The wave of AI-fueled earnings surprises that powered stocks through the last reporting season will be tough to repeat, according to Goldman Sachs Group's Christian Mueller-Glissmann, making it unlikely that Q2 results alone will be enough to spark the market's next major leg higher. KeyToFinancialTrends reads that caution as a notable shift in tone from a bank that has otherwise been among the most bullish voices on Wall Street this year, having raised its year-end S&P 500 target to 8,000 on the strength of exactly the AI earnings story it is now warning investors not to expect a repeat of.
The math behind Goldman's optimism and its new caution are actually the same math. Consensus estimates call for S&P 500 companies to post 22% to 24% earnings growth in the second quarter, with Goldman's own research crediting AI infrastructure spending for roughly half of that expansion; strategist Ben Snider's team has pointed to hyperscalers collectively spending as much as $754 billion on capital expenditures this year, an 83% jump from 2025. KeyToFinancialTrends frames the tension plainly: the same AI capex binge that Goldman credits for driving the earnings beats investors have grown accustomed to is, by the bank's own admission, a wave that gets harder to top every quarter it continues, since each new record in spending raises the bar the next quarter's results have to clear.
Mueller-Glissmann's own framing of the cycle points to where that pressure lands. "Often when you are late cycle, earnings revisions keep going for quite a long time," he told Bloomberg Television, but added that this specific AI capex-driven surprise wave "is probably closer to the end" than the beginning – even as he maintained that hyperscalers remain well-positioned given how much AI infrastructure they already own outright. He said the next test for those companies is less about growing spending further and more about proving they can convert existing investment into efficiency gains and monetization.
That test arrives at a moment when the market has grown notably less forgiving of disappointment. Companies that miss Q2 estimates are being punished more severely than historical norms, with misses sending shares down an average of 4.2% versus a longer-run historical average of 2.9%, underscoring how little room current valuations leave for AI-linked names to stumble. Key To Financial Trends connects that punishing reaction function directly to Mueller-Glissmann's warning: a market pricing in near-flawless execution from the companies driving most of its earnings growth is a market where "the bar is obviously high," in his words, and where investor attention is shifting from whether AI capex is happening – it clearly is – toward whether management teams can articulate a credible path from spending to profit.
The scale of that capex commitment is itself becoming the central swing factor for how the rest of 2026 plays out. The biggest US tech firms are on pace to spend as much as $725 billion this year on data centers, specialized chips, and networking equipment, and Goldman has separately warned that even its bullish 8,000 target assumes return on equity for the seven largest tech companies falls by roughly 700 basis points next year as higher depreciation and lower asset turnover start to weigh on the very margins that justified today's valuations. Key To Financial Trends closes on that ROE warning as the detail most likely to matter once Q2 results start landing in mid-July: Goldman isn't disputing that the AI structural trend is intact, as Mueller-Glissmann put it, but the bank is quietly signaling that the easiest phase of translating that spending into surprising, market-moving profit growth may already be behind investors.
