Broadcom's Earnings Report and AI Capital Expenditure Cycle: Structural Analysis of Revaluation in the Chip Sector and the Stage-wise Pressure on Crypto Assets

After the market close on June 3rd, Broadcom (AVGO) delivered a financial report that can only be described as "blowout"—revenue increased 48% year-over-year to $22.2 billion, AI semiconductor revenue surged 143% YoY to $10.8 billion, both exceeding market expectations. However, the next day, the stock price plummeted over 13%, with a nearly 20% decline over two trading days. Meanwhile, Bitcoin fell back toward the $60,000 mark, while the NASDAQ and S&P hit new highs repeatedly.

This is not unusual. It is a re-pricing of "expectations" and a clear signal of institutional capital flow—capital is flowing out of crypto assets and into AI infrastructure sectors. Understanding AVGO’s "performance dip" may be the most important entry point to grasp the underlying logic of the current market.

All earnings exceeded expectations, so why did "no upward revision" become the final straw crushing the stock?

First, look at the fundamentals of Broadcom’s Q2 performance. According to the company’s financial report and multiple media sources, Broadcom’s Q2 revenue for fiscal 2026 reached a record $22.2 billion, up 48% YoY; non-GAAP diluted EPS was $2.44, surpassing the consensus estimate of $2.40. AI semiconductor revenue hit $10.8 billion, a 143% increase YoY; total bookings exceeded $30 billion, extending order visibility from "2027" to "2028" compared to three months prior. Free cash flow also hit a new high of $10.3 billion.

On any individual metric, this is a financial report that would make most companies proud. But in Broadcom’s case, the market’s reaction logic is entirely different.

The real negative signals come from two "no upward revision" points:

First, the Q3 AI semiconductor revenue guidance did not meet the market’s top expectations. Broadcom forecasted Q3 AI semiconductor revenue of $16 billion, a growth of over 200%, but this was below the analyst consensus of about $17.2 billion. Some analysts pointed out that Citigroup’s Atif Malik originally expected $17.5 billion, and some buy-side models are even more aggressive. In a context where the market has already priced in extreme growth expectations, $16 billion, though still very strong, is not enough to meet higher psychological thresholds.

Second, the long-term AI revenue target was not revised upward. Broadcom reaffirmed its outlook of "over $100 billion" in AI chip revenue for fiscal 2027, consistent with the previous quarter, rather than a significant upward revision as the market hoped. Additionally, the full-year AI chip revenue estimate of about $56 billion for fiscal 2026 is below some institutional estimates of $57.6 billion. Against the backdrop of competitors like Marvell having already raised guidance, the market perceives Broadcom’s "conservative stance" as a clear psychological disappointment.

Essentially, this is a "expectation management" game. When the stock price had already risen over 15% in the five trading days before the earnings release due to optimistic AI sentiment, adding about $300 billion in market cap, a report that merely beats expectations was no longer enough.

The "Domino Effect" of AI Semiconductor: From AVGO’s Crash to $1.3 Trillion in Market Cap Vanishing in Two Days

Broadcom’s sell-off was not an isolated event. As one of the most important barometers in AI compute chips, AVGO’s sharp decline quickly propagated through the entire industry chain.

On June 4th, Broadcom’s stock plunged about 13%, the largest single-day drop in over 16 months, with a market cap evaporating approximately $286 billion in one day. The next day (June 5th), it continued to fall about 7.9%, with a total decline of nearly 20% over two days.

The ripple effects spread rapidly:

• NVIDIA (NVDA): down about 6%, with a market cap loss of over $300 billion in one day;
• Micron Technology (MU): down 13%, with a market cap loss of about $150 billion;
• Marvell Technology (MRVL): down about 17%;
• AMD and Intel (INTC): both down about 11%;
• Optical communications sector: Ciena down nearly 16%, Lumentum down nearly 6%, Coherent also experienced significant pullback.

Across the entire sector, the Philadelphia Semiconductor Index (SOX) plunged 10.26% on June 5th, the largest single-day drop since the COVID-19 outbreak in March 2020. According to Zhitong Finance data, the chip sector lost about $1.3 trillion in market value over two trading days.

A notable difference: not all tech stocks were wiped out. Google rose against the trend by 3.82%, TSMC ADR increased 1.88%, and NVIDIA rebounded to close up about 1.8%. This divergence signals two things:

The long-term demand logic for AI infrastructure remains intact—TSMC’s position as the "sole node" in manufacturing provides some defensive attributes; the market’s "punishment" is targeted rather than unconditional or across the entire sector: stocks benefiting from AI infrastructure that are not excessively overvalued have not been abandoned by capital.

Meanwhile, overall U.S. stock funds did not withdraw from the market but shifted from semiconductor stocks to financial and consumer sector components—on June 5th, the Dow Jones Industrial Average rose 874 points, hitting a new all-time high.

The True Logic Behind Institutional "Selective De-risking" of Crypto Assets

If the sharp decline in chip stocks can be attributed to a valuation and expectation mismatch, the Bitcoin decline during the same period points to a deeper structural change.

From June 2 to June 5, Bitcoin’s price fell from about $72,000 to $59,895, briefly breaking below the $60,000 level, hitting the lowest since October 2024.

The decoupling from traditional risk assets is intensifying. During the same period, the Nasdaq 100 hit a new all-time high, with a 12-month cumulative increase of about 41.5%, while Bitcoin fell approximately 37% YoY, still about 48% below its last year’s peak. This stark divergence is rare historically. K33 Research analysts explicitly stated in a report: "Bitcoin’s weakness reflects continued cooling of institutional demand, as many market participants see holding BTC as too costly in opportunity terms, while assets related to AI continue to soar."

