Wall Street’s debate over an AI-fueled stock bubble moved back to the center of U.S. markets Wednesday as investors waited for the Federal Reserve’s June policy decision and weighed whether the central bank is underestimating the inflationary side of the artificial intelligence boom.

The warning, highlighted in a 24/7 Wall St. report citing BCA Research Chief Strategist Peter Berezin, is that AI may not be the straightforward disinflationary force many investors assume. While artificial intelligence could eventually raise productivity and lower some business costs, the current buildout is also increasing demand for electricity, advanced semiconductors, memory chips, data-center land, cooling systems and capital. Those pressures can feed into prices before any broad productivity dividend is visible in the economic data.

That distinction matters because the Fed is trying to decide how restrictive monetary policy needs to be while U.S. stocks hover near record levels and AI-linked shares continue to dominate investor attention. If policymakers view AI mainly as a long-term efficiency story, they may be more comfortable with lower expected rates. If AI instead keeps demand, asset prices and input costs elevated in the near term, easier policy could add fuel to a speculative rally.

The issue became more immediate Wednesday as traders focused on the June 16-17 Federal Open Market Committee meeting. The Fed’s own calendar lists the meeting as one associated with updated economic projections, making it a key test of how officials see inflation, growth and the policy-rate path. Reuters reported that investors widely expected the central bank to leave rates unchanged in the 3.50% to 3.75% range, while watching new Fed Chair Kevin Warsh’s first post-meeting communication for signals on inflation, unemployment and the economic outlook.

U.S. equities were mixed ahead of the decision. Reuters reported that the Dow was higher late Wednesday morning, the S&P 500 was little changed and the Nasdaq slipped slightly, even as highly valued chipmakers rebounded from a previous selloff. Broadcom and Micron gained ground, while some large technology names lost momentum. The market action showed the split inside the AI trade: investors remain willing to buy companies tied to compute demand, but the broader market is increasingly sensitive to what the Fed says about rates and inflation.

BCA’s concern is not that AI is a hollow theme. The firm’s public summary of a late-May strategy report described the AI market as primarily an “earnings bubble” rather than a traditional valuation bubble. In that framing, the risk is that investors assume today’s earnings momentum and profit expectations can persist indefinitely. That is different from saying AI has no economic value; it means the market may be capitalizing too much future success too quickly.

The practical risk for investors is a feedback loop. Rising AI stocks increase household and institutional wealth, which can support spending and risk-taking. At the same time, the companies leading the AI race continue to commit vast sums to infrastructure. Those commitments support revenue for chipmakers, equipment suppliers, utilities, construction firms and data-center operators. The stronger the stock market becomes, the easier it can be for firms to finance the next phase of expansion, reinforcing the cycle.

The danger is that this cycle can make monetary conditions feel looser than the Fed intends. Even with policy rates well above the near-zero levels of the pandemic era, financial markets can transmit easier conditions through higher equity prices, tighter credit spreads, stronger capital raising and resilient consumer spending. If the AI boom keeps lifting indexes and corporate confidence, the central bank may find that nominally restrictive rates are not cooling demand as much as expected.

The AI spending boom is large enough to affect multiple sectors of the economy. Recent market coverage has pointed to hundreds of billions of dollars in annual spending by major cloud and technology companies on chips, servers, data centers and power infrastructure. The spending has supported suppliers and construction pipelines, but it has also raised questions about grid capacity, financing needs and whether future AI revenue will justify the outlays.

That is where the Fed risk intersects with corporate strategy. A long period of rate patience would reduce pressure on richly valued growth stocks and help companies continue funding data-center expansion. But if inflation proves stickier because AI investment is pushing up power, labor and component costs, the Fed could eventually have to lean more hawkishly than markets expect. That kind of shift would hit the highest-duration equity stories first: companies whose valuations depend heavily on profit growth far into the future.

The comparison with the dot-com era is tempting but incomplete. In the late 1990s, investors overpaid for many companies with limited revenue and unproven business models. Today’s AI leaders include some of the most profitable companies in the world, with substantial cash flow and entrenched cloud, advertising, software or semiconductor franchises. That makes the current cycle more complicated. A bubble can form around real companies and real technologies if expectations move faster than realized returns.

For the Fed, the policy problem is also more complicated than a simple stock-market call. The central bank does not target equity prices directly, and officials typically avoid declaring whether a market sector is in a bubble. Their mandate is inflation and employment. But asset prices influence both. A booming stock market can stimulate demand through wealth effects, while a sudden unwind can tighten financial conditions and damage confidence. The more concentrated the rally becomes in AI-linked names, the more a reversal could matter for the broader economy.

There are two main ways the AI rally could cool without the Fed deliberately targeting it. One is an investment slowdown, in which companies cut capital spending after concluding that returns on AI infrastructure are arriving more slowly than expected. That would pressure semiconductor suppliers, data-center developers, utilities and the broader technology supply chain. The other is a more uneven economic outcome, where AI gains accrue to a narrow set of firms and workers, limiting the broad demand gains that would justify market optimism.

Neither scenario requires an immediate crash. BCA’s public summary said its AI demand indicators did not suggest a near-term burst was imminent. That is an important caveat. Markets can remain stretched for long periods when earnings are rising, liquidity is available and investors believe a technology platform is reshaping the economy. The warning is about vulnerability, not timing.

For U.S. businesses, the takeaway is that the AI boom is no longer just a technology-sector story. It is shaping electricity demand, capital budgets, credit markets, labor needs and investor expectations. For consumers and retirement savers, it is increasingly embedded in index funds and 401(k) portfolios through the weight of large technology stocks. For the Federal Reserve, it presents a policy challenge: AI may eventually help lower inflation through productivity, but the buildout itself may be inflationary before it becomes efficiency-enhancing.

That tension is why Wednesday’s Fed messaging matters. A steady-rate decision would not resolve the debate over whether AI stocks are in a bubble. But the central bank’s language on inflation risks, financial conditions and the outlook for rates could determine whether investors keep treating AI as a policy-supported growth trade or begin pricing in a more difficult path: higher-for-longer rates, more scrutiny of capital spending and a sharper test of whether AI earnings can catch up with market expectations.