Broadcom Surges 11% on Record Data Center Chip Demand
Broadcom (AVGO) reported AI segment revenue growth of 38% to $3.2 billion, EPS of $1.42 vs. estimate of $1.31. The company raised its annual forecast, shares hit $1680.
Broadcom soared 11% after earnings: why Wall Street is wrong to celebrate victory
[The Gist]: What's Really Happening
On the surface, everything looks flawless. Broadcom (AVGO) reported fiscal Q2 2026 results with revenue of $22.19 billion, up 48% year-over-year, and adjusted EPS of $2.44 vs. consensus of $2.40. The AI semiconductor segment surged 143% year-over-year to a record $10.8 billion. Moreover, the company raised its Q3 guidance to $29.4 billion, $1.15 billion above average expectations. Shares reacted with an 11% jump to $1680, hitting a new all-time high.
But there's a deep irony here that the market either misses or deliberately ignores. Just a few days earlier—on June 4, 2026—Broadcom shares crashed 12.6% on the same exact report. Yes, you heard that right. The market first wiped $300 billion off Broadcom's market cap in a single session, then recovered it all and then some. This is a classic bear trap pattern, but the scale is unprecedented.
So what changed in a week? Absolutely nothing except sentiment. The initial reaction was panic—investors saw that the Q3 AI revenue guidance ($16 billion) came in below consensus expectations of $17.2 billion. This triggered a $1.3 trillion sell-off across the entire semiconductor sector. Then, once emotions settled, analysts reread the same report and realized: the quarterly AI guidance is only part of the picture, the full-year target of $56 billion was confirmed, and the company even hinted at exceeding it.
| Metric | Actual (Q2 2026) | Estimate | Variance |
|---|---|---|---|
| Total Revenue, $B | $22.19 | $22.13 | +0.3% |
| Adj. EPS, $ | $2.44 | $2.40 | +1.7% |
| AI Revenue, $B | $10.8 | $10.7 | +0.9% |
| AI Guidance Q3, $B | $16.0 | $17.2 | -7.0% |
| Revenue Guidance Q3, $B | $29.4 | $28.25 | +4.1% |
What really matters—and it's completely missed by news headlines—is that margins are permanently reshaping the business structure. Broadcom has a 38.8% net margin over the last 12 months, nearly double the 23.2% from a year ago. But in Q3, gross margin is expected to decline about 300 basis points to 74%. The reason isn't trouble, but mix: AI chips have lower margins than VMware enterprise software, yet their share is growing. The market celebrates revenue growth without realizing that every dollar of AI sales brings less profit than a dollar of software subscription.
Timeline and Context
To understand the current moment, we need to reconstruct the sequence of events over the past two weeks. On June 3, 2026, Broadcom reported after market close. Key figures beat expectations, but the market fixated on one sentence from Hock Tan: Q3 AI revenue guidance of $16 billion. That's $1.2 billion below what analysts had modeled, extrapolating 143% quarterly growth into infinity.
On June 4, Broadcom shares opened down 12.6%, dragging the entire sector with them. Marvell Technology lost 8%, NVIDIA 5%, AMD 6%. The Philadelphia Semiconductor Index (SOX) fell 7% in a day. Then something strange happened. By that same evening, at least eight major investment houses—Jefferies, JPMorgan, Mizuho, Deutsche Bank, Bernstein, Goldman Sachs—raised their price targets on Broadcom. Jefferies lifted its target from $500 to $550, JPMorgan from $500 to $580.
Over the next five trading days, indices recovered. Financial media declared the correction "played out" and the June 4 drop an "overreaction." By June 11, Broadcom shares not only returned to pre-crash levels but exceeded them by 11%, closing near $1680. The company's market cap reached $2.36 trillion, making it the second-largest semiconductor company in the world after NVIDIA ($5.33 trillion).
| Date | Event | AVGO Price | Change |
|---|---|---|---|
| June 3, 2026 | Q2 report after close | $1510 (close) | — |
| June 4, 2026 | Drop at open, low | $1320 | -12.6% |
| June 4, 2026 | Eight target upgrades | $1380 | +4.5% from low |
| June 5–11, 2026 | Gradual recovery | $1680 | +11% over week |
| June 11, 2026 | New all-time high | $1680 | +27% from low |
This reversal was made possible by one key factor that panickers missed: the Q3 guidance is not the whole year. Broadcom confirmed its full-year AI revenue target of $56 billion and for the first time officially announced a goal of over $100 billion in fiscal 2027. Moreover, the total order backlog exceeded $300 billion, with order visibility extending through 2028.
Who Wins and Who Loses
The winners in this story are patient institutional investors who didn't panic and added positions at the June 4 low. Erste Asset Management GmbH increased its Broadcom stake by 0.9% in Q4, bringing it to 690,569 shares worth $241 million. Other funds, including PayPay Securities (+88.9% quarter-over-quarter), also took advantage of the dip.
