AI Chip Stock Selloff Erased $1 Trillion: Oracle Earnings Today Offer Next Recovery Test

Broadcom’s guidance hold, not its AI revenue beat, triggered the sector reset; Oracle reports after the close

Broadcom
SAN JOSE, CALIFORNIA - JUNE 03: A sign is posted in front of a Broadcom office on June 03, 2021 in San Jose, California. Justin Sullivan/Getty Images

The Philadelphia Semiconductor Index lost 10% in a single session last Friday — its worst drop in years — not because Broadcom delivered bad news, but because it failed to deliver news good enough for a sector that had been priced for perfection. That distinction matters for every investor deciding whether to buy, hold, or reduce exposure to AI-related equities today, as Oracle Corp. reports earnings after the close and May's Consumer Price Index prints this morning.

AI chip stocks including Nvidia, Micron Technology, AMD, and Intel are falling again in pre-market trading Wednesday, extending a week of volatile trading that has erased more than $1 trillion from semiconductor valuations since the sector peaked on June 2. The S&P 500 and Nasdaq hit all-time highs that Tuesday. By Friday, both indexes had surrendered several weeks of gains.

Broadcom Beat Estimates, Then Watched $1 Trillion Disappear

On Wednesday, June 3, Broadcom released fiscal second-quarter results that by nearly every historical measure were extraordinary. Revenue reached $22.19 billion, up 48% year over year. AI semiconductor revenue hit $10.8 billion, up 143% from a year earlier. CEO Hock Tan guided third-quarter AI revenue to $16 billion — growth of more than 200% year over year.

None of that was enough.

Analysts had expected Q3 AI chip sales of $17.2 billion. Broadcom also declined to raise its full-year AI semiconductor forecast, which Wall Street had built into its models. For a company whose shares had risen roughly 40% in 2026 alone and closed at an all-time high the session before earnings, the message the market heard was not "record growth" — it was "not fast enough." By Thursday close, Broadcom shares had fallen roughly 14%. By Friday, the damage had spread to every major chip name.

This is the defining feature of the current AI trade, and the detail most likely to cost investors who miss it: these are not stocks priced on what companies are delivering today. They are priced on what companies are expected to deliver next quarter, and the quarter after that. When a company beats the present but holds the future steady, holders of a stock at a premium multiple register that as a disappointment. The correction was not irrational. It was the math catching up.

May Jobs Report Poured Fuel on the Fire

The catalyst that turned Thursday's sector selloff into Friday's rout was a Bureau of Labor Statistics report showing nonfarm payrolls grew by 172,000 in May — well above the Dow Jones consensus estimate of 80,000. Strong employment data raised the prospect that the Federal Reserve will keep interest rates elevated longer than markets had anticipated.

The consequence for AI stocks is mechanical: high-multiple growth companies are especially sensitive to interest rates because their value is concentrated in earnings projected years in the future. When the discount rate applied to those future earnings rises, valuations fall even without any change in the underlying business. The 10-year Treasury yield jumped above 4.5% and the 30-year above 5% following the jobs data. As of Tuesday, the 10-year yield remained at 4.56%.

The Nasdaq Composite dropped 4.2% on June 5 — its worst session since April 2025 — and chip stocks ended the week with their steepest losses in years. The S&P 500 lost 2.64%. The Dow fell nearly 700 points.

Five Forces Behind the Selloff

1. The expectations trade snapped. The Philadelphia Semiconductor Index had surged more than 50% in the twelve months leading into this correction. Individual names posted even more extraordinary returns. Micron Technology, which had more than tripled in 2026 and crossed a $1 trillion market capitalization in late May, fell 13.3% on June 5 alone — the largest single-session loss in the S&P 500 that day — on no company-specific news. Wells Fargo's chief equity strategist described the semiconductor sector as "way overbought" before the selling began.

2. The jobs report revived rate fears. The May payroll number more than doubled forecasts, reinforcing the case that the Federal Reserve may not cut rates as quickly as the market had priced. Rate-sensitive growth stocks absorb that signal immediately through multiple compression.

3. Capital is being called away. BNP Paribas equity derivatives strategist Greg Boutle warned that retail investors were liquidating chip positions to fund participation in SpaceX's initial public offering, which is expected to price Thursday evening at a roughly $1.75 trillion valuation — the largest IPO in US history. Micron alone had received $6.5 billion in net retail inflows over the prior month, making it a natural source of funds. When crowded positions unwind and buyers at elevated prices disappear, cascading declines follow quickly.

4. Software faces its own AI anxiety. Separate from the chip selloff, software stocks are under pressure from investor concerns that AI tools are beginning to displace traditional software revenue. Analysts at Bank of America described the resulting downturn as an "indiscriminate selloff" not reflecting a change in underlying fundamentals. William Blair analysts called it driven by "fear, not fundamentals."

5. The sector was technically overextended. The iShares Semiconductor ETF broke above its upper Bollinger Band on June 2 — a standard signal that a reversal was statistically overdue. When positioning becomes this one-sided, even a modest miss in any category can trigger stop-losses and algorithmic selling that amplifies the move well beyond what the fundamental news justifies.

Carnage Across the Sector

Stock / IndexDecline
Micron Technology (MU)13.3% (June 5)
Intel (INTC)11.3% (June 5)
AMD10.9% (June 5)
Broadcom (AVGO)~14% (June 4–5)
Nasdaq Composite~5% from June 2 high
iShares Semiconductor ETF (SOXX)~10% (June 5)

The S&P 500 remains up roughly 8% year-to-date. The Nasdaq is still positive for the year. But investors who bought at the all-time high on June 2 and held through Friday's close absorbed meaningful losses, and as of Wednesday morning those losses have not fully healed.

