What Marvell makes for data centers
public · HQ Santa Clara, United States · marvell.com
Marvell designs custom AI accelerator silicon (ASICs) for hyperscalers, including chips used in AWS Trainium/Inferentia and Microsoft Maia, and sells NICs, DPUs (OCTEON), and CXL memory-expansion controllers and switches used in AI servers. It also makes the optical DSPs, silicon photonics and co-packaged optics components, and copper cable retimers used in data center networking, plus switch ASICs (Teralynx) for scale-out fabrics.
Listing
US exchanges (NYSE/Nasdaq): MRVL· 229 USD +3.65% · mkt cap 205.4B USD
Snapshot 2026-10-25T10:41:17Z — nightly, not real-time.· Chart on TradingView ↗
Exposure
13 parts across 3 layers
No disclosed data-center revenue share on file.
Parts, by layer
Layer 1 — Compute silicon & packaging
| Part | Role | Confidence | Competitors | Note |
|---|---|---|---|---|
| Hyperscaler custom ASICs | design partner | inferred | Alphabet (Google), Microsoft, Alchip, Amazon (AWS), Broadcom, MediaTek | AWS Trainium/Inferentia; Microsoft Maia — partner on AWS Trainium/Inferentia; partner on Microsoft Maia |
| ASIC design services, IP & EDA | service | inferred | Alchip, Arm, Broadcom, Cadence, Credo, Faraday | |
| NICs / SuperNICs | manufacturer | inferred | AMD, Broadcom, Intel, Nvidia | |
| DPUs | manufacturer | inferred | Microsoft, AMD, Intel, Nvidia | OCTEON |
Layer 2 — Memory & storage
| Part | Role | Confidence | Competitors | Note |
|---|---|---|---|---|
| CXL controllers & memory expanders | manufacturer | inferred | Astera Labs, Microchip, Micron, Montage, Rambus, SK Hynix | |
| PCIe/CXL switches & retimers | manufacturer | inferred | Astera Labs, Broadcom, Microchip, Xconn | |
| Controllers | manufacturer | inferred | Kioxia, Microchip, Micron, Phison, Samsung Electronics, Silicon Motion |
Layer 3 — Networking & interconnect
| Part | Role | Confidence | Competitors | Note |
|---|---|---|---|---|
| Scale-out switch ASICs | manufacturer | inferred | Broadcom, Cisco, Nvidia, Huawei | Teralynx |
| Optical DSPs, drivers, TIAs | manufacturer | inferred | Broadcom, Credo, MaxLinear | DSPs |
| Silicon photonics & co-packaged optics | manufacturer | inferred | TSMC, Broadcom, GlobalFoundries, Intel, Nvidia, Tower Semiconductor | |
| DCI / coherent optics | manufacturer | inferred | Adtran, Ciena, Cisco, Coherent, Lumentum, Nokia | |
| Timing & telemetry silicon | manufacturer | inferred | Broadcom, Microchip, Renesas, SiTime | |
| Retimers inside AECs | manufacturer | inferred | Astera Labs, Broadcom, Spectra7 |
Where Marvell sits in the stack
Marvell's placements in this catalog cover 13 parts, from layer 1 (compute silicon & packaging) to layer 3 (networking & interconnect). None of them currently carries a sourced supply constraint tag. Across these parts it overlaps with other companies most often on the lines below.
What these parts are
- Hyperscaler custom ASICs — AI accelerator chips (often called XPUs) designed in-house by a hyperscaler for its own fleet, typically co-developed with an outside ASIC design partner and fabbed by a merchant foundry.
- ASIC design services, IP & EDA — Outside firms and licensors — design-services houses, IP blocks, and EDA software — that hyperscalers and accelerator startups use to turn a chip specification into a manufacturable design.
- NICs / SuperNICs — Network interface cards, including Nvidia's higher-bandwidth "SuperNIC" variants, that connect a server's CPU and GPUs to the cluster network.
- DPUs — Data Processing Units that offload networking, storage, and security functions from the host CPU onto a dedicated, programmable chip.
- CXL controllers & memory expanders — Chips that let a server attach an extra pool of DRAM over a CXL link rather than only through the CPU's own memory channels, expanding usable capacity beyond what fits in local DIMM slots.
- and 8 more in the table above.
Closest peers in this catalog
Seen in real machines
Marvell is named in 1 of the rack teardowns and supercomputer breakdowns on this site. Each line below is a piece of that machine.
