2026-09-15

Nvidia -3% vs Intel -6%: Chipmakers Sank While Meta, Microsoft and Adobe Rose - Inside the 'AI Pacing' Rotation

Written & reviewed by the TradeSmrt editorial team · 2026-09-15

In this article

What Happened
Why Chip Stocks and Software/Hyperscaler Stocks Moved in Opposite Directions
What to Take Away From This
FAQ
Sources

What Happened

Monday, September 14 was another turbulent session for AI-linked stocks, driven by the weekend fallout from Anthropic CEO Dario Amodei's essay calling for the industry to "pace the frontier," and OpenAI CEO Sam Altman's confirmation that OpenAI was shelving its 2026 IPO plans. On the surface, the major indexes looked relatively calm: the S&P 500 slipped 0.48% to 7,619.98, the Nasdaq Composite fell 0.56% to 26,186.41, and the Dow Jones Industrial Average dropped 0.29% to 52,421.20. But underneath those headline numbers, the picture looked very different. The Philadelphia Semiconductor Index (SOX) tumbled 5.9% on the day - far steeper than the roughly 0.8% drop in the Nasdaq 100 (QQQ), confirming the selloff was concentrated in chips rather than the broader market.

What made the session even more striking was how unevenly the pain was distributed within the chip sector itself. Nvidia (NASDAQ: NVDA), the company most dominant in AI accelerators, fell a comparatively modest 3.4% to around $212.50. AMD (NASDAQ: AMD), widely seen as Nvidia's closest AI-chip challenger, slid as much as 6% intraday before paring losses to close down roughly 4.4% at $493.41. Broadcom also fell more than 4%. But the hardest-hit name of the three was Intel (NASDAQ: INTC) - arguably the least AI-accelerator-exposed of the group. Intel fell 6.4% in pre-market trading alone, sliding from $102.94 to $96.40, and closed the regular session down 5.6% at $97.19. In other words, the company with the most dominant position in AI compute fell the least, and the company with the least-established AI-accelerator footprint fell the most - the exact opposite of what you'd expect if the selloff were purely about AI chip demand.

Zoom out past semiconductors and the picture gets more paradoxical still. The so-called hyperscalers actually rose. Alphabet climbed almost 2%, Microsoft added 1.6%, and Meta Platforms gained roughly 1.4%. Enterprise software names did even better: ServiceNow, Adobe, and Workday all rallied between 4% and 7.4%. The companies that build and sell the chips fell sharply. The companies that buy those chips to run AI services (the hyperscalers), or layer AI features on top of existing software (the enterprise software names), rose on the very same day.

Why Chip Stocks and Software/Hyperscaler Stocks Moved in Opposite Directions

The key to understanding this divergence is asking who benefits from AI capital spending accelerating, and who benefits from it merely slowing down. Nvidia, AMD, Intel, and Broadcom are all valued, at bottom, on how much capital gets poured into AI infrastructure and how fast. Amodei and Altman's calls for a pacing slowdown directly challenged the assumption of unlimited, ever-accelerating capex growth that has underpinned chip-stock valuations - so any hint of a slower pace translates mechanically into lower revenue growth expectations for the companies that sell the hardware.

For Alphabet, Microsoft, Meta, and software names like Adobe, ServiceNow, and Workday, the logic runs in reverse. These companies are the buyers in this equation, already carrying enormous AI infrastructure spending commitments. For several quarters now, a recurring worry among their investors has been whether ballooning data center and GPU spending is simply inflating depreciation charges without a matching payoff. If the industry as a whole eases off the accelerator even slightly, these companies stand to see relief on their capital expenditure trajectory - and with it, an improved outlook for free cash flow and margins. Chipmakers need AI spending to keep rising to grow; hyperscaler margins can actually improve if AI spending growth merely slows. That structural mismatch is exactly what showed up as opposite stock reactions to a single piece of news.

