In June 2026, the global semiconductor industry underwent a brutal repricing of "expectations."
The spark that ignited this shift can be traced back to June 3. After Broadcom (AVGO) released its Q2 FY2026 earnings, the company issued Q3 AI semiconductor revenue guidance of $16 billion, falling short of analysts’ consensus estimate of $17.2 billion. Even more disappointing for the market, Broadcom did not raise its full-year FY2026 AI chip sales forecast of $56 billion, while the average analyst estimate hovered around $57.6 billion. Following the earnings release, Broadcom’s stock plunged more than 15% in after-hours trading.
This "miss" in guidance triggered a chain reaction over the following days. On June 5, the PHLX Semiconductor Index plummeted 10.3% in a single day—the largest one-day drop since March 2020. The AI and chip sectors lost over $1.3 trillion in market value in a single session. Nvidia fell about 6%, erasing over $300 billion in market cap in one day; AMD dropped nearly 11%; Micron Technology tumbled more than 13%.
However, the sell-off didn’t end on June 5. By June 23, the PHLX Semiconductor Index had fallen another 7.87%, marking its steepest single-day drop since June 5. The Nasdaq closed down 2.21%, and the S&P 500 fell 1.43%. SanDisk and Micron Technology both dropped over 13%; ARM fell more than 10%; Qualcomm, Western Digital, and Applied Materials lost over 8%; ASML and NXP dropped over 7%; TSMC and Intel fell more than 6%; AMD lost over 5%; Nvidia was down 4.15%; and Broadcom fell more than 3%.
This article seeks to answer a core question: How could a roughly $1.2 billion gap between AI chip revenue guidance and market expectations trigger a $1.3 trillion wipeout in market value? Was this an overreaction driven by market sentiment, or does it mark a structural turning point in how AI chip valuations are determined?
Broadcom Earnings: Outperforming Results, "Underwhelming" Guidance
To understand the root of this market value purge, we must first look at the full picture of Broadcom’s earnings—it was far from a "bad" report.
In Q2 FY2026, Broadcom posted total revenue of $22.187 billion, up 48% year over year. Adjusted EPS reached $2.44, a 54% increase and above analyst expectations of $2.39. AI semiconductor revenue hit $10.8 billion, up 143% year over year, breaking the $10 billion quarterly mark for the first time. Semiconductor solutions revenue reached $15.01 billion, up 79% and well above the market’s $14.65 billion estimate. The company projected total Q3 revenue of about $29.4 billion, nearly 2.8% higher than analyst forecasts.
By any measure, this was a robust earnings report.
The issue lay in "expectations." In the five trading days before the earnings release, Broadcom’s market cap surged by about $270 billion, as investors priced in aggressive growth. Options markets priced in a post-earnings single-day move of about 7.8%, higher than the historical average. The market was not just looking for an earnings beat, but for a "blowout" beat.
Yet Broadcom’s Q3 AI semiconductor revenue guidance stood at $16 billion—over 200% year-over-year growth, but below the $17.2 billion analysts expected. That’s a $1.2 billion shortfall, or about a 7% miss. The full-year FY2026 AI chip sales guidance of $56 billion also lagged the $57.6 billion consensus. On the earnings call, CEO Hock Tan reiterated the FY2027 AI semiconductor revenue target of over $100 billion, but did not raise it.
CFRA Research Senior Vice President Angelo Zino summed up the market’s mood: "The bar was set way too high ahead of the earnings release." Broadcom delivered an excellent report, but the market demanded perfection. When "excellent" meets "sky-high expectations," an "underwhelming" narrative is enough to trigger widespread profit-taking.
From Guidance Miss to $1.3 Trillion Lost: Breaking Down the Transmission Chain
The reason Broadcom’s "underwhelming" guidance could trigger such a massive market value wipeout comes down to a three-layered transmission chain.
