A healthy correction underway

Although by the close Thursday the S&P 500 was down a mere 3% from its highs, let’s face it, things feel a lot worse. That is because the 3% loss is a combination of much bigger losses year-to-date in areas that had attracted the bulk of the assets — think Bitcoin (-49%), software giants (-25%) and the Mag 7 (-6%). Meanwhile, the stocks of the remaining, unloved 493 are up 1% and an even more comforting 6% from the market’s November lows. The much-despised Russell 2000 small-cap index is up 5% year-to-date, despite pulling back a tad through Thursday, and 12% from the November lows. Many clients are asking, is now the time to buy back the AI stocks? Or is this just a precursor to a much bigger and deeper problem? Or should I keep leaning into the “broadening-out trade” and wait for a bigger correction in the tech names? Only time will tell, but at Federated Hermes, we are sticking with option three. We view the current tech correction as healthy but not complete. On the other hand, we expect the market’s shift toward smaller-cap and value stocks is likely to continue for some time.

Here are our reasons:

1. The "hyperscalers" are beginning to tilt away from the "asset light" model that made them so attractive. One of the defining characteristics that propelled mega-cap technology companies over the last decade was their asset-light business model: high incremental margins, modest capital intensity, and extraordinary free cash-flow generation and growth. That picture is changing quickly. The hyperscalers are now committing enormous capital to support the AI infrastructure buildout. In 2025, the Magnificent Seven spent $400 billion on AI, against combined cash flow of $825 — still net positive free cash flow, and "only" funded with $80 billion in new debt. Still, this was the first year that the hyperscalers’ capital expenditures (capex) nearly absorbed all of their free cash flow. So far, including Amazon’s blowout capex announcement on Thursday, capex guidance for 2026 has expanded to more than $600 billion and climbing. As a result, free cash flow is deteriorating across the cohort, and investors are shifting allocation toward emerging areas of AI bottlenecks, such as memory and semiconductor capital equipment companies.
2. The software names are caught between a rock and a very hard place. The software sector faces a powerful and underappreciated combination of headwinds. First, many firms are still digesting the excess seat licenses sold during the pandemic era surge in remote work demand. Renewal cycles remain marked by optimization and downsizing, not expansion. Second, AI is threatening the traditional license-based model itself. The recent Anthropic announcement may prove to be software's "DeepSeek moment," demonstrating that agentic AI tools can automate functions historically performed inside legacy enterprise applications. That raises real questions about long-term pricing power and total addressable market assumptions. Importantly, the largest Software-as-a-Service (SaaS) platforms retain significant advantages: massive installed bases, deeply embedded workflows and high switching costs. If history is any guide, these firms will adapt and integrate AI into their solutions to defend their moats. Still, the transition will be messy and investors are selling first and asking questions later. Value is starting to get created among these stocks, but with so many question marks about their futures, we are remaining cautious in this space pending some sign of stabilization.
3. The market is broadening because fundamentals are improving outside of the heretofore market-leading tech space. The rotation we have experienced over the first month of the year is not just being driven by concerns over the Mag 7. It also reflects genuine improvement across the broader economy. GDP growth has averaged 4% over the past three quarters, and forward growth expectations continue to rise on the back of the implementation of the One Big Beautiful Bill tax cuts, larger-than-expected tax returns, and continued Federal Reserve easing. Meanwhile, early earnings season results have shown broad-based beats, with revenue growing 8.5% and earnings-per-share growing 13% versus a year ago. Importantly, earnings expectations for the full year have risen as the season has progressed — always a good sign that on balance, the incoming earnings releases are raising analysts’ optimism for the full year.
4. Many of the previously unloved value stocks are not only experiencing improving fundamentals but also sport cheap valuations. A number of value-oriented sectors look compelling on both absolute and relative bases. At Federated Hermes, we long ago stopped viewing the "value index" as a single monolith; the widely accepted category encompasses two different asset classes: economically sensitive cyclicals and defense dividend payers. In the current environment, both look attractively valued and are benefiting from the macroeconomic environment. Cyclical parts of the market like banks, energy, industrials and materials are benefiting from above-trend GDP growth, additional stimulus and the AI infrastructure build-out. At the same time, the defensive dividend-paying parts of the market like utilities, consumer staples and pharmaceuticals are being lifted by a pick-up in volatility and lower policy rates, which make their lower beta and higher dividend yields look more attractive. With fundamentals improving and valuations still depressed, we believe these areas have multiple avenues for outperformance. And with investors tiring of all the drama in technology-land, the opportunity to invest in long-ignored but stable, well-diversified businesses with attractive valuations and dividend yields is proving a welcome relief.
5. "Goldilocks" is about to arrive. We appear to be entering an unusual macro regime where easing monetary policy intersects with accelerating earnings — historically a powerful combination for equities. AI-driven productivity gains are creating benign labor market softness, as evidenced by this week’s softer labor readings, reducing wage pressures without signaling economic weakness. With policy rates well above core inflation, monetary conditions are tighter than necessary, giving incoming Fed Chair Kevin Warsh ample room to ease. We think the Fed is likely to cut rates two or three times over the course of 2026. At the same time, earnings estimates for 2026, 2027 and 2028 continue to rise as economic growth accelerates. Put together, 2026 may deliver a rare "Goldilocks" alignment: Fed cuts alongside broadening corporate earnings growth. Historically, this combination has been especially supportive of small caps, cyclicals, emerging markets, and value — precisely where leadership is emerging now.

Given all of the above, the S&P is likely entering a period of choppiness, as investors continue to rotate capital out of the mega-cap names that make up 34% of the index and into the rest of the market, which is still broadly under-owned. We are keeping our portfolios tilted toward the latter and will wait for lower levels — if we get them — to add additional capital to stocks overall, and technology in particular.

A key risk we are monitoring

A risk to our somewhat benign view of the present correction is the credit markets, especially within the tech sector. Bitcoin’s heavy losses, and the carnage among software names where fundamentals are eroding quickly, could lead to a credit event with macro-level implications.

For now, credit markets remain liquid. Despite rising measures of risk, no clear signs have surfaced that a major player is in significant distress. We are closely monitoring credit spreads and the credit default swaps markets for signs of elevated stress or contagion. Currently, the pressure is more concentrated than widespread, but we are watching for signs of credit stress spreading. Given the strong speculative run in 2025, this risk is nonzero. We are staying in close touch with our credit teams and will update our views accordingly.