Marvell Technology, Inc.
Category: AI Chips / Semiconductors
Marvell Technology designs and supplies custom silicon and semiconductor infrastructure — including custom AI accelerators, interconnects, and network switches — powering data centers for hyperscalers like Google, Amazon, and Microsoft. Marvell Technology, Inc. was founded in 1995. The company is led by Matt Murphy. Based in Santa Clara, California, USA. Team size: 500+. Latest round: Post-IPO.
- Founded
- 1995
- Headquarters
- Santa Clara, California, USA
- Team size
- 500+
Value proposition
End-to-end custom silicon co-design and the industry's broadest data-infrastructure semiconductor portfolio (custom AI XPUs, optical/electrical interconnects, switching), enabling hyperscalers to build differentiated AI compute infrastructure at scale.
Products and solutions
Custom AI accelerators/XPUs (e.g. Google Axion collaboration, Amazon Trainium, Microsoft Maia), optical and electrical interconnect DSPs, PAM4/coherent optics, Ethernet switch silicon, storage controllers, automotive Ethernet.
Unique value
Recognized co-design partner of choice for hyperscalers in custom AI silicon, holding an estimated 20-25% share of the custom-ASIC AI chip market alongside deep interconnect/optical IP.
Target customer
Hyperscale cloud providers and large enterprises (Google, Amazon/AWS, Microsoft), telecom/networking OEMs, automotive and enterprise storage/networking customers.
Industries served
Data center / cloud infrastructure, AI/ML compute, telecom & 5G, automotive, enterprise storage, networking.
Technology advantage
10,000+ patents worldwide; deep expertise across custom SoC design, high-speed SerDes/optical interconnect, and advanced packaging; long-standing hyperscaler relationships enabling multi-generation custom silicon roadmaps.
How they differentiate
Positions as the #2 custom-silicon/ASIC co-design partner behind Broadcom, differentiated via combined interconnect/optical DSP portfolio plus custom compute silicon, giving hyperscalers a second-source alternative to Broadcom and NVIDIA merchant GPUs.
Main competitors
Broadcom (dominant custom-ASIC rival, ~60-80% share), NVIDIA (merchant AI GPU/accelerator leader), Qualcomm, MediaTek (in select segments).
Key partnerships
Google (expanded custom AI chip supply agreement incl. $12.2B stock warrant, Aug 2026), Amazon/AWS (custom AI products, Trainium-related silicon, expanded strategic collaboration), Microsoft (custom silicon collaboration, Maia-related).
Notable customers
Google, Amazon (AWS), Microsoft — hyperscale cloud AI infrastructure customers via custom silicon programs.
Major milestones
1995: Founded by Sehat Sutardja, Weili Dai, Pantas Sutardja. 2000: IPO on NASDAQ (MRVL). 2016: Matt Murphy becomes CEO, pivots company toward data infrastructure/custom silicon. 2018: Acquired Cavium. Expanded AWS custom AI silicon collaboration. Aug 2026: Granted Google warrants for up to $12.2B in Marvell stock as part of expanded custom AI chip supply deal (potential $120B in chip sales).
Growth metrics
FY2026 annual revenue ~$8.2B; Q3 FY2026 net revenue $2.075B; data center segment ~76% of revenue; stock up significantly on AI/custom silicon demand.
Market positioning
Leading pure-play custom AI silicon and data-infrastructure semiconductor supplier to hyperscale cloud providers, second only to Broadcom in custom ASIC market share.
Geographic focus
Global, with primary revenue concentration among US hyperscaler customers (Google, AWS, Microsoft) and significant engineering presence in Israel (Petah Tikva) and Asia.
Patents and IP
Over 10,000 patents worldwide (per company disclosures).
About Matt Murphy
Ex-Maxim Integrated (multiple VP/GM roles); Ex-Skyworks Solutions; B.A. Franklin & Marshall College; Stanford Executive Program graduate. Led Marvell's transformation into a $100B+ data infrastructure/AI leader since becoming CEO in 2016.
Latest news about Marvell Technology, Inc.
More AI Chips / Semiconductors companies
Official website: https://www.marvell.com