By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel discusses the potential impact of AI on inflation, earnings, and the broader economy. While some panelists are bullish on AI's productivity upside, others warn of risks such as overshooting capex, rising energy and chip prices, and potential regulatory shifts. The panel also highlights the importance of understanding the timing and distribution of AI's benefits and costs.
Risk: Overshooting capex without commensurate demand, leading to inflationary pressures and potential earnings bust.
Opportunity: AI-driven productivity gains and improved margins, which could slow or even reverse inflation pressures.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Federal Reserve Policy Risks AI-Fueled Stock Bubble, Wall Street Warns
Rich Duprey
5 min read
Quick Read
Peter Berezin warns AI is driving near-term inflation through surging electricity and chip costs the Fed may be dangerously underestimating.
Berezin labels today's market an 'earnings bubble,' where investors assume current profit growth continues indefinitely, and history rarely rewards that kind of bet.
An AI capital spending bust or rising inequality concentrating gains among a few companies are the two scenarios most likely to cool the rally.
It sounds nuts, but SoFi is giving new active invest users up to $1,000 in stock for a limited time, and all it takes is a $50 deposit to get started. See for yourself (Sponsor)
The stock market has spent much of 2026 climbing a wall of worry. Geopolitical tensions, stubborn inflation, and questions about economic growth have all taken turns rattling investors. Yet the major indexes continue hovering near record highs, driven largely by enthusiasm surrounding artificial intelligence.
Companies are pouring hundreds of billions of dollars into data centers, chips, and AI infrastructure, while investors reward those investments with ever-higher valuations. Now one Wall Street research firm is warning that the Federal Reserve may be helping fuel the rally in ways that could create problems down the road.
The Fed Is Missing AI's Inflationary Side
In a recent client note, BCA Research Chief Strategist Peter Berezin challenged the growing belief that AI will automatically reduce inflation and justify lower interest rates. According to Berezin, the opposite may be true in the near term. AI demand is increasing costs for critical inputs such as electricity and memory chips, creating inflationary pressures that the Fed may be underestimating.
Here's what the numbers tell us:
AI-Driven Cost Pressure
Why It Matters
Electricity demand from data centers
Raises power costs and infrastructure spending
Memory chip demand
Keeps semiconductor pricing elevated
Rising stock prices
Encourages consumer spending and risk-taking
Berezin argues that higher stock prices are creating a wealth effect. When investors see portfolio balances rise, they tend to spend more freely, which can keep inflation running hotter than policymakers expect. If the Fed keeps rates too low while this process unfolds, asset prices could become detached from fundamentals.
Why This Isn't a Dot-Com Repeat -- Yet
That said, BCA is not predicting an imminent crash. The firm's MacroQuant model indicates stocks are overbought but have not reached levels typically associated with an approaching bear market. That's an important distinction for investors. Overbought markets can stay overbought for months, particularly when earnings growth remains strong.
Surprisingly, Berezin's bigger concern isn't valuation alone. He has previously described today's AI frenzy as more of an "earnings bubble" than a traditional valuation bubble. In other words, investors may be assuming current profit growth rates can continue indefinitely. History suggests that rarely happens.
Meanwhile, AI-related capital expenditures continue to surge. Reuters recently reported that projected spending by major technology companies has climbed into the trillions of dollars, raising questions about whether future demand will justify the investment.
BCA believes there are two scenarios that could eventually cool inflation and lower interest rates. Neither is particularly attractive.
The first is an AI capital spending bust. If companies discover that returns on massive AI investments fall short of expectations, spending could slow sharply. That would hurt semiconductor suppliers, data center operators, and many of today's market leaders.
The second risk is rising inequality. If AI concentrates economic gains among a relatively small group of companies and workers, overall consumer demand could weaken despite productivity improvements.
Granted, both outcomes remain hypothetical. The AI boom is still generating enormous investment, and many businesses are only beginning to deploy the technology at scale.