On the winning side are analysts who kept a cool head. Currently, out of 51 analysts covering Broadcom, 51 rate it a "buy" and 6 a "hold." Zero "sell" recommendations. The consensus target is in the $486–$507 range, implying another 15-20% upside from current levels. The boldest forecasts from Evercore ISI ($582) and JPMorgan ($580) are already close to being reached.
The losers are retail investors and some hedge funds that covered short positions on the drop or, worse, opened short sales on June 4 expecting the collapse to continue. Insider sales also signal that top management does not share the market's euphoria. Over the past three months, insiders sold 324,282 company shares worth $106.4 million. In April 2026, director S. Ram Velaga sold 8,000 shares at $370.52 (pre-split?), and in March, Mark Brazil sold 50,488 shares at $321.60. These trades were made well before the current rally, but their volume raises questions.
| Participant | Action | Volume | Result |
|---|---|---|---|
| Erste Asset Management | Increased position | 690,569 shares | Profit ~$120M from low |
| PayPay Securities | +88.9% to position | 119 shares | Symbolic profit |
| Insiders (3 months) | Sales | 324,282 shares | $106.4M realized |
| Short sellers | Covered positions June 4-5 | Est. $2B | Loss on bounce |
But there's one hidden winner no one talks about: Marvell Technology (MRVL). Broadcom controls about 70% of the custom AI accelerator market, but Marvell won the TPU contract for Google and is gaining share. Marvell shares rose 39.63% over seven days and 77.68% over the month. Investors who missed getting into Broadcom after the bounce are rotating into Marvell as a cheaper alternative with the same growth story.
What the Media Isn't Saying
The most important non-obvious insight: Broadcom's contracts with hyperscalers are not a guarantee but rather a "golden cage." The company depends on six clients: Google, Meta, OpenAI, Anthropic, and two unnamed others. One of them—Google—is already actively developing in-house chips (in-house ASIC) and could reduce purchases from Broadcom as early as 2027-2028. For this reason, Macquarie downgraded Broadcom to "neutral" immediately after the report.
A second fact missing from news feeds: the 2027 AI revenue forecast of $100+ billion assumes all six clients will simultaneously double their capacity. But supply chain sources report issues with CoWoS packaging availability at TSMC. Broadcom, according to HSBC, has reserved 260,000 wafers for 2026 and 480,000 for 2027. That's enough for growth, but any delay at TSMC or Samsung would hit the forecast.
Third—hidden margin pressure. In Q3, gross margin will fall to 74% from the current 77%+. This is not a one-off but a trend. The more AI chips in the sales mix, the lower the overall margin, because VMware software has margins around 80-85%, while physical silicon is around 60-65%. The market values Broadcom as a software company with a 35x PE multiple. But if in two years AI semiconductors make up 70% of revenue, the multiple should compress to 20-25x. The market has not priced in this compression yet.
| Risk | Probability | Potential Impact on AVGO |
|---|---|---|
| Google moving to in-house ASIC | 40% | -15-20% to 2028 revenue |
| P/E multiple compression | 60% | -25% to price at same EPS |
| CoWoS issues at TSMC | 30% | -10% to 2027 forecast |
| Slowing hyperscaler capex growth | 25% | -5-7% to quarterly metrics |
Forecast: Next 30 Days and 90 Days
The next 30 days will be driven not by fundamentals (already known) but by technical factors and capital flows. Broadcom is historically volatile between reports, and the current gap between price and the 50-day moving average is nearly 12%—a sign of overbought conditions. I expect consolidation in the $1600–$1720 range over the next month. Reason: many institutional investors who missed the rally from $1320 won't buy at all-time highs, and burned short sellers won't dare open new positions.
Over a 90-day horizon, the key catalyst will be the September Q3 report. If actual AI revenue comes in closer to $16.5-17 billion than $16.0 billion, shares could rewrite all-time highs around $1900-2000. If the company prints a number exactly at the low end of guidance or below, a correction will begin because the market has already priced in outperformance. The second scenario is more likely: Hock Tan is known for conservative forecasting and tends to lowball then beat.
| Period | Pessimistic Scenario | Base Scenario | Optimistic Scenario |
|---|---|---|---|
| 30 days | $1550–$1620 | $1600–$1720 | $1680–$1780 |
| 90 days | $1450–$1580 | $1600–$1800 | $1800–$2000 |
| Key factor | P/E compression to 28x | Stable P/E 32x | P/E expansion to 36x |
| Probability | 25% | 55% | 20% |
Editorial Forecast
Broadcom (AVGO) shares will continue a moderate rise over the next 24-72 hours, reaching resistance at $1720–$1750, after which a technical correction of 3-5% is likely. The key milestone—breaking the all-time high of $1680—has already occurred; the next target is $1740 (Fibonacci extension 1.618 from the June 4 low). Confidence level is moderate (55%) because the market is overbought and trading volumes are declining. The main risk to the forecast is sudden news of new US export restrictions on chip shipments to China, which would affect some Broadcom clients. This is an editorial opinion, not investment advice.
— Editorial Team
