Microsoft's coverage in our briefs is less spot-tape-visible than AWS's (Azure does not publicly list Trainium-equivalent spot prices), but the company sits inside several material theses through customer relationships and the Maia ASIC roadmap.
The Azure Maia / Broadcom partnership. Microsoft is one of four hyperscalers (alongside Google, Meta, Anthropic+OpenAI) named as a Broadcom XPU partner. The May 13 thesis on Broadcom's $20bn 2025 → $100bn+ 2026 AI revenue trajectory rests on these four customer relationships. The size of Microsoft's Maia commitment over the next 18 months drives a material share of Broadcom's upside case.
CoreWeave customer concentration. CoreWeave's Q1 2026 10-Q discloses "significant customer concentration risk" with Customer A and Customer B representing material portions of revenue. Microsoft is widely believed to be one of these two — making MSFT's in-house ASIC progress (Maia + Cobalt) a direct headwind for CoreWeave's H100 rental utilisation. Per the May 12 brief: "Every dollar spent on in-house chips is a dollar not spent renting CoreWeave GPUs." Microsoft's buying behaviour effectively determines half of CoreWeave's revenue trajectory.
The training-cluster demand side. Per the May 17 brief: hyperscaler capex (Morgan Stanley estimate) projected at +80% to $805bn in 2026, dominated by training infrastructure. Azure is a meaningful share. Microsoft's OpenAI partnership and its own frontier-model work (Phi family, MAI / Copilot infrastructure) continue to absorb Nvidia GPUs at scale through the medium term, even as Maia ramps for inference workloads.
The implicit B200 reserve. Although Microsoft does not publicly quote B200 spot pricing, the company has reportedly committed to large B200 / GB200 NVL72 rack volumes. Limited public visibility makes this a "wait for earnings" read rather than a spot-tape read.
The composite picture: a hyperscaler with significant but partially-opaque positioning across all four weekly themes — supplier (CoreWeave customer), substitute (Maia), training-side demand (Azure capacity buildout), power (Azure regional capacity plans). Most material near-term catalyst is Q1 2026 Azure AI revenue commentary in next Microsoft earnings.