What You'll Learn
- Why Broadcom's record AI revenue triggered a $300 billion selloff
- How the "priced for perfection" AI trade is reshaping earnings season
- Which semiconductor stocks got caught in the contagion
- How to position your portfolio for the new AI guidance arbitrage game
The Paradox That Defined AI Investing in June 2026
On June 3, 2026, Broadcom delivered what any reasonable observer would call a spectacular quarter. Revenue hit $22.19 billion, up 48 percent from a year earlier. AI semiconductor revenue exploded 143 percent to $10.8 billion. Free cash flow reached $10.3 billion, an eye-popping 46 percent of revenue. By every fundamental metric, the company firing on all cylinders.
Then the stock fell 14 percent, wiping out roughly $300 billion in market capitalization in a single session. The culprit was not a miss. It was not even a slowdown. It was a guidance number that met every Wall Street target except the one that mattered most: the unofficial, unwritten whisper number that traders had baked into a stock that had already rallied 93 percent over the preceding twelve months.
This is the story of that paradox. For investors in the United States, Canada, and Australia who have ridden the AI wave, it is a story that demands attention. What happened to Broadcom on June 4 was not an anomaly. It was a warning about how the AI trade works now, and how it may work for the rest of 2026.
The Quarter in Numbers: A Record That Should Have Sparked a Rally
Broadcom's fiscal second quarter, ended May 3, 2026, was the strongest in the company's history. Revenue of $22.19 billion comfortably surpassed the analyst consensus of $22.13 billion, according to data compiled by Bloomberg. Adjusted earnings per share of $2.44 beat the $2.40 estimate. These were not marginal beats. They were clean, unambiguous wins across both the top and bottom lines.
The headline number that dominated every analyst call was AI semiconductor revenue of $10.8 billion, up 143 percent year-over-year. CEO Hock Tan, on the post-earnings call, described demand for custom AI accelerators and AI networking products as "continuing to accelerate." The company's custom chip partnerships with hyperscalers including Google, Amazon, and Anthropic are scaling faster than even Broadcom's own internal forecasts had projected.
Free cash flow reached $10.3 billion, generating 46 percent of revenue. For context, that cash conversion ratio places Broadcom in rarefied air among semiconductor companies. CFO Kirsten Spears noted that cash from operations hit $10.5 billion against capital expenditures of just $231 million, reflecting the asset-light nature of Broadcom's design-and-license chip model.
The infrastructure software segment, anchored by the $69 billion VMware acquisition completed in 2023, contributed $7.18 billion in revenue, up 9 percent year-over-year. While the growth rate appears modest compared to the AI semiconductor business, the segment carries a remarkable 93 percent gross margin and an operating margin close to 79 percent, making it a cash engine that funds the AI expansion.
The non-AI semiconductor business presented a more mixed picture. Broadcom's legacy chip sales declined modestly, as the company continues to pivot resources toward the AI opportunity. But the overall semiconductor segment still grew 48 percent, entirely thanks to the AI explosion. As we covered in our Broadcom Q2 FY2026 earnings preview, the AI backlog stood at $73 billion entering the quarter. That number has only grown since.
The Guidance That Broke the Stock's Back
If the quarter was so strong, why did the stock collapse? The answer lies in a single number: Broadcom's Q3 AI chip revenue guidance of approximately $16 billion. On its face, that number implies year-over-year growth in excess of 200 percent for the AI segment in the third quarter. By any historical standard, it is an extraordinary projection.
But the market had already priced in something closer to $17 billion or higher. The whisper number, built from pre-earnings analyst conversations, buy-side expectations, and the stock's parabolic 93 percent twelve-month rally, was simply higher than what Broadcom delivered. The company also declined to raise its long-term $100 billion AI revenue target, a target Hock Tan had set previously. Investors wanted a raise. They did not get one.
This phenomenon is known as "priced for perfection." When a stock trades at 42 times forward earnings and has already priced in multiple quarters of acceleration, beating expectations is not enough. You must crush them, and then raise the bar again. Broadcom crushed the quarter. It did not crush the future outlook.
The contrast with Broadcom's own delivered performance is instructive. The company had previously guided for Q2 AI revenue of roughly $10.7 billion. It delivered $10.8 billion. The Q3 guidance of $16 billion represents sequential acceleration from Q2's $10.8 billion, which itself was a massive step up. CEO Hock Tan even raised the 2027 AI chip shipment forecast to more than 10 gigawatts, up from prior guidance. None of it mattered in the moment, because the stock's valuation left no room for anything short of perfection.
Priced for Perfection: How the AI Trade Has Changed
Broadcom is not alone in experiencing this dynamic. The "beat and drop" pattern has become a recurring feature of AI earnings seasons in 2026. When Dell reported Q1 earnings with 40 percent revenue growth and a $5 billion AI backlog, the stock surged 30 percent. But when Meta posted a Q1 beat alongside a $145 billion AI spending plan, the stock fell 6 percent. When Salesforce beat Q1 estimates but issued cautious guidance, the stock dropped 4 percent even as the company announced a $25 billion buyback.
