04Buildout

The gigawatt land-grab

Follow the money: what hyperscalers are spending, what NVIDIA is booking, and where the gigawatt-class campuses are actually rising.

Big-4 capex guidance · 2026

$710B

≈1.7× 2025 actuals · MSFT + GOOG + META + AMZN midpoints

NVIDIA DC revenue · FY27 Q1

$75.2B

~$300B/yr run-rate

Tracked frontier pipeline

~32 GW

Sum of planned capacity, clusters below

Largest single campus

Hyperion

Meta · up to 5 GW

Big-tech capital expenditure

US$B per calendar year (2026 = guidance midpoints)

Static chart frame · exact values in table view

Earnings releases & SEC filings, as of Jul 2026. Microsoft incl. finance leases; Oracle fiscal years mapped to prior calendar year.

NVIDIA data-center revenue

US$B by fiscal quarter (FY ends late January)

Static chart frame · exact values in table view

NVIDIA quarterly releases. Next-quarter guidance ~$91B total revenue, assuming zero China DC sales.

Gigawatt-class clusters & campuses

Frontier training/inference sites, status as of Jul 2026. Power and chip counts are best public estimates — open Context for project notes.

Frontier-scale AI datacenters
		Location		Chips
Stargate US expansionest — estimated / secondary-source figure Context for Stargate US expansion >$450B committed; Michigan site added Jun 2026	OpenAI · Oracle · SoftBank	TX · NM · OH · MI (6+ sites)	— → 8 GW	GB300-era NVIDIA	Building
Hyperionest — estimated / secondary-source figure Context for Hyperion Entergy power deal expanded Mar 2026 (~5 GW of new gas generation)	Meta	Richland Parish, Louisiana	2 GW → 5 GW	Multi-GW campus (chips TBD)	Building
Anthropic–Google TPU fleetest — estimated / secondary-source figure Context for Anthropic–Google TPU fleet Oct 2025 deal; expanded with Google/Broadcom Apr 2026	Anthropic · Google Cloud	Multiple US sites	1 GW → 5 GW	Up to 1M TPU v7 (Ironwood); +3.5 GW next-gen TPUs from 2027	Building
Stargate UAEest — estimated / secondary-source figure Context for Stargate UAE First 200 MW due Q3 2026; anchor of a 5 GW campus	G42 · OpenAI · Oracle	Abu Dhabi, UAE	200 MW → 5 GW	GB300-class	Building
Project Rainierest — estimated / secondary-source figure Context for Project Rainier Amazon–Anthropic expansion: >$100B over 10 years, up to 5 GW	Anthropic · AWS	New Carlisle, Indiana	1 GW → 2.2 GW	>500K Trainium2; ~1 GW Trainium2+3 by end-2026	Operating
Colossus 2est — estimated / secondary-source figure Context for Colossus 2 First single ~gigawatt-class AI datacenter; targeting 1M GPUs	xAI	Memphis, Tennessee	1 GW → 2 GW	110K+ GB200; ~555K GPUs across the Memphis complex	Operating
Fairwater (WI + Atlanta)est — estimated / secondary-source figure Context for Fairwater (WI + Atlanta) Sites linked into one “AI superfactory” WAN; Wisconsin facility live Jun 2026	Microsoft	Mount Pleasant, WI · Fayette County, GA	900 MW → 2 GW	Hundreds of thousands of GB200/GB300	Operating
Stargate Abileneest — estimated / secondary-source figure Context for Stargate Abilene ~4 of 8 buildings live mid-2026; flagship of the Stargate program	OpenAI · Oracle · Crusoe	Abilene, Texas	1.2 GW	~450K GB200 at full buildout	Building
Prometheusest — estimated / secondary-source figure Context for Prometheus First cluster in service since late 2025; ~1 GW through 2026	Meta	New Albany, Ohio	1 GW	GB200 + GB300 systems	Operating
Colossus 1est — estimated / secondary-source figure Context for Colossus 1 Built in 122 days in 2024	xAI	Memphis, Tennessee	300 MW	~230K GPUs (150K H100, 50K H200, 30K GB200)	Operating