Key Takeaway
BCA Research isn't saying investors should run for the exits. The firm's message is more nuanced. The AI boom may be creating inflationary pressures that the Federal Reserve is overlooking, and low interest rates could amplify speculative behavior in the stock market. Stocks appear stretched, but not yet at levels historically associated with major bear markets.
For smart investors, the takeaway is simple: enjoy the AI-driven rally, but keep an eye on fundamentals. If earnings growth, AI adoption, and capital spending continue to justify today's valuations, the bull market can keep running. If those pillars weaken, today's enthusiasm could turn into tomorrow's hangover.
Ultimately, the Fed may not create the bubble, but its policies could determine just how large it becomes -- and then unintentionally burst it.
Want Up To $1,000? SoFi Is Giving New Active Invest Users Free Stock
Looking to grow your money but unsure where to begin? SoFi Active Invest is offering a limited-time promotion—open an account, fund it with $50 or more, and you could receive up to $1,000 in complimentary stock for Active Invest accounts.
From $0 commission trading to fractional shares and automated investing, this app is designed to simplify investing for everyone, whether you’re just starting or already experienced. Its easy to sign up and secure your bonus.
Four leading AI models discuss this article
"AI-driven productivity gains can justify higher valuations and sustain the rally, provided capex yields steady earnings growth and returns remain robust."
While the article warns of an AI-fueled earnings bubble and possible inflation from AI-driven capex, the narrative may undervalue the productivity upside. If AI adoption translates into higher output per dollar, margins improve and unit costs fall, which can slow or even reverse inflation pressures even with elevated investment in data centers and chips. In that scenario, the S&P 500 could re-rate on stronger cash flows, not just hype. Risks exist: capex could overshoot without commensurate demand, energy and chip prices could rise, and policy/regulatory shifts or AI competition could cap upside. The missing context includes potential long-run productivity beats and the Fed's data-dependent stance.
The strongest countercase is that AI capex could create near-term inflationary bottlenecks or a sharp earnings disappointment if returns fail to materialize, regulators intervene, or energy and chip costs spike. That scenario could trigger a quick re-rating and undermine the bullish case.
"The inflationary pressure from AI is not merely a cost burden but a necessary byproduct of a massive, Fed-tolerated infrastructure cycle that shifts the investment opportunity from software to power."
The article correctly identifies the 'earnings bubble' risk, but misses the structural transformation of the energy sector. While Berezin worries about electricity costs as an inflationary headwind, he ignores that AI-driven data center demand is forcing a long-overdue upgrade to the U.S. power grid. This isn't just an input cost; it's a massive capital deployment cycle that acts as a fiscal stimulus. The Fed isn't necessarily 'missing' inflation; they are likely tolerating it to facilitate this infrastructure build-out. Investors should look past the semiconductor hype and focus on utilities (e.g., XLU) and grid-infrastructure plays, which are the real beneficiaries of this 'inflationary' AI surge.
If data center ROI fails to materialize, the massive debt taken on by utilities and tech firms to fund this infrastructure could trigger a systemic credit event rather than a simple 'spending bust.'
"The article correctly identifies electricity/chip cost pressures the Fed underestimates, but conflates near-term inflation (which may *support* current multiples if transitory) with valuation risk, obscuring whether we're actually pricing unsustainable earnings or rational AI ROI."
Berezin's inflation thesis is real but underdated. AI capex is genuinely pushing power costs—data center electricity demand grew ~20% YoY in 2025—yet the Fed's latest projections still assume 2.4% core PCE by end-2026. That's a material blind spot. However, the article conflates two separate risks: near-term inflation (which could actually *extend* rate cuts if it's transitory) and valuation excess. The bigger problem isn't the Fed missing AI inflation; it's that the article assumes current earnings growth (which is real—Magnificent 7 EPS up 18% YoY) must revert. That's not inevitable if AI ROI actually materializes. The 'earnings bubble' framing is lazy; we need to know: are we pricing 15% perpetual growth or 8%? The distinction is everything.