First Layer: Concerns Over Slowing Marginal AI CapEx Growth. Broadcom is a core supplier of AI infrastructure, designing custom AI accelerator chips (XPUs) and networking chips for hyperscale data center clients like Google, Meta, OpenAI, and Anthropic. Hyperscale data center AI spending is projected to reach $650 billion in 2026. When Broadcom chose to maintain rather than raise its AI revenue guidance, the market began to ask a deeper question: While AI capital spending is still growing, is its marginal growth rate starting to slow?
Second Layer: Rigid Expectations at High Valuations. According to LSEG, the Magnificent Seven trade at an average forward P/E of about 28x. This valuation isn’t extreme in itself, but the pricing logic has shifted—the market is no longer satisfied with "profitable," but demands "consistently beating expectations." Broadcom’s case shows that when a company’s stock price has already surged on AI optimism ahead of earnings, any guidance that falls short of a "blowout" can trigger systemic selling.
Third Layer: Structural Fragility of Crowded Trades. Since the start of 2026, the PHLX Semiconductor Index had soared as much as 90%. Morgan Stanley Investment Management Senior Portfolio Manager Andrew Slimmon noted, "AI beneficiaries are being indiscriminately sold. I don’t think they’re expensive, but they’re too crowded." The essence of a crowded trade is that once the catalyst reverses, selling becomes self-reinforcing and amplifies the decline.
The June 23 Retest: The Crowded Trade Continues
After the June 5 crash, the market saw a brief rebound—on June 8, US chip stocks bounced back, and on June 9, Asian chip names rallied strongly, with Korea’s Kospi jumping 8.2% as Samsung Electronics and SK Hynix surged 9% and 15.9%, respectively. But the recovery didn’t last.
On June 23, the PHLX Semiconductor Index plunged another 7.87%, closing at 13,482. The Nasdaq fell 2.21% to 25,587, the S&P 500 dropped 1.43% to 7,365, while the Dow Jones Industrial Average was nearly flat, down just 0.09%. This divergence made it clear: selling pressure was highly concentrated in tech and semiconductors, while traditional industries held up relatively well.
Multiple factors triggered this round of selling. After the Fed’s FOMC meeting last week sent hawkish signals, the interest rate futures market reflected a roughly 70% probability of at least a 25-basis-point hike by the end of September. Rising rate expectations increased funding costs, putting pressure on AI infrastructure investments that rely on ongoing capital expenditures. Meanwhile, Asian markets led the decline—Korea’s Kospi dropped nearly 10%, with Samsung Electronics leading the fall. Quarter-end portfolio rebalancing also accelerated institutional outflows.
A key structural factor was the amplifying effect of leveraged ETFs. The SOXL 3x leveraged long semiconductor ETF, for example, triggers forced liquidations of derivatives during market downturns, further magnifying index declines. Some analysts noted that certain ETFs had already fallen nearly 20% before the market even opened: "This isn’t just a bad morning—it signals that structural factors are amplifying the drop."
From June 5 to June 23, the PHLX Semiconductor Index experienced eight single-day swings of more than 5% in one month—the most volatile stretch since the pandemic-driven sell-off in March 2020. This level of volatility signals that the market is searching for a new pricing anchor.
The Deeper Logic of "Expectation Gaps": Shifting Valuation Anchors
The core issue exposed by the Broadcom episode isn’t a crack in the fundamentals of the AI chip industry, but rather a shift in the market’s valuation anchor for AI chip companies.
From 2025 through the first half of 2026, the market’s pricing logic for AI chip companies could be summed up as "the faster the growth, the higher the valuation"—as long as revenue growth consistently beat expectations, valuation tolerance was virtually unlimited. Broadcom’s AI semiconductor revenue jumped from about $20 billion in FY2025 to a projected $56 billion in FY2026, an increase of roughly 180%. Stocks like Nvidia, Broadcom, and AMD repeatedly hit new highs under this logic.
