Version 5

AI infrastructure spending has reached a scale that exceeds the operating cash flows of the industry's most profitable companies. Google reported its first-ever negative free cash flow quarter in Q2 2026 against record revenue of $119.8 billion [1], with full-year 2026 capex projected at $205 billion [2] and Alphabet's total forward spending commitments at $811 billion [3]. OpenAI is building a $20 billion data center campus in Effingham County, Georgia — Project Camellia — with a 3.2-gigawatt power agreement and pledges covering community benefits, ratepayer protections, and minimal water use [9][11]. The Trump EPA has formally issued permitting guidance framed as promoting data center development while limiting community participation in air pollution reviews [12][13], drawing formal opposition from environmental groups [19][20].

Hyperscalers are committing capital that outpaces their ability to generate cash, on a timeline where the revenue needed to close that gap depends on enterprise AI adoption assumptions that remain unvalidated. The permitting changes shift who gets a say in where and how the associated power infrastructure is built, compressing the window for community review precisely as construction accelerates.

The scale of AI infrastructure spending has reached a point where even the most profitable technology companies cannot fund it from operating cash flows. Google reported record Q2 2026 revenue of $119.8 billion and Cloud revenue of $24.8 billion — up 23.8% from Q1 — yet capital expenditures exceeded all of it, producing the company's first-ever negative free cash flow quarter [1]. Google's full-year 2026 capex is projected at $205 billion [2], and Alphabet's total forward spending commitments reached $811 billion [3]. Evercore analysts warned in February 2026 that hyperscaler capex could exceed operating cash generation before revenue caught up [4]; Semianalysis CEO Dylan Patel predicted in April 2026 that Google's free cash flow would hit zero [5][6]. Both forecasts proved accurate ahead of schedule. Across the industry, hyperscalers are expected to commit at least $625 billion to AI infrastructure in 2026 [7], and IDC projects annual spending will exceed $1 trillion by 2029 [8].

On July 22, 2026, OpenAI disclosed Project Camellia, a $20 billion data center campus planned for Effingham County, Georgia, near Savannah [9]. The project has secured a 3.2-gigawatt power agreement with Georgia Power, to be delivered in phases between 2028 and 2032 [9][10]. OpenAI pledged $80 million in community benefits directed toward schools, public safety, healthcare, and workforce training, cited Georgia Public Service Commission rules as prohibiting infrastructure cost pass-through to residential ratepayers, and committed to minimal water use for the campus [9][11]. An independent annual public audit is promised to verify these commitments [9]. All disclosed details precede groundbreaking.

The Trump EPA has formally issued permitting guidance it describes as furthering the president's agenda of promoting data center development while safeguarding communities [12]. The guidance's dual-track approach includes a proposal to give states authority to determine the extent of public participation in air pollution permitting reviews — potentially limiting public notice for gas plants and diesel generators that support data center operations [13][14] — alongside a companion proposal to allow construction to begin before air permits are issued [15][16]. A law firm analysis frames the agency as stepping back from federal data center regulation and handing authority to states [17], while the Environmental Data and Governance Initiative has documented elevated air pollution near EPA-regulated data centers [18]. The Environmental Defense Fund and Sierra Club, joined by other health and environmental groups, formally opposed the pre-permit construction proposal, arguing it removes the practical window in which community and regulatory review can affect a project's design or siting [19][20].

The build-out is extending into telecom infrastructure. Verizon signed a deal worth more than $1 billion with Google to connect data centers using its existing dark fiber network, and is repurposing legacy central offices as edge data centers optimized for AI inference workloads targeting robotics, remote surgery, and autonomous driving [21]. Verizon's CEO expects additional AI-related deals worth multiple billions to be announced by year end [21]. The arrangement shows AI capital deployment reaching into sectors with existing physical infrastructure that can be repositioned without greenfield construction.