Private equity giant KKR sees artificial intelligence already reshaping the economy in measurable ways. In a recent analysis, the firm highlighted how AI-related capital spending now accounts for roughly 5% of U.S. GDP and grows at a pace matching the late 1990s technology surge. The four largest hyperscalers alone plan more than $350 billion in capital expenditures this year. Add in other tech players, and the total approaches half a trillion dollars. That outpaced consumer spending as a contributor to GDP growth in the first half of 2025.
Such figures come as economists debate whether the surge marks the start of a sustained expansion or risks building on shaky foundations. Data from the Stanford AI Index 2026 shows global corporate AI investment more than doubled in 2025. U.S. consumer surplus from generative AI reached an estimated $172 billion annually by early 2026, up 54% from the prior year. Yet broad productivity improvements remain patchy. Task-level gains look impressive. Macro evidence stays early and mixed.
KKR points to a 5% rise in revenue per employee at large companies since ChatGPT launched. The bump suggests AI has moved beyond experiment and begun to hit profit-and-loss statements. But many executives still report modest returns. A Bain survey cited in recent coverage found 40% of firms seeing cost savings below 10%. One CFO reportedly spent half a billion dollars on a single AI service in a month. Returns vary sharply by sector and implementation.
Billions flow in. Results arrive slowly.
The investment wave recalls earlier infrastructure buildouts. Railroads in the 19th century. Telecommunications in the 1990s. Vanguard analysts compare the current cycle to those periods and assign a 60% chance the U.S. achieves 3% real GDP growth in coming years. For 2026 they forecast a more modest 2.25% expansion, supported by AI outlays and fiscal measures. First-half softness may linger from tariffs and demographics. Productivity gains from workers have yet to appear at scale.
But this spending already props up the numbers. A Harvard economist calculated that investment in information processing equipment and software drove 92% of U.S. GDP growth in the first half of 2025. Without the data-center boom, recent growth figures would look far weaker. The Guardian reported this concentration carries risks. Any slowdown in expenditure could dent both economic and political momentum.
Productivity data tells a nuanced story. U.S. labor productivity grew 2.7% in 2025, nearly double the prior decade’s average. European firms adopting AI saw a 4% labor productivity increase, with each percentage point spent on training adding nearly 6 points to the gain. Software developers report 26% faster output. Customer support teams achieve 14% to 15% improvements. Marketing teams sometimes double output. Gains shrink in roles demanding complex reasoning. Heavy reliance on AI may even create learning penalties that hinder long-term skill building.
The International Monetary Fund raised its 2026 global growth forecast, crediting the AI boom for offsetting trade tensions. In a note published earlier this year, the IMF projected AI could add as much as 0.3 percentage points to global growth this year and between 0.1 and 0.8 points annually in the medium term. Advanced economies stand to capture most benefits because of better infrastructure and readiness. The IMF report also warns of uneven distribution across countries and within labor markets.
So far the boom concentrates among a few players. Investment flows heavily to a small group of countries, companies, and deals. The Stanford report notes generative AI reached 53% adoption in three years, faster than personal computers or the internet. Adoption correlates strongly with GDP per capita. Wealthier nations and workers reap rewards first. Younger employees in AI-exposed occupations face steeper disruption.
Private markets feel the tension. KKR executives told New Private Markets that the wider use of AI complicates exits. It creates a divide between firms vulnerable to disruption and those benefiting from the infrastructure buildout. Some portfolio companies race to integrate AI tools. Others lag and lose ground. The “big AI question” now hangs over valuations and deal flow.
Optimists see a J-curve ahead. Initial costs weigh on results. Then efficiency gains compound. Morgan Stanley analysts project AI will reshape China’s growth trajectory in similar fashion, starting near neutral before turning positive after 2027. In the U.S., Apollo’s Torsten Slok describes three engines for 2026: data-center spending, AI-related energy demand, and eventual productivity lift. The first two already deliver. The third remains prospective.
Critics caution against overhyping. A Fortune article from early June detailed estimates that AI generated about $250 billion in economic activity in 2025, comparable to the entire U.S. airline industry. The authors argue official statistics fail to capture rapid quality improvements and falling costs. If adjusted properly, U.S. growth in 2025 might appear 4 percentage points higher. Yet they present this as an upper bound, not a central case. Measurement problems abound when capabilities advance 2,600% annually while prices plunge 94%.
Federal Reserve officials factored AI investment into their December forecasts, lifting the 2026 U.S. growth projection to 2.3%. Chair Powell acknowledged the role of data-center and business spending. Still, the central bank watches for signs that capital spending simply substitutes for other demand without raising underlying potential growth. Inflation risks persist if the boom overheats energy and construction markets.
Energy emerges as a constraint. Data centers and AI training require massive power. Goldman Sachs projects AI-related spending climbing from $765 billion this year toward $1.6 trillion by 2031. Delays in power plants or grid upgrades could slow deployment. Some regions already face tight electricity supplies. The boom simultaneously drives investment in new generation capacity, creating another growth tailwind but also potential bottlenecks.
McKinsey estimates companies will pour nearly $7 trillion into global data-center infrastructure by 2030. That sum equals the combined GDP of Japan and Germany. The scale underscores both opportunity and exposure. Hyperscalers bet that demand for AI computation will keep rising exponentially. If enterprise adoption accelerates and use cases deliver consistent value, the bet pays handsomely. If not, excess capacity could weigh on returns for years.
Recent IPO filings add urgency. Anthropic’s S-1 and expected listings from OpenAI and others arrive amid questions about unit economics. Some customers report runaway costs. Others point to clear savings in narrow applications. The market sorts winners from experiments in real time. Revenue-per-employee gains tracked by KKR offer one positive signal. Sustained profit growth at the frontier labs would reinforce it.
But this is no guaranteed boom. History shows general-purpose technologies take time to diffuse. Electricity and computers both required complementary investments in skills, organization, and infrastructure before productivity surged. AI may follow the same path. Or new capabilities in autonomous research and recursive improvement could compress the timeline dramatically. Economists at the Peterson Institute and others model scenarios ranging from gradual adoption to transformative acceleration by 2030.
Either way, the current spending already moves the needle. U.S. growth holds steady despite tariffs, demographic drags, and geopolitical shocks. AI capital expenditure acts as a buffer. Whether it evolves into broad-based prosperity depends on translation from data centers into office productivity, factory efficiency, and service innovation. Early task-level wins suggest the potential exists. The next 12 to 24 months will reveal how quickly organizations close the gap between pilot and profit.
Investors, policymakers, and executives watch closely. Vanguard warns that AI investment’s outsized role represents the chief risk for 2026 forecasts. Too much optimism could lead to policy mistakes or market corrections. Too little risks underestimating a genuine shift in potential growth. For now the data supports cautious optimism. Spending surges. Adoption spreads. Productivity flickers but has not yet ignited across the economy.
The race continues. Nations and firms that convert AI infrastructure into measurable output fastest will claim the largest share of gains. Those that treat the technology as a simple cost center may fall behind. KKR’s analysis captures the moment well. The infrastructure bet is placed. The productivity harvest remains to come.
WebProNews is an iEntry Publication