Intel's outsized decline reflects a more layered exposure than either Nvidia's or AMD's. Intel had recently been building a fresh growth narrative around its CPUs serving as "orchestrators" for AI agents - a story that depends on continued enthusiasm for AI buildout. At the same time, Intel's foundry business depends on external customers signing long-term manufacturing commitments, which they're only willing to do if they're confident AI chip demand keeps climbing. That gives Intel two separate exposures to the same "pace of AI growth" variable, rather than one, which is part of why pacing fears hit it twice as hard. Compounding that, Piper Sandler had initiated coverage on Intel just days earlier with a cautious Neutral rating and a $110 price target, leaving the stock with fewer bullish voices to lean on during the selloff. Nvidia, by contrast, had its own CEO pushing back directly - Jensen Huang dismissed the AI doomsday scenarios as "totally unfounded" - while SemiAnalysis analyst Joey Brookhart offered a similarly reassuring view that "there's not going to be a material slowdown," both of which likely helped cap Nvidia's losses.

24/7 Wall St. characterized this ordering - Intel falling hardest despite the least AI-accelerator exposure, Nvidia falling least despite the most - as evidence of a positioning unwind rather than an actual demand shift. The interpretation: investors had built up AI-growth bets across the chip sector broadly, and when the pacing warning hit, those bets got unwound fastest and hardest in the names carrying the heaviest valuation risk or the least-proven growth story, which in this case was Intel.

What to Take Away From This

Within a single theme, separate the spenders from the sellers. Chipmakers need AI capex to keep growing to post revenue growth; hyperscalers and software companies can see their margins improve even when that same capex growth merely slows. The same headline can produce opposite stock reactions depending on which side of the spending relationship a company sits on.
When the order of declines doesn't match fundamental exposure, suspect positioning rather than a demand story. Intel falling hardest despite being the least AI-accelerator-exposed of the three chipmakers is a classic sign of a crowded trade unwinding, not a reassessment of actual chip demand.
Compare the sector index against the broad market index. SOX falling 5.9% against a roughly 0.8% drop in the Nasdaq 100 was a clear signal that Monday's selloff was chip-specific, not a market-wide risk-off event - a distinction worth checking any time a "market fell" headline appears.
Executive pushback and dissenting analyst views can meaningfully cushion a stock's decline. Jensen Huang's immediate rebuttal and SemiAnalysis's "no material slowdown" call likely helped keep Nvidia's losses smaller than its peers'. Watching who defends a stock publicly during a selloff - and who stays silent - is worth doing.

FAQ

Does Intel's bigger decline mean its AI business is worse than Nvidia's?

Not necessarily. The gap in declines looks more like a function of positioning and valuation than a fresh downgrade of either company's actual AI revenue outlook. Intel carries a dual exposure - its CPU-orchestration story and its foundry business both depend on continued AI buildout confidence - and had less bullish analyst support heading into the selloff. Whether Intel's underlying growth story has actually weakened will only be confirmed by future earnings and foundry order announcements.

Why did chip stocks fall while Big Tech and software stocks rose on the same news?

Chipmakers are the suppliers in the AI capex relationship - their revenue grows when AI infrastructure spending grows. Hyperscalers and software companies are the buyers, already shouldering enormous AI spending commitments. A signal that industry-wide AI development might pace more slowly reads as a growth risk for chipmakers, but as potential relief on capital spending - and a boost to margins - for the companies footing the AI infrastructure bill.

How long could this chips-down, software-up rotation last?

That's not yet clear. Whether it persists depends on whether Monday's pacing concerns evolve into an actual slowdown in AI development, or fade faster than expected. Chipmakers' order backlogs and hyperscalers' capital expenditure guidance in upcoming quarters will be the clearest signals of whether this rotation is temporary positioning noise or the start of a more structural shift.

Sources

This article is an original synthesis and analysis based on the reporting below, not a reproduction of the original articles. Please check the source articles directly for the most current figures.

⚠️ This article is for informational purposes only and is not investment advice. Market conditions change constantly - always verify the latest data before making investment decisions.