Is Broadcom’s AI Selloff a Buying Opportunity or a Reality Check?

NEW YORK — Broadcom posted what should have been a triumphant quarter — AI chip revenue up 143% year-over-year, free cash flow at $10.26 billion, margins that most semiconductors companies can only envy — and the stock dropped 15% anyway.

The Business Behind the Selloff

Broadcom is not a pure-play AI chipmaker in the Nvidia mold. It is a sprawling semiconductor and infrastructure software company whose AI revenue line — custom accelerators and networking silicon for hyperscale customers — has become the engine the market cares about most. That AI segment generated $10.8 billion in Q2 FY2026, a figure that grew 143% year-over-year and arrived above analyst forecasts. The problem was the next number: Q3 guidance of $16 billion, against a Street consensus of $17.2 billion.

That $1.2 billion gap sent AVGO down from a June 3 close of $479.23 to an intraday low of $405.51 on June 4 — the steepest single-session decline in more than 16 months. By the end of the week the stock had settled near $385, leaving it 22.3% below its 52-week high of $495 and, notably, trading beneath its June 2 all-time closing high of $481.57 by nearly $100.

The competitive narrative adds texture. Macquarie’s downgrade on June 4 flagged Google’s strategic shift toward developing its own application-specific integrated circuits internally — a move that could gradually reduce Broadcom’s role as a sole-source ASIC supplier to Alphabet’s data centers. That is a slow-moving risk, not an immediate revenue threat, but it introduces a customer concentration dynamic that bulls have historically dismissed. As we examined earlier this week, the guidance miss rippled across the entire chip sector, raising questions about whether AI infrastructure spending is beginning to moderate at the hyperscaler level.

Data Visual

Broadcom AI Semiconductor Revenue: Quarterly Actuals vs. Guidance ($B)

Shows the acceleration in Broadcom’s AI chip revenue from Q2 FY2025 through Q3 FY2026 guidance, highlighting the quarter where guidance fell short of consensus.

Values in $B

The Numbers: Strong Quarter, Uncomfortable Guidance

Strip away the guidance noise and the Q2 FY2026 results were genuinely strong. Total revenue of $22.19 billion grew 48% year-over-year. Adjusted EBITDA reached $15.24 billion, representing a 69% margin — a figure that few companies of any size can match. Non-GAAP diluted EPS of $2.44 beat consensus estimates of $2.39 to $2.40 and compared against $1.58 in Q2 FY2025, a 54% year-over-year improvement.

The valuation picture is nuanced. On a GAAP basis, AVGO trades at 64.21x earnings — elevated by any traditional measure. But the non-GAAP forward P/E of 24.55x, applied against consensus FY2026 EPS estimates of $11.61, tells a different story. The semiconductor industry has traded at a 3-year average P/E of 59x. At 24.55x forward on non-GAAP numbers, with EPS forecast to grow 32.8% per annum, Broadcom is not obviously expensive by sector standards.

The gross margin trajectory deserves attention. Q2 came in at 77.1%. The Q3 implied step-down toward 74% is small in absolute terms but directionally unwelcome for a stock that commands a premium precisely because of its margin profile. If that compression proves transitory — driven by product mix during a transition quarter — the bull case holds. If it reflects pricing pressure from hyperscalers pushing back on ASIC contracts, the story changes materially. Marvell faces a structurally similar question about whether AI networking margins are sustainable as customer bargaining power grows.

Data Visual

AVGO Analyst Price Target Range vs. Current Price (June 7, 2026)

Illustrates the spread between the current share price, the consensus target, and the outlier targets, showing where the Street thinks fair value sits after the selloff.

Values in $

What Wall Street Is Actually Saying

The analyst community responded to Thursday’s selloff with a near-uniform verdict: buy the dip. Of 27 analysts tracked by S&P Global, 24 carry Buy or equivalent ratings and three hold Neutral — zero sells. The consensus price target sits at $513.84, implying 33.2% upside from the current $385 level.

The target revisions that landed on June 4 were broadly constructive. Jefferies analyst Blayne Curtis raised his target to $550 from $500, explicitly calling the selloff a buying opportunity. Mizuho’s Vijay Rakesh lifted his target to $530 from $480, maintaining Outperform. JPMorgan went to $580 from $500 — the most aggressive revision — while Goldman Sachs moved to $525 from $500 and Deutsche Bank raised to $515 from $430. Morgan Stanley’s revised estimate of $502 was the most measured of the upgrades.

The outlier on the downside is the lowest analyst target on record: $215.88, representing a 44% discount to current prices. That figure almost certainly predates the AI revenue inflection and reflects a bear case built on traditional semiconductor cyclicality rather than the current growth profile. Still, its existence as an active target is a reminder that not every analyst has capitulated to the AI-demand-forever thesis. The macro backdrop matters too — a surprise inflation print heading into the June 17 Fed meeting could pressure the entire high-multiple tech cohort regardless of company-specific fundamentals.

The Levels That Will Define the Next Move

Broadcom’s technical picture deteriorated during the week. The MACD turned negative on June 5, a bearish momentum signal that typically precedes further selling pressure when confirmed by volume. Daily volume on June 4 hit 51.14 million shares against a 36.5 million average — 40% above normal, indicating institutional participation in the selloff rather than retail panic. That distinction matters: institutions selling with conviction are harder to reverse than retail traders chasing headlines.

The 200-day moving average at $349.84 represents the first major technical floor of consequence. A close below that level would constitute a regime shift — from a stock in a healthy correction within an uptrend to a stock breaking down. Between $385 and $350, the tape is ambiguous. Below $350, the burden of proof shifts firmly to the bulls. On the upside, a reclaim of $413 — the intraday high from June 7 — would be the first indication that dip buyers are regaining control. Full technical repair requires a sustained move back above $450, where the stock spent most of May before the earnings-related breakdown.

The fiscal 2027 AI revenue target of $100 billion-plus remains the central long-term catalyst. CEO Hock Tan’s decision not to raise that figure — despite Q2 AI revenue beating estimates — is what rattled investors most. As markets processed the Broadcom miss alongside last week’s payrolls data, the broader AI spending narrative absorbed its first credibility test of the year. The next proof point arrives with Q3 FY2026 results, expected in September. Between now and then, any public commentary from hyperscale customers — Google, Meta, Apple — about AI infrastructure capex will function as a leading indicator for whether the $16 billion Q3 guide was a floor or a ceiling.

Broadcom is not a broken company. Its fundamentals — 48% revenue growth, 69% EBITDA margins, $10.26 billion in quarterly free cash flow — remain among the strongest in large-cap technology. The selloff was a guidance story, not an earnings story, and the distinction matters for how quickly sentiment can recover. What changes the thesis in either direction: an upward revision to the FY2027 AI revenue target before September earnings would be unambiguously bullish. A public signal from Google that it is accelerating its internal ASIC program — or any hyperscaler announcing a pull-back in AI infrastructure spend — would validate the Macquarie downgrade and put the 200-day moving average under immediate pressure. The broader AI trade and the Fed’s June 17 meeting provide the macro backdrop against which all of that plays out. For now, AVGO sits in the uncomfortable middle: too cheap to ignore on forward multiples, too uncertain on guidance trajectory to chase.