BAKU, Azerbaijan, July 4. Artificial intelligence has become the defining investment story of the global economy. Technology companies are spending unprecedented sums on chips, data centers and cloud infrastructure, governments are rolling out national AI strategies, and investors continue to pour money into companies linked to the sector. But while much of the discussion focuses on AI's technological breakthroughs, another question is becoming increasingly important: Who is actually profiting from the AI boom? The answer is not as straightforward as it may seem.
Although companies developing large language models remain the public face of artificial intelligence, many analysts argue that the biggest financial winners so far are the businesses supplying the infrastructure behind AI, from semiconductor manufacturers and cloud providers to data-center operators and electricity producers. The scale of investment continues to grow.
According to an analysis by Bridgewater Associates, Alphabet, Amazon, Meta and Microsoft are expected to invest about $650 billion in AI-related infrastructure in 2026, up sharply from around $410 billion a year earlier. Bridgewater says demand for computing power continues to outpace supply, forcing technology companies to accelerate investment even further. For AI developers, this creates a difficult balancing act.
Training increasingly sophisticated AI models requires enormous computing resources, while serving millions of users every day generates additional operating costs. Competition has also intensified as companies race to release more capable models while keeping subscription prices attractive. As a result, many AI companies remain focused on expanding market share rather than maximizing short-term profits.
If AI developers are spending hundreds of billions of dollars, someone is earning that money. That has increasingly shifted investors' attention toward what many describe as the "picks and shovels" of the AI economy—the companies providing the hardware and infrastructure that every AI model depends on.
Chipmakers such as NVIDIA and AMD have benefited from soaring demand for AI processors, while Taiwan Semiconductor Manufacturing Co. (TSMC) has become a critical manufacturer for many of the industry's most advanced chips. Cloud providers including Amazon Web Services, Microsoft Azure and Google Cloud have also emerged as major beneficiaries as businesses increasingly rent computing power instead of investing in their own infrastructure.
Unlike AI developers, whose commercial success depends on attracting and retaining users, infrastructure providers generate revenue regardless of which AI model eventually dominates the market. Another sector experiencing rapid growth is data-center construction. As AI models require far more computing power than traditional software applications, technology companies have accelerated investments in new facilities capable of supporting advanced workloads.
Goldman Sachs estimates that between 2026 and 2031, roughly $7.6 trillion could be invested globally in AI infrastructure, including computing capacity, data centers and energy systems, reflecting what the bank sees as a shift in AI investment toward the broader physical economy.
For investors, data centers have increasingly become long-term infrastructure assets rather than simply real estate projects. Perhaps the least visible beneficiaries of AI are electricity producers. Modern AI data centers consume enormous amounts of power, forcing utilities to expand generating capacity and modernize transmission networks.
The impact is already becoming visible in financial markets. In the United States, mergers and acquisitions involving electricity producers and energy infrastructure have accelerated as utilities position themselves for rising demand from AI-related projects. Industry analysts expect electricity consumption from AI data centers to continue increasing over the coming decade. This has renewed interest in nuclear power, natural gas, renewable energy projects and battery storage, as technology companies seek reliable long-term energy supplies.
One of the biggest questions surrounding AI is why company valuations have risen much faster than measured productivity. Economists note that this pattern is not unusual during periods of technological transformation. Electricity, personal computers and the internet all required years—sometimes decades—before their full impact appeared in national productivity statistics. Businesses first had to redesign operations, retrain employees and integrate new technologies into existing processes.
Financial markets, however, price expectations rather than current economic output. Investors are betting that AI will eventually transform industries ranging from healthcare and manufacturing to banking and logistics, even if those gains are not yet fully reflected in official economic data. That debate has intensified in recent months. In its latest annual report, the Bank for International Settlements (BIS) warned that more than $1 trillion in AI-related capital expenditure planned for 2025-2026 could create financial vulnerabilities if future returns fail to justify today's level of investment. The BIS drew parallels with previous technology investment booms, while stopping short of calling AI a speculative bubble.
Not everyone agrees. JPMorgan argues that the current investment cycle remains fundamentally stronger than previous technology bubbles because the companies leading AI spending are already profitable and continue generating substantial cash flows. According to the bank, the challenge is less about whether AI will create value than whether adoption can keep pace with the trillions of dollars now flowing into infrastructure. The economics of artificial intelligence increasingly suggest that today's AI boom extends well beyond software. While AI developers remain at the center of public attention, much of the industry's financial value is currently flowing toward companies manufacturing chips, operating cloud platforms, building data centers and supplying electricity. That does not mean AI developers will not ultimately become the biggest winners. But at this stage of the investment cycle, infrastructure companies appear to be capturing a significant share of the returns. Whether today's extraordinary spending ultimately produces the productivity gains investors expect remains one of the defining economic questions of the decade. For now, the AI economy is proving to be as much about power grids, semiconductor factories and data centers as it is about algorithms.
