Wall Street is reshaping the investment narrative around artificial intelligence (AI).

Top financial institutions, including BlackRock and Bank of America, argue that debating whether AI is a "bubble" is now irrelevant. The real focus should shift to addressing the mismatch and lag between massive upfront investments and commercial returns, a phase that could introduce new market volatility.

Jean Boivin, head of the BlackRock Investment Institute, recently stated that evaluating AI investments through a "bubble" framework is unhelpful at this stage. He emphasized that given the unprecedented scale and speed of AI infrastructure development, any retrospective assessment based on historical metrics would be "incomplete." BlackRock noted that the sheer scale of capital expenditures in AI constitutes a macroeconomic event in itself.

Bank of America issued a more specific warning. Savita Subramanian, its head of U.S. equity and quantitative strategy, also dismissed the "bubble" theory but cautioned about a potential "air pocket" in the market. She described this as a phase where capital expenditure growth outpaces revenue growth, with the lag between investment and returns—especially amid power and infrastructure bottlenecks—potentially shaking investor confidence in the short term.

Despite near-term risks, both institutions remain optimistic long-term. They argue the current AI boom is driven by genuine corporate earnings, investment, and productivity growth, fundamentally differing from the early-2000s dot-com bubble.

BlackRock: "AI Bubble" Narrative Misleading, Real Challenge Lies in Spending-Revenue Gap Wall Street giants are steering market attention away from the "AI bubble" debate. Jean Boivin emphasized that labeling the current AI surge as a bubble oversimplifies the situation. He noted healthy skepticism in markets, with widespread awareness of risks—a stark contrast to the irrational exuberance of past bubbles.

BlackRock's outlook report described AI-related capital expenditure ambitions as so vast that "micro becomes macro." The firm believes this scale of investment could push U.S. GDP growth sustainably above the 2% trend dominating recent decades, positioning AI as both a technological revolution and a macroeconomic force reshaping entire economies.

Behind AI’s grand narrative lie staggering capital expenditure plans and real-world constraints. BlackRock estimates global corporate AI spending could reach $5–8 trillion by 2030, mostly in the U.S. The report questioned whether these expenditures would align with potential AI revenues, highlighting structural pressures from physical limits—from computing power to grid capacity—with AI data centers projected to consume 15–20% of U.S. electricity by 2030.

Bank of America: Beware the "Air Pocket" — AI Investment-Return Lag Risks Bank of America offered sharper warnings about the next phase of AI investment cycles. Savita Subramanian framed potential disruptions as a "pause" rather than a collapse, introducing the "air pocket" concept to describe temporary disconnects between spending and returns.

Data shows this risk materializing on balance sheets. Hyperscalers’ capital expenditures now consume 60% of operating cash flows—double the 30% level a decade ago—though still below the 140% dot-com bubble peak. The bank projects hyperscaler spending (Microsoft, Amazon, Google, Meta, Oracle) will hit $400 billion in 2025 and $510 billion in 2026.

Why This Isn’t 2000 Redux While narrowed market breadth and high valuations echo 2000, Subramanian highlighted key differences: lower equity allocations, earnings-backed valuations, smaller IPO markets, and less extreme speculation on profitless companies versus the late 1990s. These underpin Bank of America’s long-term bullish outlook, including a S&P 500 (^GSPC) target of 7,100 by end-2026, affirming AI’s structurally sound upward trajectory despite near-term turbulence.