Recovery Has Begun — But Is Not Complete

Monday, June 8, brought a partial rebound. Micron gained 9.9% after its Friday plunge. Marvell Technology climbed 9.6% on its first day after S&P Dow Jones Indices announced the company would join the S&P 500. The S&P 500 gained 0.3% and the Nasdaq added 0.86%.

Intel provided the session's clearest company-specific catalyst: the stock surged more than 11% after reports that Alphabet had contracted the company to manufacture more than 3 million custom AI chips — a contract that, if confirmed, would represent a significant validation of Intel's foundry strategy and its effort to re-establish itself as a manufacturing force in AI silicon.

Tuesday was choppier. The S&P 500 fell 0.26% to close at 7,386, and the semiconductor ETF SMH shed 1% after its Monday rebound, confirming that single-session recoveries have not yet re-established the sustained uptrend that defined the sector through most of the first half of 2026.

As of Wednesday morning, Nvidia, Micron, AMD, and Intel are all trading lower in pre-market again.

Is This a Buying Opportunity?

Not all market participants are treating the selloff as a warning. "A lot of the sell-off from our perspective is an opportunity to buy some really essential, critical AI infrastructure stocks at cheaper prices," said Rob Thummel, portfolio manager at Tortoise Capital.

Bank of America pointed to the closest historical parallel: the DeepSeek panic of January 2025, which briefly erased more than $1 trillion in market cap. Global cloud capital spending still grew 69% through 2025 following that episode, and major indexes not only recovered but finished the year up by double digits.

Morningstar analysts, updating their models as of June 8, rated Nvidia as approximately 25% undervalued relative to a $280 fair value estimate. Nvidia CEO Jensen Huang called the selloff a buying opportunity at his Computex keynote in Taiwan.

The counterargument is valuation. AMD traded at a forward price-to-earnings ratio of roughly 84 times heading into the correction — more than double its five-year median. For a market that rewards companies not for meeting earnings expectations but for exceeding them, sustaining current multiples requires not just record growth but accelerating record growth. The Intel Alphabet contract win offered a reminder that company-specific catalysts can cut through sector-wide sentiment: Intel's 11% single-session gain on a real order demonstrates that AI infrastructure demand is still real enough to move stocks sharply higher on concrete news.

What Oracle Reports Tonight Will Tell Investors

The AI rally's next significant test arrives after the market closes Wednesday. Oracle Corp., whose cloud infrastructure business has grown directly alongside AI spending, reports earnings. Its remaining performance obligations hit $553 billion at the end of fiscal Q3 — up 325% year over year — representing signed contracts that have not yet been recognized as revenue. That figure is either a revenue bridge stretching years into the future or a delivery problem waiting to surface, and tonight's report will begin to resolve which.

A strong Oracle result with raised guidance could help stabilize the semiconductor sector by confirming that enterprise AI demand is as robust as the hyperscalers have claimed. A cautious outlook — on margins, on the pace of backlog conversion, or on cloud infrastructure capital expenditures — could extend the selloff into a second week.

Nvidia's next full earnings report is scheduled for August 26, 2026. As the company that built the infrastructure powering the generative AI build-out, its guidance will function as a sector-wide verdict on whether AI spending is pausing or plateauing.

The question the market is answering in real time — on every tick of Oracle's earnings call tonight — is whether the AI infrastructure build-out is merely catching its breath, or whether expectations for this trade finally ran ahead of the business.


Frequently Asked Questions

Why did AI chip stocks fall so sharply despite Broadcom reporting record revenue?

Broadcom reported AI semiconductor revenue of $10.8 billion, up 143% year over year, beating analyst estimates. The problem was guidance: the company projected third-quarter AI chip sales of $16 billion, roughly $1.2 billion below what analysts had modeled. For a stock priced at a substantial premium to reflect hypergrowth assumptions, holding guidance steady functions as a miss — the market does not reward meeting expectations; it rewards exceeding them. The sector had priced in continuous acceleration, and flat guidance was enough to trigger a 14% single-session decline.

What caused the semiconductor market correction in June 2026?

Three factors converged: Broadcom's below-consensus AI chip guidance triggered the initial selloff on June 4–5; a stronger-than-expected May jobs report revived Federal Reserve rate concerns and raised the discount rate applied to high-multiple growth stocks; and a sector that had already surged more than 50% over the preceding twelve months had limited tolerance for any shortfall in the growth narrative. When algorithmic stop-losses and crowded retail positions began unwinding simultaneously, the moves were amplified well beyond what the fundamental disappointment alone justified.

Are AI chip stocks a buy or a sell after this selloff?

The answer depends on entry price and time horizon. Bullish analysts, including Morningstar and Tortoise Capital's Rob Thummel, view the selloff as a valuation reset in a structurally intact AI spending cycle; major cloud providers have committed more than $725 billion in AI infrastructure spending in 2026 alone. Bearish signals include AMD trading at more than 84 times forward earnings and Micron's recent price well above long-term intrinsic value models. The closest precedent — the January 2025 DeepSeek panic — resolved with a full recovery and continued cloud capital spending growth of 69% through year-end.

What role does Oracle's June 10 earnings report play in the AI stock recovery?

Oracle's cloud and data center business tracks closely with enterprise AI infrastructure demand, and its $553 billion backlog of signed but unrecognized contracts makes it a direct readout on whether that demand is converting into actual revenue. A strong Oracle result with raised guidance could confirm the AI spending cycle is intact and stabilize the semiconductor sector; a cautious outlook on margins, delivery pace, or capital expenditures could extend the selloff into a second week.

ⓒ 2026 TECHTIMES.com All rights reserved. Do not reproduce without permission.

Join the Discussion