CPU

Where the book uses it

• Chapter 4 — One Rack

  How a datacenter is actually built, from a converted factory in Memphis to a greenfield campus in Texas, and what the clock from first ground to first power really measures.

Sources that use it

• AMD, “AAI 2026: AMD Delivers Full-Stack Compute for the Agentic AI Era,” 23 July 2026 (AMD Helios rackscale solutions “now in production to be deployed by leading AI companies at gigawatt scale”, 72 Instinct MI455X GPUs and 18 sixth-generation EPYC “Venice” CPUs to a rack; “OpenAI expects to bring Helios online beginning in the fourth quarter of 2026, with deployments accelerating throughout 2027”; the release names the MI400 Series and the MI455X, and no MI450). In the Sources
• Microsoft, “FY26 Q4 Earnings Conference Call,” 29 July 2026 (the company’s own posted transcript; Amy Hood, chief financial officer, in prepared remarks: “Commercial remaining performance obligation grew 84% to $678 billion” and “for the full year, our cloud revenue surpassed $214 billion, with nearly 90% from customers outside of frontier model companies”; Hood in the question session: “The investment into land and data center builds is actually quite flexible”; Satya Nadella in the question session, answering UBS on model choice and the protection of corporate IP: “the goal is to have the firm be in control of their own destiny, in terms of what I describe as building their human capital and their token capital … And the models are an input, not some extraction of the knowledge of the enterprise”, and of what Microsoft sells into that goal, “the enterprise design architecture that we are going to evangelize”; Amy Hood, chief financial officer, in prepared remarks on the cloud segment: “Segment gross margins were also impacted by growing GitHub Copilot usage, though margins improved through the quarter with the June business model change to usage-based pricing”; Satya Nadella, chief executive, in prepared remarks on GitHub Copilot: “This quarter, we introduced usage-based billing, and have continued to see Business and Enterprise seat growth, and also significant consumption revenue after the new model went into effect”; Nadella in prepared remarks: “we now have over 30 million paid Microsoft 365 Copilot seats, with net seat adds more than doubling quarter-over-quarter”; Hood in prepared remarks: “paid M365 Commercial seats grew 6% year-over-year”, with no count of the base given on this call; Hood in prepared remarks, on the lives it gives its buildings from fiscal 2027: “effective at the start of FY27, we are extending the estimated useful lives of our datacenters and office buildings, from 15 to 25 years, reflecting our operating history and expected use of these assets”, and “This change affects only the timing of future depreciation and is expected to have a minimal benefit to FY27 operating income.”; Hood on the quarter’s capital expenditure: “Roughly two thirds of our capex was for short-lived assets, primarily CPUs and GPUs as customers increasingly build solutions that leverage both AI and non-AI infrastructure. The remaining spend was for long-lived assets.”; Hood in the guidance for fiscal 2027, “the increased monetization opportunity from adding usage-based billing products alongside per-seat licensing in July” and “acceleration in M365 Commercial cloud revenue growth through this fiscal year”). In the Sources
• Microsoft, “Microsoft announces new supercomputer, lays out vision for future AI work,” 19 May 2020 (a single system with more than 285,000 CPU cores, 10,000 GPUs and 400 gigabits per second of network connectivity “for each GPU server”, built for OpenAI; “one of the top five publicly disclosed supercomputers in the world”, a ranking the release attributes to Microsoft itself; Altman: “If we could design our dream system, what would it look like?”; Altman, the release’s next quoted words after its attribution: “what would it look like? And then Microsoft was able to build it”). In the Sources
• Microsoft, third-quarter fiscal 2026 earnings call, 29 April 2026, transcript published by the company (Satya Nadella: “earlier this week we announced our move to a usage-based pricing model for GitHub Copilot as we align pricing to actual usage and costs”; “Nearly 140,000 organizations now use GitHub Copilot, and enterprise subscribers have nearly tripled year over year”; the products in LinkedIn Talent Solutions that automate sourcing, screening and drafting “have already surpassed a $450 million annualized revenue run-rate”; Amy Hood, chief financial officer, in her prepared remarks on the Intelligent Cloud segment: “gross margin dollars increased 19% and 18% in constant currency” and “gross margin percentage decreased year-over-year driven by continued AI investment and increased GitHub Copilot usage, partially offset by ongoing efficiency gains in Azure”; Hood’s fourth-quarter outlook: Microsoft Cloud gross margin percentage “roughly 64%”, “down year-over-year driven by continued investments in AI and increased GitHub Copilot usage”, then “Just this week we announced a business model transition in GitHub Copilot that will align pricing with usage and value that takes effect on June 1st of this year”; the company’s aka.ms/transcriptfy26q3 short link resolves to this document, a Word file, read again on 25 September 2026; Amy Hood on calendar-2026 capital expenditure, “roughly $190 billion ... includes approximately $25 billion from the impact of higher component pricing”; Nadella, asked by Rishi Jaluria for the mix of seats and consumption three to five years out, “seat-based pricing is just entitlement to some consumption” and “IT budgets are going to have to be reshaped by a combination of business outcomes making their way into IT budgets and maybe reallocation from other line items on the income statement, like OpEx”; Hood, on the quarter’s capital spending by asset life: “roughly two thirds of our capex was for short-lived assets, primarily GPUs and CPUs. The remaining spend was for long-lived assets that will support monetization over the next 15 years and beyond”). In the Sources
• NVIDIA, “Hardware — NVIDIA DGX GB Rack Scale Systems User Guide” (GB200 NVL72: 72 GPUs, 36 CPUs, 18 compute trays, 9 NVLink switch trays; approximately 120 kW), accessed 29 August 2026. In the Sources

And 13 more: search the Sources.

Related terms

• Accelerator
• 122 days
• Abilene
• Blackwell
• Colossus
• GB200 NVL72
• GB300
• GPU
• Greenfield
• H100
• Lien
• Lithography