What Microsoft makes for data centers
public
Microsoft is a publicly listed company. It appears on 4 parts of the AI data center — Hyperscaler custom ASICs, DPUs, Hyperscalers, neoclouds & model labs and Hyperscaler Arm CPUs — in the roles of manufacturer, in house and operator. Those parts sit in Layer 1 — Compute silicon & packaging and Layer 9 — Building shell, site & construction. Its shares trade on US exchanges (NYSE/Nasdaq) as MSFT, with 1 further listing.
Listing
US exchanges (NYSE/Nasdaq): MSFT· 525 USD +0.48% · mkt cap 3.89T USD
Also listed: PPA
Snapshot 2026-10-25T10:41:17Z — nightly, not real-time.· Chart on TradingView ↗
Exposure
4 parts across 2 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 | manufacturer | inferred | Alphabet (Google), Alchip, Amazon (AWS), Broadcom, Marvell, MediaTek | Maia |
| DPUs | in house | inferred | AMD, Intel, Marvell, Nvidia | in-house |
| Hyperscaler Arm CPUs | manufacturer | inferred | Alphabet (Google), Amazon (AWS) | Cobalt |
Layer 9 — Building shell, site & construction
| Part | Role | Confidence | Competitors | Note |
|---|---|---|---|---|
| Hyperscalers, neoclouds & model labs | operator | inferred | Alphabet (Google), Alibaba, Amazon (AWS), Apple, Baidu, KT |
Where Microsoft sits in the stack
Microsoft's placements in this catalog cover 4 parts, from layer 1 (compute silicon & packaging) to layer 9 (building shell, site & construction). 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.
- Hyperscaler Arm CPUs — Arm-based server CPUs designed in-house by a hyperscaler for its own fleet, such as AWS Graviton, Google Axion, and Microsoft Cobalt.
- DPUs — Data Processing Units that offload networking, storage, and security functions from the host CPU onto a dedicated, programmable chip.
- Hyperscalers, neoclouds & model labs — The hyperscalers, neoclouds, and AI labs that occupy data center capacity as tenants or owner-operators, and whose capital spending ultimately drives demand for every layer below this one.
Closest peers in this catalog
Seen in real machines
Microsoft is named in 1 of the rack teardowns and supercomputer breakdowns on this site. Each line below is a piece of that machine.
- Eagle (Microsoft Azure) · supercomputer
- System design, operation and the ND H100 v5 node (8x H100 GPUs, 96 Xeon cores, 1,900 GiB DDR5, 8x 400 Gb/s InfiniBand adapters per VM; about 1,800 nodes in the TOP500 run) (designs, builds and operates the Azure ND H100 v5 cluster and submitted it to TOP500 as "Microsoft NDv5"; TOP500 lists the manufacturer as Microsoft Azure)
Disclosed suppliers
Latest feed
- China-aligned hackers deploy AI agents in massive ConnectWise supply chain attack — Tom's Hardware, 2026-10-23
- Microsoft files for data center campus in Floyd County, Georgia — Data Center Dynamics, 2026-10-22Microsoft has filed plans for a data center campus in Floyd County, Georgia, which will consist of three separate buildings.
- Nscale set to lose Microsoft as anchor tenant over UK power delays, adding uncertainty to IPO — DigiTimes, 2026-10-20The UK neocloud provider Nscale may lose Microsoft as an anchor tenant due to anticipated power delays from the National Grid, which may not have sufficient electricity for the site until after 2030.
- Microsoft abandons UK AI supercomputer as grid struggles to keep up — The Register, 2026-10-19Microsoft has withdrawn from its planned UK AI supercomputer because power grid connection issues are expected to delay the nation's AI infrastructure goals until the 2030s.
- Microsoft signs 20-year power deal with Indiana utility NiSource for planned data center in the state — Data Center Dynamics, 2026-10-19Microsoft has secured a 20-year power agreement with the utility NiSource to support a planned data center facility in Indiana.
- DIGITIMES Insight: Microsoft Maia 300 leads 2nm race, TSMC's US strategy shifts, Intel EMIB-T faces yield challenge — DigiTimes, 2026-10-19Microsoft's Maia 300 is highlighted as a leader in the 2nm race, while TSMC's US strategy is changing and Intel's EMIB-T packaging is reportedly encountering yield issues.
- Meeza and Microsoft launch sovereign private cloud in Qatar — Data Center Dynamics, 2026-10-16Microsoft is launching a sovereign private cloud service using Azure Local infrastructure within Meeza's data centers in Qatar.
- Microsoft plans multi-gigawatt rollout of next-gen Azure Maia AI accelerator racks — The Register, 2026-10-13Microsoft is planning a multi-gigawatt deployment of its next-generation Azure Maia AI accelerator racks.
- Microsoft's 2nm Maia 300 completes cloud AI ASIC lineup, setting up 2027 server supply-chain boom — DigiTimes, 2026-10-13Microsoft has finalized its cloud AI ASIC lineup with the completion of the 2nm Maia 300 chip.
- Microsoft Boosts HBM Bandwidth by 2.5x with Maia 300 Accelerator — HPCwire, 2026-10-12Microsoft announced the Maia 300 accelerator, which features a 2.5x increase in HBM bandwidth for AI workloads.
Questions about Microsoft
- What does Microsoft make for AI data centers?
- This catalog places Microsoft on 4 parts: Hyperscaler custom ASICs (manufacturer), DPUs (in house), Hyperscalers, neoclouds & model labs (operator) and Hyperscaler Arm CPUs (manufacturer).
- Is Microsoft publicly traded?
- Yes — Microsoft is listed on US exchanges (NYSE/Nasdaq) under MSFT, plus 1 other listing. Prices shown on this page are a nightly snapshot, not real-time, and nothing here is investment advice.
- Who else makes hyperscaler custom ASICs for AI data centers?
- Alphabet (Google), Alchip, Amazon (AWS), Broadcom, Marvell 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 Microsoft the mix is 4 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.