Everything in Markets Is Now Moving Incredibly Fast

Markets are moving at extraordinary speed, but the S&P 500’s proximity to a record masks a sharp split beneath the index. Luke Kawa said it was up about 12% for the year and roughly 1.5% below its high, while only 51.2% of its stocks were above their 200-day averages. He compared that weak breadth with the day after the dot-com peak, but noted that similarly weak breadth in 1998 preceded another two years of gains. Breadth can flag risk, but it is neither necessary nor sufficient to predict a lasting market turn.

AI spending is a central force behind that divergence. A Goldman Sachs note attributed about half of expected earnings growth this year to hyperscaler capital spending, while another Goldman analysis found that three-years-forward sales estimates explained Russell 1000 valuations more than usual; one-year-forward estimates and revisions mattered less. Investors have rewarded expensive companies with high sales expectations, including Bloom Energy and optical stocks, as forecasts for hyperscaler capex have repeatedly risen. The bet is that AI growth continues until credit, lending capacity, regulation, or another constraint intervenes. But the spending must eventually translate into productivity, labor savings, and income for customers. Hannah Lustig estimated that a handful of companies would need revenue equivalent to about 9% of GDP to justify the investment, a target one speaker called ambitious.

The rates backdrop makes those bets harder to assess. Strong global growth and industrial activity are supporting high bond yields, while growth can also lift inflation and prompt central-bank action. Rising real yields make bonds more competitive with stocks, but investors are weighing that return against AI opportunities and appear reluctant to hold longer-term bonds, potentially favoring cash, short-term assets, stocks, or commodities. One speaker said U.S. inflation breakevens had moved little despite past misses of the Fed’s inflation target, perhaps reflecting confidence that central banks will contain inflation—a credibility that still has to be demonstrated. Higher rates are weighing on the equal-weight S&P 500 and Russell 2000 even as AI beneficiaries lead. Many consumer stocks have also struggled: Nike fell from about $180 in 2021 to about $35 at the time of the discussion.

The speed is visible in violent reversals across semiconductors and software, as well as unusually large moves in silver. Retail buying has also swung quickly: Robinhood’s rolling 21-day peak in net single-stock purchases came in late June around Micron’s earnings, and September buying rebounded to about 89% of that peak. Recent interest included Micron and SpaceX, while some unfamiliar names appeared among the top 25. Investors were even using options expiring in four days or less to express views on long-term corporate earnings. Tighter stops at pod shops, more retail participation, zero-day options, and algorithmic trading may contribute to sharp single-stock swings and faster mean reversals, even while index volatility stays lower. Agents could alter order flow, stops, and market visibility, but the speaker experimenting with a paper-trading bot stressed that the effects remain unclear.

Credit markets may offer an important warning if AI financing becomes stretched. As hyperscalers’ free cash flow turned negative, the apparent alternatives were slower growth or more creative financing; financing did become more creative. Debt-funded capex alongside a large fiscal deficit was described as a dependable source of profits “until it isn’t.” Oracle stood out as relatively indebted among major hyperscalers, its credit-default swaps were reportedly spiking, and Nvidia had been trying to “credit wrap” AI borrowers. Meanwhile, AI companies are no longer treated uniformly as unquestionable beneficiaries, and the market is not rewarding the fastest earnings revisions across the board. The discussion pointed toward a possible shift from hypergrowth expectations to growth at a reasonable price. Earnings remain central to stock performance, but revisions lag: in most of the past six or seven bear markets, waiting until S&P 500 earnings revisions fell 5% below their 52-week high still left healthy gains over the following two years; COVID was the major exception.

The cycle also carries inflationary and demand risks. AI investment, rising government entitlements, wars, rearmament, and reshoring semiconductor capacity were all identified as inflationary forces; markets often focus on oil when assessing war headlines and give less attention to defense spending. U.S. consumer discretionary stocks have been a punching bag despite roughly 6% nominal growth, and the sector was estimated to contribute 2.8 percentage points to growth in the quarter discussed. The speakers said AI capex and stock-market wealth effects may now be taking on roles once associated with housing, whose weakness seems to matter less than it used to. They cautioned that the AI cycle cannot last indefinitely without downstream income and demand. Investment and stock gains can reinforce spending in a circular loop, but the eventual outcome could range from broad gains in real wealth and lower costs to major labor-market disruption or worse.