CoreWeave's top customer went from 67% of revenue in FY2025 to 36% in Q2 2026. Diversification, or the same few buyers?

JPMorgan upgraded CRWV to Overweight on Sept. 24 (target **$125** from $120), citing AI compute demand and premium pricing on short-term contracts. Before reading much into that, look at who pays the bills.

**FY2025.** The 10-K (filed 2026-03-02) says: "We recognized an aggregate of approximately **67%** of our revenue from our top customer, Microsoft, for the year ended December 31, 2025." That was up from **62%** in 2024.

**2026.** The 10-Qs give percentages but no names:

* Q1: Customer A: **45%**, Customer B: **20%**.
* Q2: Customer A: **36%**, B: **26%**, C: **10%**.

The filings say the letters "may represent different customers" across periods, so I infer that Customer A is Microsoft.

**What I take from it.**

The top customer's share roughly halved in two quarters, but its dollars look flat: **45%** of Q1 revenue ($2.08B) and **36%** of Q2 revenue ($2.58B) both come to about $0.93B. The share fell because other customers grew.

**Less diversified than it looks.**

The top three still accounted for about **72%** of Q2 revenue, and the Q2 10-Q says roughly **93%** of the revenue increase came from existing customers. One very large buyer has become three large buyers.

**Filings:**

* [10-K](https://www.sec.gov/Archives/edgar/data/1769628/000176962826000104/0001769628-26-000104-index.htm)
* [Q1 10-Q](https://www.sec.gov/Archives/edgar/data/1769628/000176962826000222/crwv-20260331.htm)
* [Q2 10-Q](https://www.sec.gov/Archives/edgar/data/1769628/000176962826000366/crwv-20260630.htm)

For people who follow this name closely: **Are the premium short-term contracts going to new buyers, or to the same three on different terms?**