BofA: AI Confidence Now the Real Risk to US Stocks

BofA Global Research strategists said the biggest risk to US equities is not rising bond yields but a possible loss of investor confidence in artificial intelligence, a dynamic they call the "AI put". The team's report, published via Dow Jones and Market Watch, notes the 20 best-performing S&P 500 stocks added about $1.7 trillion in market value since August 31, while the other 480 lost about $1.9 trillion. The S&P 500 was on course to finish September roughly flat.

Context — why the AI put matters more than yields now

The term borrows from the Fed put, the market belief that central banks will step in to support prices when markets struggle. BofA's argument is that, as the Fed embarks on another round of rate hikes, enthusiasm for AI is performing a similar cushioning role for stocks. The bull market reaches its fourth anniversary in October, and the strategists see its fate as resting largely on whether that enthusiasm holds.

Recent market action fits that picture. As bond yields climbed to their highest levels in decades, a previously broad rally narrowed sharply. Small and midcap stocks, which tend to be more sensitive to interest rates, turned lower, while financials and utilities also came under pressure.

The Dow Jones Industrial Average, which has less exposure to AI than the S&P 500 or the Nasdaq composite, has struggled too. Gains in AI-linked shares masked much of that weakness. The strategists noted a key difference from the original Fed put: that rests on the decisions of one institution, whereas the AI put depends on the confidence of millions of investors, which is harder to gauge.

A central question is where the return on all the AI spending will come from. Analysts at Goldman Sachs and elsewhere estimate that more than $1 trillion has gone into data centre build-outs since late 2022.

Data — what the numbers show

The concentration data is the core of BofA's case. The 20 best-performing S&P 500 stocks, mostly technology and industrial names, have added about $1.7 trillion in market value since August 31. The other 480 have lost about $1.9 trillion over the same stretch. The S&P 500 was on course to finish September roughly flat.

Group	Market value change since Aug 31
20 best-performing S&P 500 stocks	+$1.7 trillion
Other 480 S&P 500 stocks	-$1.9 trillion

That gap is the "AI put" in numbers. The index looks resilient at the headline level while the median constituent is under pressure. Small and midcap stocks, financials and utilities have come under pressure as yields rose to multi-decade highs. The Dow Jones Industrial Average, with less AI exposure than the S&P 500 or the Nasdaq composite, has struggled too.

On the spending side, Goldman Sachs and other analysts estimate more than $1 trillion has gone into data centre build-outs since late 2022. One economist pointed out that analysts covering technology expect cash flows to surge by 2028, while those covering the sectors that would pay for AI services are far more cautious, suggesting forecasts across sectors are out of step.

Analysis — what it means for markets and sectors

If the AI put is doing the work of the Fed put, then the swing factor for US equities is not the direction of yields but the durability of AI enthusiasm. BofA says higher Treasury yields would eventually weigh on stocks, but probably at a higher level than most expect. That reframes the yield threshold as a secondary concern until confidence itself cracks.

The concentration cuts both ways. Small and midcap stocks, financials and utilities have already come under pressure as yields rose to multi-decade highs, so the weakness underneath the index is visible in those groups. The Dow Jones Industrial Average has struggled too, given its lighter AI exposure than the S&P 500 or the Nasdaq composite.

The second-order risk sits in the spending chain. If the analysts covering technology are right that cash flows surge by 2028 while the sectors paying for AI services remain cautious, the forecast mismatch could surface in capex updates or earnings. A wobble there would hit the same 20 names carrying the index.

BofA acknowledged the limitation in its own framing: the AI put depends on the confidence of millions of investors, which is harder to gauge than a single institution's decisions. Investor psychology can shift quickly, and this is one team's framing rather than a consensus view. Positioning follows the same logic — fear of missing out on AI is driving aggressive dip-buying and crowding out attention to macro risks.

Outlook — what to watch next

The catalysts to watch are AI-related earnings and capex updates, plus any sign of slower spending. On BofA's reading, those could become bigger swing factors for US equities than moves in yields, at least until yields climb further. The strategists see a level at which higher Treasury yields would start to weigh on stocks, but say it is probably higher than most investors expect.

Traders may want to watch the gap between AI leaders and the broader market, including small caps, as a gauge of how much the AI put is doing the work. A widening gap signals the cushion is still holding; a narrowing one signals it is not. The bull market's fourth anniversary in October is a natural checkpoint for whether the narrow leadership persists. Fed commentary on the rate path remains a background factor, since the strategists tie the AI put's importance to the Fed embarking on another round of rate hikes.

Frequently Asked Questions

What does the "AI put" mean for retail investors?

The AI put is BofA's term for the idea that enthusiasm for AI is cushioning US stocks the way the Fed put once did. For retail investors, the practical read is that headline index levels may overstate broad market health, because the 20 best-performing S&P 500 stocks added about $1.7 trillion since August 31 while the other 480 lost about $1.9 trillion. The strategists say the AI put depends on the confidence of millions of investors, which is harder to gauge than a central bank's decisions.

Why did the S&P 500 finish September roughly flat if most stocks fell?

The index was on course to finish September roughly flat because gains in AI-linked shares masked weakness elsewhere. The 20 best-performing S&P 500 stocks, mostly technology and industrial names, added about $1.7 trillion in market value since August 31, while the other 480 lost about $1.9 trillion. Small and midcap stocks, financials and utilities came under pressure as yields rose to multi-decade highs, and the Dow Jones Industrial Average struggled too given its lighter AI exposure.

What could make the AI put stop working?

A loss of investor confidence in AI is the scenario BofA flags. The strategists say a central question is where the return on AI spending will come from, noting that analysts covering technology expect cash flows to surge by 2028 while those covering the sectors that would pay for AI services are far more cautious. If confidence faltered, it could collide with an uncertain macro backdrop and magnify losses across markets.

Bottom Line

BofA says the AI put, not the Fed put, now cushions US stocks — and its durability is the risk to watch.

Disclaimer: This article is for informational purposes only and does not constitute investment advice. CFD trading carries high risk of capital loss.