The market’s faith in artificial intelligence was shaken, tested, and then reaffirmed this week exposing just how much global risk appetite still depends on a handful of megacaps and the promise of AI-driven growth. But when Nvidia’s results arrived, they delivered both vindication and a reminder of the fragility of sentiment. The chipmaker’s blowout forecast briefly pulled indexes higher, but enthusiasm turned to anxiety and back again in the span of a single session. We believe Nvidia might be the sole winner in the current AI capex megacycle. We look at the F3Q26 results and analyse what these numbers say what we believe might be the story of next 5 years. Key observations are as following: 1. Management provided solid evidence that addressed several key investor questions. 2. Management sees potential upside to its prior Data Center revenue outlook of $500 bn in 2025/26. 3) Rubin remains on track for introduction in mid-2026 and a strong revenue ramp in 2H26. 4) The company sees an extended useful life for its past generation GPUs, with most of its Ampere series GPUs (shipped up to six years ago) still running near full utilization. Nvidia reported revenue of $57.0 bn, well above street estimates at $55.4 bn. Gross margin came at 73.6% against street at 73.7%. In fact the company expects to be able to hold its gross margin at ~75% throughout FY27 as well. AI GPUs are likely to represent on average high 80% to 90% of total NVDA data center sales in FY2026-FY2027 which is a high margin business. There is huge demand for Blackwell & Rubin as can be seen in total inventory which is up 132% Q/Q at $19.8Bn. The company has positioned itself much ahead in time for the anticipated demand. The core of the company, the data center business delivered quarterly revenues of ~$43b versus market estimates of ~ $42b (with Blackwell at +46% q/q to ~$41b in revs and Hopper/all else down from ~$6b to $2b in the quarter). The management reiterated its prior expectation that it sees over $500 bn of customer demand for its Data Center products across both Compute and Networking products for its Blackwell and Rubin generations and the company sees the potential for upside to this figure based on incremental customer orders. Nvidia continues to see a path to $3 -4 TN in annual AI infrastructure spending by 2030, and it expects to capture significant share of this market. Just like the dot com era, many companies will go under but AI is here to stay just like the internet stayed. Growing hardware demands of AI infrastructure are driving a shift from asset-light to asset-heavy business models, marking a significant change in the technology sector. With robust CapEx spending expected to continue and significant investments needed to support an AI-enabled future, Nvidia is at the right place at the right time with the right mix of technology and capital. We might debate the likes of Oracle & Palantir’s place in AI’s future but Nvidia is different. They are the factory of building blocks of what will be required for AI applications in future.