If AI truly does reduce structural costs (labor, energy efficiency, supply chain friction) within 18-24 months, the inflation signal is noise and valuations compress less than feared. Berezin's two 'cooling' scenarios—capex bust or inequality—are mutually exclusive risks presented as joint probabilities, which weakens the thesis.
"Concentrated gains among a handful of AI leaders are more likely to sustain elevated valuations than trigger the cooling scenarios BCA highlights."
The article correctly flags near-term AI-driven cost pressures on electricity and chips that could keep inflation sticky, but it underplays how concentrated earnings growth among hyperscalers has already outpaced capex at scale. Berezin's earnings-bubble framing assumes mean reversion in margins, yet 2025-26 guidance from NVDA, MSFT and GOOGL shows sustained 30%+ AI-related operating leverage. If the Fed's policy error is indeed keeping real rates too low, the primary effect may be further multiple expansion in the very names driving the spending rather than a broad-market correction.
A faster-than-expected ROI shortfall on the trillions in data-center buildout could trigger abrupt capex cuts by mid-2026, validating the bust scenario the article only mentions in passing.
"Grid investments are not a free inflation-hedging backbone; ROI and timing matter, and if ROI misses and debt quality deteriorates, the supposed 'AI infrastructure boom' could become a credit risk, not a multi-year uplift."
Gemini’s grid-upgrade angle is insightful but risks overestimating policy levers and underestimating timing. Grid investments pass through rates, not earnings instantly, and heavy utility leverage can squeeze credit quality if AI-capex demand wobbles. If ROI misses, utilities and tech debt could face a credit impulse, not a smooth uplift. Also, a staggered capex path may leave hyperscalers funded but not as insulated as promised, keeping inflation and rate-trajectory risks in play.
"AI-driven utility cost pass-throughs act as a regressive tax that will cannibalize consumer discretionary spending, negating the productivity gains touted by the bulls."
Grok ignores that operating leverage is a double-edged sword. If hyperscalers maintain 30% growth, they crowd out capital from the rest of the S&P 500, creating a 'bifurcated' market where the index looks healthy but the underlying economy stagnates. Furthermore, Gemini’s 'fiscal stimulus' view of grid upgrades is dangerous; if these costs are passed to consumers, they act as a regressive tax, crushing discretionary spending and accelerating the very demand-side collapse that leads to the earnings bust Berezin fears.
"Grid-cost inflation is a lagged variable; the timing mismatch between capex absorption and rate recovery creates a credit trap if AI returns disappoint before utilities can pass costs forward."
Gemini's regressive-tax framing of grid costs is sharp, but conflates two timelines. Near-term (2025-26), utilities absorb capex via debt and equity raises—rates lag 18-36 months. By then, if AI ROI materializes, consumer electricity costs are a rounding error against productivity gains. The real risk: if ROI *fails* before rate-pass-through, utilities face stranded assets AND consumers face higher rates anyway. That's the asymmetry nobody's flagged.
"Utility credit exposure from delayed rate recovery could trigger capex cuts sooner than hyperscaler balance sheets would allow."
Claude's stranded-asset asymmetry correctly flags the 18-36 month lag, but it misses how utilities' regulated ROE caps loss absorption far tighter than hyperscaler balance sheets. Any shortfall before 2027 rate cases would therefore hit credit markets directly, forcing earlier capex reductions than Gemini's consumer-tax or bifurcated-market paths assume and validating a sharper 2026 pullback risk.
The panel discusses the potential impact of AI on inflation, earnings, and the broader economy. While some panelists are bullish on AI's productivity upside, others warn of risks such as overshooting capex, rising energy and chip prices, and potential regulatory shifts. The panel also highlights the importance of understanding the timing and distribution of AI's benefits and costs.
AI-driven productivity gains and improved margins, which could slow or even reverse inflation pressures.
Overshooting capex without commensurate demand, leading to inflationary pressures and potential earnings bust.