Overview
- Nvidia reported a record quarter on Aug. 26 with $96.22 billion in revenue and $89 billion from data centers, and for the first time issued a year‑ahead FY28 revenue guide of about 70%.
- Management says that guidance reflects supply limits rather than demand, explicitly excluding China data‑center revenue and pointing to rising memory costs and packaging capacity as the main bottlenecks.
- Nvidia is moving beyond standalone GPUs into a platform that includes the Vera Rubin accelerator (about 4,000 TFLOPS FP16 per GPU), server CPUs in the Grace family, and rack systems that bundle compute, networking and software.
- The company is mobilizing third‑party capital and deals with large asset managers to speed data‑center builds and has disclosed partnerships to channel hundreds of billions in infrastructure financing while accepting greater counterparty and balance‑sheet exposure.
- Competitors and suppliers are winning parts of the expanding market—AMD, Intel and others are growing their data‑center businesses while firms such as Seagate, Micron and Dell benefit from the broader AI capex wave—leaving Nvidia dominant at the highest end but exposed to margin, supply‑chain and geopolitical risks that will shape how durable its platform lead proves.