Artificial intelligence companies are racing to build vast data centers, but the revenue generated is not keeping pace. That is the warning from Bain & Co., which says the revenue hole may be far wider than many assumed earlier.
In its annual Global Technology Report, Bain estimates that by 2030, AI providers will need about $2 trillion in combined yearly revenue to support the computing power required to meet demand. The firm says actual revenue is expected to decline far short of that target by roughly $800 billion as efforts to generate revenue trail the pace of investment in data-center capacity and related infrastructure.
The forecast raises fresh doubts about valuations and business models across the sector. Interest in products like Google’s Gemini and OpenAI’s ChatGPT, along with AI pushes by companies worldwide, is sending the need for computing capacity and electricity sharply higher. However, the savings and new income that AI can deliver are not growing as quickly as the costs, Bain says.
“If the current scaling laws hold, AI will increasingly strain supply chains globally,” said the chairman of Bain’s global technology practice, David Crawford.
AI spending soars as OpenAI prioritizes growth over profit
OpenAI is incurring multi-billion-dollar losses each year with a focus on growth rather than profit for now, while expecting to become cash-flow positive by 2029. Bain did not assess what might happen to major AI players if profitability remains elusive as 2030 approaches. A day earlier, Nvidia and OpenAI announced a partnership to build massive data centers, as reported by Cryptopolitan.
Spending plans continue to accelerate. Amazon, Microsoft, and Meta are set to push their combined annual AI outlays to more than $500 billion by the early 2030s, according to Bloomberg Intelligence. A wave of new models from OpenAI and China’s DeepSeek, among others, is fueling demand for AI services and prompting the entire industry to invest more.
According to Bain, the incremental global AI computing needs could jump to 200 gigawatts by 2030, with the United States accounting for roughly half of that total. While breakthroughs in hardware and algorithms could ease the load, supply chain bottlenecks or limited power availability could still slow progress, the firm says.
Alongside spending on compute, leading AI companies are pouring money into product development. One focal point is autonomous AI agents that can carry out multi-step tasks with limited guidance, in ways that mimic parts of human workflows.
Over the next three to five years, Bain estimates companies will dedicate as much as 10% of overall tech budgets to building core AI capabilities, including agent platforms.
Bain predicts quantum growth and early robot trials
Bain anticipates growth in quantum computing, an emerging field that it says could unlock about $250 billion in market value across finance, pharmaceuticals, logistics, and materials science. Rather than a single dramatic breakthrough, the firm expects a gradual adoption curve, with early use in narrow domains over the next decade, followed by wider uptake.
Humanoid robots are drawing capital and appearing more often in pilots, yet real-world deployment remains early and depends heavily on human oversight, Bain says. Commercial success will hinge on whether the surrounding ecosystem is ready hardware suppliers, software platforms, and customer operations, and companies that run pilots sooner are likely to set the pace for the field.
Taken together, Bain’s findings describe a fast-rising need for computing power and energy, paired with revenue that may not keep up. The picture is one of rapid build-outs, monetization, and new technologies arriving in steps, not all at once, with early movers positioned to set direction next.
If you're reading this, you’re already ahead. Stay there with our newsletter.
