It has been a tough few weeks for the AI theme, even as the technology companies themselves highlighted the insatiable customer demand for AI during the third quarter earnings season. After exceptional returns during the second quarter, especially for semiconductor and hardware companies, since late June the ride has gotten bumpier for the so-called “picks and shovels” of the AI buildout. Both July as well as October through mid-November have seen “AI tantrums”, with global semiconductors and the momentum factor going through corrections of 30% and 20%, respectively, during each episode1. If the AI thesis is intact and technology earnings have been solid, what’s going on?
While the AI story is very much alive and well, investors are moving through different chapters of the story increasingly quickly. After having rewarded the AI capex builders over the AI capex spenders during the first half of the year (proxied by semiconductors and hardware over hyperscaler stocks), investors are looking beyond the AI buildout phase itself to the AI adoption and monetization phase:
- One more big year of capex: To be clear, the AI buildout continues into 2027, with current consensus estimates looking for $1.2 trillion of capex from the five major hyperscalers after this year’s $800bn investment tab, a figure which is more likely to surprise to the upside than the downside. This would represent another big year of accelerating capex growth, building on this year’s 93% year-over-year growth rate.
- The second derivative problem: After that though, the pace of capex is likely to slow meaningfully. Hyperscalers and neoclouds increasingly need to finance this buildout in capital markets, especially the debt market, and the cost of capital has been going up. It’s not only investors asking these companies to slow down, voters are too as they made clear during the recent Midterm Election.
- What’s it all for: Investors are quickly moving on to the actual enterprise adoption of AI and its monetization. This includes industries within the tech sector itself, with a focus back on hyperscalers and software. That said, it’s not a one size fits all. Within hyperscalers, investors are rewarding companies whose free cash flow trend is bottoming, as monetization from new and existing businesses picks up. Within software, the peak “SaaSpocalypse” panic has rightfully passed, but not all software companies are equally competitive and able to capitalize on aiding companies in their AI adoption. Crucially, the winners also include non-tech companies as AI adoption leads to tangible productivity benefits and new goods and services. It’s not too early to start thinking about the AI adoption winners within sectors like financials, health care and traditional industrials.
- Quality over momentum: Importantly, investors are rewarding the quality factor over the momentum factor once again. After an extreme gap of 32%pts between momentum and quality returns in late June, the gap has now fully closed (and recently turned negative). Momentum now sits 25% below its late June peak, while quality continues to make steady gains. As investors turn more rational rather than euphoric over the AI theme, quality and active management should reassert themselves once again. Lastly, investors should analyze their portfolios holistically to ensure there’s enough diversification from the AI theme as this more volatile period continues.