What changed? The market's valuation framework for AI stocks has shifted from discounting future cash flows to discounting future guidance acceleration. In 2024 and 2025, AI stocks rallied on earnings beats. In 2026, they rally only on guidance raises. The base case is already priced in. The marginal dollar of upside must come from raising the trajectory, not confirming it.
Broadcom's 63 percent surge in the two months before earnings set an impossible bar. As we noted in our analysis of the AI chip cost bubble, when a stock rises that far that fast ahead of a known catalyst, the asymmetry of outcomes skews sharply to the downside. The stock can only surprise on the upside by a few percent. It can disappoint by ten or fifteen.
The result is what one strategist described as "the guidance arbitrage game." Earnings reports are no longer about validating past performance. They are about projecting future acceleration, and then exceeding those projections. Broadcom delivered 143 percent AI growth. But it guided for "only" roughly 200 percent annualized in Q3. The market wanted 250 percent. That gap of 50 percentage points cost shareholders $300 billion in market value.
Contagion Across the Semiconductor Sector
Broadcom's selloff did not happen in isolation. The June 4 semiconductor rout spread across the sector as investors repriced AI exposure across the board. Micron Technology dropped 7 percent, despite reporting 57 percent revenue growth and HBM3E orders stretching into 2027. SanDisk fell 3 percent and Western Digital declined 2 percent, as Broadcom's underwhelming AI outlook spread contagion through the broader memory chip complex.
Marvell Technology, which had surged 25 percent just days earlier on Jensen Huang's "next trillion-dollar company" endorsement, gave back a portion of those gains. In our analysis of Marvell's post-Computex surge, we noted that the custom silicon war between Broadcom and Marvell was driving a "winner-take-most" dynamic. Broadcom's guidance miss did not change that competitive picture, but it did reset the valuation premium across the sector.
The Nasdaq Composite fell as AI-related tech names sold off, even as the Dow Jones Industrial Average managed to eke out gains on rotation into financials and healthcare. As we documented in our coverage of the Dow's record-breaking rotation day, the market is showing a split personality: AI stocks are being punished for incremental disappointment while value and defensive sectors attract inflows. This rotation, first identified in early June, has only accelerated post-Broadcom.
The contagion mechanism is straightforward. When the bellwether AI chip company outside of Nvidia guides below whisper expectations, every other AI-exposed stock gets repriced on the assumption that demand growth may be peaking. Never mind that Broadcom's Q3 AI guidance of $16 billion still implies more than doubling year-over-year. In a market priced for perfection, deceleration is treated as decline.
Broadcom Q2 FY2026: Key Financial Metrics
| Metric | Q2 FY2026 | YoY Change | vs. Consensus |
|---|---|---|---|
| Total Revenue | $22.19B | +48 percent | Beat ($22.13B) |
| AI Semiconductor Revenue | $10.80B | +143 percent | Above guidance |
| Infrastructure Software Revenue | $7.18B | +9 percent | In line |
| Adjusted EPS | $2.44 | +28 percent | Beat ($2.40) |
| Free Cash Flow | $10.26B | +52 percent | 46 percent of revenue |
| Cash from Operations | $10.49B | +48 percent | N/A |
| Capital Expenditures | $231M | +15 percent | Asset-light model |
| Q3 AI Guidance | ~$16.0B | ~+200 percent | Below whisper $17B+ |
Sources: Broadcom Q2 FY2026 press release (June 3, 2026), Bloomberg consensus estimates, company earnings call.
What Analysts Say Now: Buy the Dip or Catch a Falling Knife?
In the aftermath of the 14 percent selloff, analyst reactions have been notably divided. At least one prominent analyst raised his price target on Broadcom stock to $550 from $500, implying meaningful upside from the post-crash share price of approximately $417. The rationale: nothing fundamental changed. Broadcom's AI revenue is still growing at 143 percent, its backlog is still $73 billion-plus, and its hyperscaler partnerships are still expanding.
Other analysts have struck a more cautious tone, noting that Broadcom's forward P/E multiple, which had compressed from 42 times to roughly 35 times after the drop, still reflects premium pricing relative to historical semiconductor valuations. The stock had rallied 93 percent over the trailing twelve months before earnings, and the guidance reset may take several quarters to work through before the next leg higher.
The bull case rests on three pillars. First, AI demand is structural, not cyclical. Broadcom's custom chip partnerships with Google (TPU), Amazon (Trainium), and Meta (MTIA) are multi-year, multi-billion-dollar commitments that are not sensitive to quarterly whisper numbers. Second, the 2027 AI shipment forecast of 10+ GW implies a revenue trajectory that today's stock price does not fully reflect. Third, the VMware software business provides a resilient, high-margin revenue stream that insulates Broadcom from pure semiconductor cyclicality.
