Descriptions:
Bloomberg Tech hosts a Goldman Sachs analyst for a post-earnings breakdown of Microsoft’s AI business, focusing on what the latest quarterly results reveal about the company’s ability to monetize its massive infrastructure investments. The conversation centers on Azure’s AI revenue trajectory, the shift toward consumption-based pricing models (notably GitHub Copilot’s move to per-usage billing for power users), and the challenge of reading ROI signals from the outside given limited visibility into Microsoft’s full monetization stack.
Key data points discussed include Microsoft Copilot reaching 30 million installed seats by quarter end — up from 20 million in the March quarter — and Satya Nadella’s disclosure that GPU data-center live times have been cut by nearly 50% over the past fiscal year, with a full gigawatt of capacity added in the quarter alone and overall capacity on track to double within two years. The analyst also walks through Microsoft’s recent CapEx accounting change and its implications for modeling long-lived versus short-lived assets.
The broader framing is Microsoft’s “discovery value” thesis: the argument that the market has been underpricing Microsoft’s AI progress relative to Google and others, and that the next 12 months will surface improvements that are currently invisible to outside observers. The segment also touches on the industry-wide shift toward nuanced model selection — mixing open-weights, frontier, and internal models — as companies optimize what they extract from their infrastructure spend.
📺 Source: Bloomberg Tech · Published July 30, 2026
🏷️ Format: Interview
