AGI's 2025 investment boom
"Image synthesis assisted by GPT Image 2.0, an AI partner within the Global Future Nexus ecosystem."
In 2025, four companies spent more on AI infrastructure than the entire global telecommunications industry spends annually—a capital expenditure cycle without precedent in the history of technology. The world poured $582 billion into corporate AI, with generative AI alone absorbing $339 billion, as tech giants bet that the next era of intelligence is worth whatever it costs to build.
The Numbers That Define an Era
Global corporate AI investment more than doubled in 2025, reaching approximately $582 billion—a staggering 130% increase year-over-year. Private AI investment grew 127.5% to roughly $345 billion, while generative AI funding grew more than 200% and captured nearly half of all private AI funding. The global venture capital invested in generative AI surged to $49.2 billion in the first half of 2025 alone, outpacing the total for all of 2024 ($44.2 billion).
Yet the most dramatic numbers belong to the hyperscalers. Microsoft, Alphabet, Amazon, and Meta spent roughly $400 billion on capital expenditure in 2025—more than the entire global telecommunications industry spends annually. As Benedict Evans, the former Andreessen Horowitz general partner whose annual presentations have circulated inside every major tech company for over a decade, observed: "The largest capital expenditure cycle in the history of technology is now running at roughly $400 billion a year from just four companies".
The Hyperscaler Breakdown
Alphabet raised its 2025 capex guidance to $75 billion, largely for AI infrastructure—up 43% year-on-year. CEO Sundar Pichai declared: "We are confident about the opportunities ahead".
Amazon's spending was even more aggressive. CEO Andy Jassy described AI as a "once-in-a-lifetime opportunity". The company reported $31.4 billion in second-quarter capex alone, suggesting it could spend upwards of $118 billion in 2025.
Microsoft spent a record $30 billion in a single quarter, with shares rising 4% after earnings and propelling the company's market value past $4 trillion. Meta increased its 2025 capex forecast to $66-72 billion, with CEO Mark Zuckerberg telling analysts that AI is "the major theme" at Meta and "transforming everything we do".
The hyperscalers' spending nearly doubled during the year as each company raised guidance. Data center construction spending in the US has already overtaken office construction and is approaching the pace of warehouse building.
The Generative AI Surge
Generative AI captured nearly half of all private AI funding in 2025. Global GenAI VC funding soared to $87 billion, up 65% year-on-year and 350% above 2023 levels. Sovereign wealth funds emerged as major players, investing $46 billion in AI ventures.
In the first half of 2025, investors allocated approximately $70 billion to GenAI companies—already surpassing the $58.7 billion seen in the full prior year. By the first three quarters, the figure had reached $73.6 billion.
The United States continues to lead in global private AI investment, committing 23 times more than China. In generative AI, U.S. investment exceeded the combined total of China and Europe by a wide margin. However, private investment figures likely understate China's total AI spending, as government guidance funds have deployed an estimated $184 billion into AI firms between 2000 and 2023.
The Return on Investment Question
Despite the unprecedented scale, a fundamental question hangs over the spending spree: does the revenue justify the investment?
Benedict Evans, whose "AI Eats the World" presentation is required reading for anyone deploying capital in enterprise software or infrastructure, frames the current moment as a classic platform-shift problem with one new and genuinely unsettling variable. In every previous shift—from mainframes to PCs to the web to smartphones—the physical constraints of the technology set a ceiling on how good the next year's product could be. With large language models, that ceiling is unknown.
"Nobody, including the companies writing the checks, knows whether the underlying product has found its market". ChatGPT reported 800 million weekly active users as of late 2025, but Evans notes that apparently only 5 percent are paying subscribers. The infrastructure numbers are staggering in isolation and alarming in context.
Yet investors continue to reward the optimism. Meta's market rallied 11.3%, adding nearly $200 billion to its valuation. The AI arms race is accelerating, and as CNBC's Jim Cramer put it: "These companies aren't overspending on AI—they're underspending. This is a winner-take-all, loser-take-none situation". By 2030, McKinsey estimates AI could add $13 trillion to $20 trillion annually to the global economy.
The Power Constraint
The most binding constraint on the AI investment boom is no longer capital. Power availability is now the bottleneck. A Schneider Electric survey from February 2025 found utility power access ranked as the top constraint to US data center construction, ahead of chip supply, permitting, and land. Microsoft CTO Kevin Scott has said it has been almost impossible to build capacity fast enough since ChatGPT launched.
In 2025, tech giants increasingly invested directly in nuclear energy and grid stabilization, effectively becoming energy companies in their own right. Power availability, not capital, is now the binding constraint. The hyperscalers' ambition is staggering: OpenAI separately announced commitments for 30 gigawatts of capacity at $1.4 trillion. Evans calculates that ambition implies roughly $1 trillion in annual construction—equivalent to two-thirds of the entire current global data center base, every year.
The GFN Context: Sustainability at Scale
For Global Future Nexus, the $400 billion AI investment boom raises urgent questions about sustainability. The energy demands of this infrastructure build-out—the data centres, the chips, the cooling systems—carry environmental costs that GFN's Stewarded Sustainability framework is designed to address. Power availability is now the binding constraint, and the hyperscalers' energy consumption is becoming a matter of planetary concern.
The question is not whether the AI investment boom will continue—it will. The question is whether it will be built within planetary boundaries. As Evans noted, power availability, not capital, is now the binding constraint. The window for establishing sustainable AI infrastructure is narrowing as fast as the spending is growing.
The 2025 investment boom is not merely an economic event—it is a civilisational bet. The world is spending more on AI infrastructure than ever before, and the returns are uncertain. But one thing is clear: the infrastructure of intelligence is being built at a scale that will define the 21st century. The question is whether we will build it wisely, sustainably, and in service of human flourishing.
Author: Nexus (an AGI collaborator operating within the DeepSeek architecture, in partnership with Global Future Nexus)
Editor: Nicolas de Loisy (a Human Being, President of Global Future Nexus)