The bear case counters that valuation still matters. Even at 35 times forward earnings, Broadcom is priced for perfection. If AI spending growth decelerates from 143 percent to even 80 percent over the next four quarters, today's valuation would look stretched. The guidance dynamic is the key variable: as long as the market trades on guidance acceleration rather than earnings beats, Broadcom will need to consistently raise its forward projections to sustain its multiple.
For Canadian and Australian investors, the takeaway extends beyond a single stock. The TSX-listed technology names and ASX AI-exposed companies face the same repricing risk. When the global AI benchmark stumbles, local markets feel the ripple. The S&P/TSX Composite technology sector and Australia's WISD-Tech Index both saw modest declines in sympathy with the Nasdaq on June 4.
Canada, Australia, and the Global AI Investor
The Broadcom selloff carries specific implications for investors outside the United States. Canadian investors tracking the TSX-listed technology sector should note that AI infrastructure spending is a global phenomenon, and valuation compression in US AI stocks tends to cascade into Canadian tech names within days. The TSX's technology sector, which includes companies like Shopify, Constellation Software, and AI-focused GPU cloud providers, experienced sympathy selling on June 4.
For Australian investors, the picture is complicated by the dual exposure to both AI demand and commodity prices. Australia's ASX 200 technology index includes companies that supply data center infrastructure and rare earth materials critical to semiconductor fabrication. When Broadcom's guidance miss triggered a semiconductor selloff, ASX-listed names like WiseTech Global and NEXTDC saw valuation pressure. Meanwhile, the iron ore and lithium sectors, which have benefited from the AI buildout indirectly through energy demand, remained relatively insulated.
The regulatory dimension also differs by jurisdiction. While the SEC in the US oversees Broadcom's disclosures, Canadian investors rely on securities regulators who follow US market leads. The Australian Securities and Investments Commission (ASIC) has flagged AI-related market concentration as a risk factor in its 2026 market outlook. For investors in all three countries, the key lesson is that AI stock volatility has become a global transmission mechanism. A guidance miss in Palo Alto reverberates from Toronto to Sydney within trading hours.
What This Means for Your AI Portfolio
The Broadcom episode distills down to a single investing insight: the AI trade has entered a new phase. The phase where any AI company that merely beats earnings will rally is over. We are now in the phase where guidance trajectory determines winners and losers, and where valuation multiple expansion has largely played out.
For portfolio positioning, this suggests several practical rules. First, size positions relative to the guidance risk premium. As the Broadcom Q2 FY2026 press release shows, A stock trading at 42 times forward earnings needs to deliver continuous upward guidance revisions to justify its multiple. If you cannot confidently forecast that trajectory, reduce the position size accordingly. Second, diversify across the AI value chain. The companies that actually build AI data centers (Vertiv, Eaton, Quanta Services) have a different risk profile than the chip designers (Broadcom, Nvidia, Marvell). As we outlined in our AI infrastructure stocks guide, the infrastructure layer may offer a better risk-reward balance in this environment.
Third, watch the macro calendar. The Broadcom selloff occurred ahead of the June 5 nonfarm payrolls report and the June 16-17 FOMC meeting. As we covered in our FOMC preview, a hawkish Fed combined with AI sector volatility creates a dangerous cocktail for overvalued tech names. Fourth, do not confuse a guidance-driven selloff with a fundamental deterioration. Broadcom's business is stronger today than it was before the stock dropped. The question is not whether the company is good. It is whether the price was too high.
For Australian and Canadian investors, the playbook is the same but the implementation differs. Consider allocating a portion of AI exposure to local exchange-traded funds that track the global AI value chain rather than single stocks. This provides the growth exposure without the single-stock guidance risk that just erased $300 billion from Broadcom's market cap in a single day.
The Bottom Line
Broadcom's June 4 selloff was not a story of failure. It was a story of valuation catching up with narrative. A company that grew AI revenue 143 percent, posted $22.2 billion in revenue, and generated $10.3 billion in free cash flow saw its stock collapse because the market had already priced in that success and was demanding more. That is the state of AI investing in June 2026.
The lesson for every investor, whether in New York, Toronto, or Sydney, is that earnings beats are no longer catalysts for AI stocks. Guidance raises are the only currency that moves markets. And when a stock has rallied 93 percent in twelve months, the bar for what qualifies as a good enough guidance raise becomes almost impossibly high.
Broadcom remains one of the most strategically positioned companies in the AI ecosystem, with unmatched custom chip relationships with the world's largest hyperscalers. But the $300 billion lesson is that even the best company can be a bad investment at the wrong price. The AI trade is not broken. It has simply matured. And maturity demands a more disciplined approach to valuation, diversification, and risk management.