- Nvidia GB300 NVL72 · rack teardown
- 800G / 1.6T OSFP transceivers (scale-out)
Disclosed customers
| Customer | % | Metric | Period | Filing |
|---|---|---|---|---|
| largest customers (combined) | 75% | receivables | 2026-05-02 | 10-Q |
| largest customers (combined) | 10% | revenue | 2026-05-02 | 10-Q |
| largest customers (combined) | 73% | receivables | 2026-01-31 | 10-K |
| largest customers (combined) | 10% | revenue | 2026-01-31 | 10-K |
| largest customers (combined) | 70% | receivables | 2025-11-01 | 10-Q |
| largest customers (combined) | 10% | revenue | 2025-11-01 | 10-Q |
Latest feed
- Marvell Unveils 2nd-Gen Structera X with Six-channel CXL 3.2 Memory Expanders for AI Inference — HPCwire, 2026-10-12Marvell has announced the second generation of its Structura X family, featuring six-channel CXL 3.2 memory expanders designed to support AI inference workloads.
- Marvell Structera X 3406 and 3506 CXL 3.2 Memory Expanders — ServeTheHome, 2026-10-12Marvell has launched its Structura X 3406 and 3506 memory expanders, which support CXL 3.2 for high-speed memory expansion and pooling.
- CyberTAN enters 1.6T AI switch race with Marvell-backed platform — DigiTimes, 2026-10-12CyberTAN has launched a new platform designed for 1.6T AI switching, supported by Marvell Technology.
- Marvell's custom AI chip upgrade points to a broader fight over peripherals — DigiTimes, 2026-10-08Marvell is highlighting the importance of custom silicon and high-speed interconnect peripherals in the broader AI infrastructure landscape.
- Marvell raises 2028 revenue outlook to US$20 billion as AI connectivity expands — DigiTimes, 2026-10-07Marvell Technology raised its 2028 revenue outlook to $20 billion, driven by growing demand for AI connectivity infrastructure.
- GlobalFoundries and Marvell sign multi-year development agreement for optical technologies — Data Center Dynamics, 2026-09-21GlobalFoundries and Marvell have entered into a multi-year agreement to expand silicon germanium capacity at the GlobalFoundries facility in Burlington, Vermont.
- Marvell pushes GlobalFoundries to light up wafer production — The Register, 2026-09-17Marvell stated that GlobalFoundries must improve its supply side to support its ambitious growth plans.
- Weekly news roundup: AI memory race heats up, Taiwan manufacturing rebounds, Nexperia rift deepens — DigiTimes, 2026-09-07The week's AI infrastructure news covered semiconductor supply chains, including Nexperia's legal issues, Sandisk's HBF, Micron's HBM expansion, and Marvell's silicon photonics roadmap.
- CPO material race widens as Taiwan eyes key role — DigiTimes, 2026-09-02Industry experts discussed that silicon photonics bottlenecks are shifting from materials to advanced packaging precision, platform selection, and necessary manufacturing tools.
- Marvell sees silicon photonics scaling from late 2027, with Taiwan and TSMC at the center — DigiTimes, 2026-09-02Marvell Technology expects silicon photonics to begin scaling in late 2027, highlighting the importance of the Taiwan supply chain and TSMC's manufacturing capabilities.
Questions about Marvell
- What does Marvell make for AI data centers?
- This catalog places Marvell on 13 parts: Hyperscaler custom ASICs (design partner), ASIC design services, IP & EDA (service), NICs / SuperNICs (manufacturer), DPUs (manufacturer), CXL controllers & memory expanders (manufacturer), PCIe/CXL switches & retimers (manufacturer), Scale-out switch ASICs (manufacturer) and Optical DSPs, drivers, TIAs (manufacturer), and 5 more.
- Is Marvell publicly traded?
- Yes — Marvell is listed on US exchanges (NYSE/Nasdaq) under MRVL. Prices shown on this page are a nightly snapshot, not real-time, and nothing here is investment advice.
- Where is Marvell based?
- Santa Clara, United States, per the company record in this catalog.
- Who else makes hyperscaler custom ASICs for AI data centers?
- Alphabet (Google), Microsoft, Alchip, Amazon (AWS), Broadcom and MediaTek are also placed on Hyperscaler custom ASICs in this catalog.
- How well sourced are these placements?
- Every part→company placement carries a confidence tag. For Marvell the mix is 13 inferred. "disclosed" means the company or a filing says so; "reported" comes from trade press or a reference-design list; "inferred" is a lead from the taxonomy with no citation yet. Treat reported and inferred as leads, not facts.
Answers are generated from this catalog's own rows — see sources & method.
Verification: enriched, last verified 2026-08-27. Sources: www.marvell.com · en.wikipedia.org.· Markdown · JSON