The tech rally that powered markets through 2025 is being tested in 2026.
In early February, a broad tech selloff hit markets, fueled by various elements, including aggressive artificial intelligence (AI) capital spending guidance from hyperscalers, as well as the rapid release of new AI models, which sparked disruption concerns within the software sector. This powerful combination forced investors to separate durable AI leaders from stocks whose gains were driven mainly by sentiment and stretched valuations.
Technology benchmarks saw significant losses. From December 31, 2025, to its February 5 year‑to‑date low, the S&P Technology Index (INDEXSP:SP500-45) dropped by nearly 7 percent. Software-focused measures were hit especially hard; the iShares Expanded Tech-Software Sector ETF (BATS:IGV) declined by almost 25 percent.
Meanwhile, semiconductor‑focused peers like the iShares Semiconductor ETF (NASDAQ:SOXX) remained up more than 5 percent over the same stretch. The divergence underscored how quickly a broad AI theme can split into clear winners and laggards depending on where revenues and profits are actually showing up.
Indexes have since returned some of their losses, but investors with a multi‑year horizon need portfolio construction that can withstand the volatile nature of a sentiment-sensitive sector like tech. In this kind of environment, the challenge becomes building exposure to long‑term AI growth without drifting into a concentrated valuation risk trade.
James Learmonth serves as co-chief investment officer at Harvest ETFs and oversees strategies including the Harvest Tech Achievers Growth & Income ETF (TSX:HTA). Over the same period, it declined only by about 7 percent, underscoring the difference between a diversified, income‑oriented structure and a pure software basket.
Why did tech stocks sell off in early February?
After piling into AI‑linked software and services names on strong cloud and AI‑related revenue growth, the technology sector underwent a steep correction from its October 2025 high. The decline followed earnings reports that included guidance pointing to sustained, capital‑intensive buildouts and longer payback periods.
After hyperscalers signaled aggressive 2026 infrastructure spending, market participants began to question return‑on‑investment timelines, even as fundamentals largely held up.
Companies with less certain paths to monetization saw their share prices decrease rapidly, while those showing profitable AI‑driven growth and measurable returns on invested capital were hit less hard. Disruption‑driven headlines, such as the launch of Anthropic’s Claude Cowork tools and new AI assistants aimed at legal and accounting workflows, added to the perception that many software business models are at risk, even if long‑term AI adoption remains intact.
The move exposed the limits of a purely thematic AI basket approach; in this environment, a passive, set‑and‑forget AI allocation can quickly morph from a growth‑oriented bet into a concentrated valuation risk trade, which is where active managers like Learmonth are trying to draw a sharper line between structural growth and speculation.
For Harvest ETFs, that line starts with business quality rather than a story about AI.
“Obviously it’s a rapidly evolving landscape across AI right now,” he said. “I think having competitive moats in place is paramount for companies maintaining their leadership position over time. From a valuation perspective, we like to look at P/E with that growth multiplier peg applied to us, so you have that growth lens applied to the valuation.”
Several lenses help distinguish structural winners from speculative names.
Learmonth pointed to growing margins, return on equity and return on invested capital as key markers that AI‑driven capex is actually creating value, rather than just inflating a headline growth story.
“You want to make sure companies are actually growing profitably, and not just generating revenue for the sake of generating revenue, but not able to pass that through in terms of bottom‑line growth as well. I think return on equity and return on invested capital, along those same lines, are key metrics to look at too,’ he noted.
Companies with clear, recurring AI‑related revenue streams, such as infrastructure or enabling hardware, tend to fare better than those whose AI exposure is largely driven by narrative.
“We have for a long time argued that the hardware and semiconductor side of the business is where we want to be (more heavily focused) right now, because it is seeing the revenue and profit generation directly from the infrastructure investment. That being said, particularly with the severity of the declines that we’ve seen in the software side over the past few weeks, I think (some opportunities) might be starting to spring up there,’ said Learmonth.
“We have reduced our software exposure a little bit over the past few quarters, but we are still maintaining some software exposure in those companies where we think they have competitive moats, whether that’s specialized areas like tax preparation and accounting, things like that,’ the expert elaborated.
Following the earlier correction, which Wedbush Securities analyst Dan Ives says may have been an overreaction, AI‑sensitive stocks are now trading at more reasonable multiples than at their October 2025 peak.
For the S&P 500 Software & Services group, the average forward P/E multiple has fallen from about 32.6 times to 22.7 times expected profits, even though analysts still forecast double‑digit revenue and earnings growth, plus net margins close to 30 percent. That average hides a wide gap between names that still trade on premium “AI story” multiples and others that have rerated much more sharply, which is where stock picking becomes critical.
In a recent note, Morgan Stanley (NYSE:MS) spotlighted Atlassian (NASDAQ:TEAM), Shopify (NYSE:SHOP) and Palo Alto Networks (NASDAQ:PANW) as some of the most compelling software opportunities for investors looking to buy the dip.
Investor takeaway
Against this backdrop, the focus is shifting from “how much AI” to “how AI is structured.’
For investors who want to stay exposed to AI‑driven tech, but are wary of sharp, headline‑driven swings, vehicles like the Harvest Tech Achievers Growth & Income ETF could offer a middle ground by combining active stock selection in structural winners with a covered‑call overlay.
“That’s how we generate enhanced yields — by selling calls on our long equity positions to generate option premiums, which we then pay as distributions on a fixed monthly basis,” explained Learmonth.
“That sale of options can help to mitigate some of the month‑to‑month volatility across the fund, with the tradeoff being some foregone upside in a strong bull market.”
As the AI trend evolves, success will likely favor those who view AI as a long-term, multi-year structural shift rather than a short-term theme. Winners will employ active management, prioritize income and utilize a disciplined structure to separate signal from noise.
Securities Disclosure: I, Meagen Seatter, hold no direct investment interest in any company mentioned in this article.