Nvidia is reportedly in talks to backstop OpenAI's US$500 billion data center campus in Pike County, Ohio with a US$250 billion lease guarantee, going beyond its role as the campus's GPU supplier; the development is part of the Stargate AI infrastructure initiative and would become the world's largest data center by power capacity if completed

US-based neocloud and inference startups building the compute layer for AI drew more than US$9bn in venture capital in 2025-2026, concentrated outside the major hyperscalers.

The US chip designer competing for second place in AI accelerators, with a US$34.6bn 2025 revenue base and a 6GW GPU supply pact with OpenAI.

Hyperscalers plan US$725 billion in 2026 capex for GPU campuses, chip contracts, and power deals that determine who can train and serve frontier AI models.

AI data centres globally consume billions of gallons annually for server cooling, triggering water-rights conflicts across the US Southwest, Spain, and other drought-stressed regions.

Texas's independent grid operator managing 90% of the US state's electricity load for 26 million customers, central to data-center demand and grid reliability debates.

The US GPU designer whose chips run more than 90% of global AI model training, making it the world's most strategically contested semiconductor company.

The largest US electric grid operator, coordinating wholesale power for 67 million people across 13 states and DC, now under acute strain from AI data-center demand.

Nuclear reactors up to 300 MWe per module, factory-built for speed; the United States, United Kingdom, and China are racing to commercial operation as AI data centres create demand.

Champlain Hudson energises and Cimarron Link and Southern Spirit progress, but planners say new lines still lag load growth and renewables

A 18 May emergency order lets the operator cut large loads before rolling blackouts; PJM warns the 2026 summer leaves little reliability margin

Data-center developers and European utilities lock in small modular reactors years before first concrete, betting on factory-built nuclear to feed compute and grids

Six months after Nvidia's $20B licence-and-hire deal stripped out its founders, Groq bets its second act on cloud, not chips

The PUCT approved a new large-load process on 18 June; it allocates connections by transmission capacity and financing, not first-come-first-served

A developer sues the Imperial Irrigation District for ~260m gallons/yr from the drought-stricken river, one of a wave of suits, wells and settlements over data-center thirst

Wholesale power on the largest US grid nearly doubled year-over-year; states impose large-load tariffs while a foreign-influence subplot complicates the politics

Data-center load pushes the 2026 summer forecast past the 85.5GW record; ERCOT puts the odds of a June emergency at 0.09%

Related Digital breaks ground on three buildings on 250 acres; one of five new US sites pushing Stargate to ~7GW planned and >$400bn over three years

Data-center revenue ~$72.8bn; Vera Rubin ramps to full production by fall; guidance assumes zero China data-center compute, and the stock still slid

After the council rejected the Amazon-linked complex over desert water use, the city revoked a meter for improper use; Arizona regulators still cleared two project wells

Developers bypass grid queues with behind-the-meter gas: a 7.7GW West Texas permit, NextEra's 10GW, an Oracle–Bloom 2.8GW fuel-cell deal, and GE Vernova's 100GW turbine backlog

A $27bn Blue Owl-funded megacampus in Richland Parish triples its on-site gas to ~7.7GW, about seven times New Orleans' peak demand, via an expanded Entergy deal

Jensen Huang says orders are in and manufacturing is 'restarting'; BIS revised the license policy, but as of mid-2026 only a small number of H200s have actually shipped

A 50km spool that cost $300 now runs ~$2,500, the same bend-insensitive fibre feeds both FPV cables and hyperscaler racks

Nvidia invests in OpenAI; OpenAI commits to Oracle clouds; Oracle buys Nvidia chips, a self-referential money wheel as Oracle/Microsoft CDS spreads roughly double

Two Fairwater sites stitched by a dedicated AI wide-area network train one model in tandem, hundreds of thousands of Nvidia GB200/GB300 GPUs, 140kW racks, no UPS or gen-sets

Indiana's $11bn AWS megacampus runs ~500,000 Trainium2 chips for Anthropic; the partners now target up to 5GW, with Amazon guiding to ~$200bn capex in 2026

Meta's Talavera site would draw ~504m litres/yr mostly from the Tagus; Amazon seeks a 48% water rise in Aragón; Lleida banned data centers as drought collides with the AI build

First 1GW of Instinct MI450 deploys 2H 2026; the Helios rack targets ~3 AI exaflops with 72 MI455X accelerators, AMD's bid to break Nvidia's data-center grip

The 1.2GW Texas campus runs on Oracle Cloud with first GB200 racks delivered; Crusoe's $1.38bn Series E funds the expansion as SoftBank completes its $41bn into OpenAI