‑science company that creates category‑defining products from glass, ceramics, and optical physics. Its core AI‑relevant business is Optical Communications, where it produces optical fiber, cable, and high‑density connectivity systems that form the physical backbone of hyperscale data center networks. These products enable AI model builders to pack more bandwidth into existing infrastructure, reduce total installed cost, and connect tens of thousands of accelerators. The company also supplies advanced optics for EUV lithography tools — essential for manufacturing leading‑edge AI chips — and is the dominant manufacturer of display glass substrates.

Electrification data center orders hit $2.4B in Q1 alone, exceeding the full-year 2025 total, while gas turbine gigawatts under contract jumped to 100 GW with 20% directed at data centers.

GE Vernova designs, manufactures, and services the heavy‑duty gas turbines, HVDC transmission systems, transformers, and switchgear that form the backbone of the world’s electricity grid. With an installed base generating approximately 25% of global electricity, the company’s equipment is critical to meeting the surge in power demand from AI data centers. Its gas turbines, in particular, are increasingly deployed for dedicated data‑center campuses, while its Electrification segment provides the substations and grid automation that connect those campuses to the grid and optimize power flows.