But after the Broadcom event, the pricing logic has shifted from "the faster the growth, the better" to "can growth consistently beat expectations?" When a company’s AI revenue has reached $10.8 billion in a single quarter and $56 billion for the year, maintaining triple-digit growth rates becomes exponentially more difficult. Analyst models for Broadcom’s FY2027 AI revenue have already been raised to $114 billion, while the company’s reiterated target is merely "over $100 billion." Some institutions are projecting next year’s AI revenue at over $130 billion.
This gap between expectations and guidance essentially reflects diverging views on the "AI capital expenditure cycle inflection point." Bulls argue that AI infrastructure investment is still in its early stages—Goldman Sachs projects S&P 500 companies’ capital expenditures will rise 33% in 2026. Bears counter that while hyperscale data center AI spending is still growing, its marginal growth rate can’t remain exponential indefinitely.
Bernstein analyst Stacy Rasgon raised Broadcom’s price target to $550 after earnings, arguing that the market overreacted to cautious guidance. In contrast, Macquarie analyst Arthur Lai downgraded Broadcom from "Outperform" to "Neutral," cutting the price target from $513 to $437. This analyst divergence highlights the current lack of consensus on AI chip valuations—a lack of consensus that makes markets especially fragile.
Conclusion: Lessons from the Expectation Gap
A roughly $1.2 billion gap between Broadcom’s AI chip revenue guidance and market expectations ultimately triggered a $1.3 trillion market value shakeout. This stark contrast reveals the core feature of today’s AI chip market: valuations are no longer a function of fundamentals, but of expectations.
Broadcom’s earnings were not poor—revenue grew 48%, AI chip revenue soared 143%, and EPS beat estimates. But in a bull market that had pushed the PHLX Semiconductor Index up 90%, "not poor" wasn’t nearly enough. The market wanted not just "growth," but "growth that beats expectations"; not just "in line with guidance," but "raised guidance."
From June 5 to June 23, the semiconductor sector underwent a systemic reset of expectations. This doesn’t mean the long-term growth story for AI chips is broken—AI capital spending continues to expand, demand for custom chips remains strong, and hyperscale data center procurement plans are intact. But the market is learning to distinguish between "industry growth" and "stock price growth," and is reassessing what level of growth justifies what kind of valuation.
For investors, the lesson from Broadcom may be this: In a market where "consistently beating expectations" is the pricing baseline, the greatest risk isn’t declining results, but results that "merely" meet expectations. When expectations become the sole anchor for valuation, any signal less than "perfection" can trigger a disproportionate repricing.
The long-term AI chip narrative hasn’t changed, but the market’s short-term tolerance has. The $1.3 trillion market value purge didn’t erase the future of AI—it wiped out expectations that had been overextended during the past two years of the bull market.
FAQ
Q: What is Broadcom’s Q3 FY2026 AI chip guidance, and how does it compare to market expectations?
Broadcom’s Q3 FY2026 AI semiconductor revenue guidance is $16 billion, representing over 200% year-over-year growth but falling short of the consensus estimate of $17.2 billion—a gap of about $1.2 billion (roughly 7%).
Q: How much market value was wiped out in the semiconductor sector on June 5?
On June 5, 2026, the PHLX Semiconductor Index plunged 10.3% in a single day, with the AI and chip sectors losing over $1.3 trillion in market value.
Q: What was the PHLX Semiconductor Index’s decline on June 23?
On June 23, 2026, the PHLX Semiconductor Index fell 7.87%, closing at 13,482—the largest single-day drop since June 5.
Q: What is Broadcom’s full-year FY2026 AI chip revenue guidance?
Broadcom expects FY2026 AI semiconductor revenue to reach $56 billion, up about 180% year over year, but below the analyst consensus of $57.6 billion.
Q: Is this semiconductor sell-off a bursting of the AI bubble or just a short-term correction?
The prevailing market view is that this is a "correction of crowded trades" rather than a "collapse of industry fundamentals." AI capital expenditures continue to expand, but the market’s pricing logic for "consistently beating expectations" is being recalibrated.
