The Dominance of Mega-Cap Tech Fades as Earnings Broaden

For years, a handful of mega-cap tech stocks drove almost every meaningful gain in the US market. That grip is loosening and the earnings data are to thank for it.
The equal-weight S&P 500 (a version of the index that treats every company the same regardless of size) is outperforming the standard market-cap-weighted gauge for the first time since 2022.
Morgan Stanley strategists, led by Michael Wilson, point to the data behind this move. The median S&P 1500 constituent is generating earnings-per-share growth of more than 10%, the best reading since the post-Covid recovery.
Analysts are still upgrading profit estimates for consumer discretionary and transport stocks, two sectors closely tied to economic growth.
> "We expect the broadening to continue, driven by earnings resiliency from the median stock."
> Michael Wilson, Morgan Stanley
Second-quarter earnings season just kicked off. Analysts expect S&P 500 companies to report a 23% jump in profits, among the best readings outside of major recession recoveries.
The tech giants are not struggling, but their path to earnings growth is getting harder to read. The core issue is capital expenditure, or capex, which is the money companies spend on physical infrastructure like data centers and chips.
Microsoft invested roughly $65B in cloud and AI infrastructure in fiscal 2025 while generating annualized AI revenue of roughly $37B. That gap is what's making investors nervous.
Rising depreciation costs from all that spending are expected to slow earnings-per-share growth and reduce the cash available for share buybacks — a key support mechanism for stock valuations.
Microsoft fell 20% in the first half of 2026 while Oracle dropped 27%. Alphabet gained 14%, but the group as a whole has largely missed this year's broader S&P 500 rally.
The so-called hyperscalers (companies building and running the biggest cloud infrastructure) are now under scrutiny over whether AI revenue will eventually justify the scale of investment.
There is also a new hardware risk. Morgan Stanley researchers flagged what they call "chipflation," where memory chip prices rise sharply as demand outstrips supply. This adds costs across AI supply chains and squeezes margins for anyone building or buying infrastructure.
Geopolitics are adding another layer. Escalating US-Iran strikes sent Asian chipmakers sharply lower on Monday. SK Hynix fell 15.4% in Seoul in a single session. Micron dropped 5.2% in US trading. The Nasdaq 100 slid 1.4% as oil prices jumped on renewed Strait of Hormuz tensions.
Outside the Mag 7 orbit, several companies are executing on distinct business pivots that do not depend on AI infrastructure spending.
Lyft is expanding into robotaxis in markets like Nashville through a partnership with Waymo, while also moving into Europe through acquisitions.
Coinbase is building what CEO Brian Armstrong calls an "everything" app, integrating AI agents that let users manage crypto portfolios through plain-English instructions.
Axon, the maker of Tasers and security drones, is developing a next-generation weapon it expects will outperform a nine-millimeter bullet, a product still in testing but with a projected completion by next winter.
None of these stories hinges on hyperscaler capex or chip supply chains. That is precisely the point.
For investors who have spent the last two years watching Nvidia and peers, the rotation signal is already visible in the data.
Broadcom offers exposure to custom silicon and high-speed networking, the infrastructure layer that scales with AI deployments but sits outside the pure-GPU trade.
RBC Capital Markets recently upgraded the broader tech sector to overweight, citing strong earnings estimate upgrades and renewed fund inflows. The firm noted that tech valuations are only slightly above their long-term average on a median basis.
The Mag 7 is not disappearing, but the market is no longer waiting for them to lead.