What's happening
The Philadelphia Semiconductor Index fell 10.3% on June 4–5, 2026, its largest single-session decline since March 2020, as Broadcom's fiscal second-quarter earnings report triggered a broad rotation out of AI-related semiconductor names. Although Broadcom reported record Q2 revenue and earnings, its forward guidance fell short of elevated market expectations: the company issued fiscal Q3 AI semiconductor revenue guidance of $16 billion against analyst consensus of $17.2 billion, and Chief Executive Hock Tan confirmed the company was leaving its long-range AI chip sales forecast unchanged at $100 billion. The combination of a guidance miss and a flat long-range outlook on a metric that had been a primary driver of sector enthusiasm catalyzed the selloff.
Broadcom shares fell 12.59% to close at $418.91 on June 4, after a pre-market decline that exceeded 14% and erased more than $310 billion in market value. The pressure radiated across the sector: Nvidia fell approximately 6.2%, erasing more than $300 billion in market capitalization; Micron plunged 13%, wiping out roughly $150 billion in market value; AMD dropped 11%; and Marvell Technology declined 17%. The aggregate loss across the index reached approximately $1.3 trillion, according to Reuters reporting dated June 5, 2026.
Why it matters for markets
The scale of the single-session loss — $1.3 trillion in combined semiconductor sector market capitalization — underscores the degree to which AI-driven revenue expectations had been embedded in chip stock valuations heading into Broadcom's report. Broadcom's $16 billion fiscal Q3 AI semiconductor revenue guidance, coming in $1.2 billion below the $17.2 billion consensus estimate, was sufficient to call into question whether the pace of AI infrastructure spending can sustain the growth trajectories priced into the sector. The flat maintenance of the $100 billion long-range AI revenue forecast, rather than an upward revision, amplified that concern.
The breadth of the decline illustrates the interconnected nature of AI chip exposure across the semiconductor landscape. Nvidia, with a market capitalization of approximately $4.97 trillion and revenue of $253.49 billion, absorbed a loss of more than $300 billion in a single session. Micron, whose high-bandwidth memory products are central to AI accelerator deployments, saw roughly $150 billion erased from its market cap. AMD, trading at a price-to-earnings ratio of 156.5, faces particular scrutiny given the premium its valuation assigns to AI-related growth. The selloff reflects how tightly correlated these names have become through their shared exposure to AI infrastructure spending cycles.
The steepness of the Philadelphia Semiconductor Index's 10.3% decline — a threshold not crossed since the early-pandemic market dislocations of March 2020 — signals that the event moved beyond routine earnings volatility into territory that may prompt reassessment of sector positioning. Broadcom's 52-week trading range of $241.11 to $495.00 provides context for the magnitude of the single-session move, with the stock closing at $418.91 after the report.
Sectors and assets to watch
The primary tickers directly affected span the full spectrum of AI semiconductor exposure. Broadcom (AVGO), the catalyst for the selloff, designs custom ASICs and networking chips central to hyperscaler AI infrastructure; its fiscal Q3 guidance and unchanged $100 billion long-range forecast will remain the reference point for near-term sector sentiment. Nvidia (NVDA), whose H100 and A100 data center GPUs and CUDA software ecosystem have made it the dominant supplier of AI training hardware, saw more than $300 billion in market cap erased in the session. Micron (MU), which supplies high-bandwidth memory used in AI accelerators and has a 52-week range of $103.38 to $1,089.29, fell 13% as investors reassessed demand trajectories for memory tied to AI workloads. AMD (AMD), which competes in AI accelerators through its Instinct product line, dropped 11%, with its elevated P/E ratio of 156.5 making it particularly sensitive to any revision in AI growth expectations.
Marvell Technology, which also designs custom AI chips for hyperscalers and fell 17% in the session, represents another node in the custom ASIC supply chain that warrants monitoring alongside Broadcom. The breadth of the decline — spanning GPU suppliers, memory manufacturers, custom ASIC designers, and CPU vendors with AI exposure — indicates that the repricing was sector-wide rather than confined to Broadcom's specific customer relationships.
What to watch next
Market participants will be focused on whether Broadcom's $100 billion long-range AI revenue forecast holds through subsequent guidance updates, and whether the $16 billion fiscal Q3 AI semiconductor revenue figure proves to be a floor or a ceiling as hyperscaler capital expenditure plans are disclosed in coming earnings cycles. Upcoming quarterly results from Nvidia, AMD, and Micron will be closely scrutinized for any corroborating or contradicting signals on AI chip demand; Micron's high-bandwidth memory order trends in particular will serve as a leading indicator for AI accelerator build rates. The Philadelphia Semiconductor Index's recovery trajectory from its June 4–5 low will also be watched as a gauge of whether the sector stabilizes around revised AI revenue expectations or continues to reprice.