The scale of capital outflow from crypto can be validated by ETF data:

• Since mid-May, US spot Bitcoin ETFs have net outflows of about $4 billion, with 17 out of 19 trading days recording net outflows;
• By early June, the last three weeks saw a total outflow of 62,794 BTC, the second-largest recorded outflow;
• Since mid to late May, the daily outflow from BlackRock’s IBIT reached as high as $528 million;
• CME Bitcoin futures open interest has fallen to its lowest level since October 2023—often a sign of waning institutional participation.

Simultaneously, large capital flows are heading into AI infrastructure investments. MicroStrategy founder Michael Saylor explicitly stated on X: "In the past six months, about $400 billion has been invested in AI projects. This is not a devaluation of Bitcoin but a capital rotation." Wintermute’s latest market analysis report in early June also pointed out that the correlation between crypto markets and US stocks is experiencing its most severe break this year, with significant capital outflows from cryptocurrencies, while AI semiconductor sectors are becoming the new institutional inflow destination.

Curve founder Michael Egorov’s recent analysis at Gate Square echoed this view, suggesting that the core factor is not deterioration in crypto fundamentals but a phase shift in capital preferences—by 2026, AI stocks have become the main market theme, aligning with multiple independent market studies.

An often-overlooked piece of evidence: during the same week Bitcoin continued to decline, US May non-farm payrolls added 172k jobs, far exceeding expectations, reinforcing the possibility of the Fed maintaining high interest rates or even raising again. Interest rate futures show the market’s probability of a December rate hike rising from 48% to 63%. For crypto assets, further tightening of the rate environment means rising holding and opportunity costs, providing macro-level support for institutional de-risking.

Inside Out: Wall Street’s "Reverse Voting"

Contrary to the intense market sentiment swings, major Wall Street institutions almost uniformly raised their target prices and reaffirmed buy ratings after Broadcom’s plunge.

• Goldman Sachs: raised target price from $500 to $525, maintained "Buy," and reaffirmed Broadcom on its "Conviction List," expecting nearly 30% upside from the pre-market around $408;
• Exane (BNP Paribas): raised target from $600 to $640, maintained "Outperform," calling the post-earnings decline "shortsighted";
• Morgan Stanley: raised target from $470 to $485;
• TipRanks average target price: $512.88 based on 24 "Buy" and 3 "Hold" ratings, implying about 28.45% upside.

The consensus logic on Wall Street is: the fundamental reason for this sell-off is the market overestimating AVGO’s valuation—by the earnings report day, the P/E ratio had exceeded 90, leaving little room for error. When the company issued guidance conservatively, the market took profits rather than doubting the long-term AI story. Jefferies analyst Blayne Curtis attributed the decline to the market "pricing in huge AI growth signals and expecting further upward revisions," but when Broadcom simply reaffirmed rather than beat expectations, valuations were re-priced.

Using Capital Flows to Anticipate the Future: A Two-Track Framework and Key Indicators

The core variable now is no longer "whether the AI boom will continue," but the rebalancing of capital among cryptocurrencies, AI infrastructure, and macro sectors. Here are three logical frameworks for market participants to build their judgment:

First, how long will the "rotation pressure" in crypto assets last?

Current capital outflows are a structural signal, not just short-term sentiment. Capital expenditure on AI chips is at a historic expansion cycle. Goldman Sachs data shows Broadcom has six custom chip clients (including Google, Meta, Anthropic, OpenAI), with multiple gigawatt-scale data center deployments scheduled over the next two years. Given this scale of industry expansion, the likelihood of crypto assets regaining dominance in "high-growth narratives" in the short term is low. However, large institutions are also conducting OTC positions off-exchange, indicating that long-term holders have not fully exited.

Second, how will the "valuation reset" of AVGO unfold?

The current 14% decline corresponds to a reversion from a near-100 P/E to a more reasonable range, not a disproof of AI fundamentals. Broadcom’s management has extended AI order visibility to 2028—an important hard indicator in any market cycle. Therefore, the key variable is not whether Broadcom "still grows," but how much growth premium the market is willing to assign. If guidance in Q3 materially raises outlooks, the current gap could become a starting point for re-pricing.

Third, how will the correlation between Bitcoin and the AI sector evolve?

Currently, they are in a phase of "decoupling," contrasting with Bitcoin’s long-standing role as a "tech risk appetite indicator." But if institutional returns on AI infrastructure investments underperform (e.g., profit realization cycles are delayed after large capex), capital may reassess crypto as an alternative asset class. When that happens, the divergence in their current trajectories could serve as a "pre-signal" for the next narrative shift.

Conclusion

The case of Broadcom’s Q2 earnings reveals a recurring pattern in the 2026 tech asset pricing: when valuations inflate to the "perfect expectation" stage, the market’s desire for "upward revisions" surpasses the absolute quality of the earnings themselves. Beating expectations is no longer enough to lift the stock—what the market craves is "super-expectations" that explode.

AVGO’s decline is not a turning point for AI investment logic. Broadcom’s AI orders are scheduled through 2028, and its six major clients’ data center expansion plans are proceeding as planned. What is truly being re-priced is the valuation level that was once driven to over 90 P/E by extreme optimism—this is a rational valuation correction, not a wholesale negation of AI infrastructure investment logic.

Meanwhile, the $4 billion ETF capital outflow and Bitcoin’s drop below $60,000 should not be interpreted in isolation. The internal consistency of these phenomena points to a common narrative: capital is flowing from crypto markets into the AI infrastructure supply chain, exhibiting systemic institutional characteristics.

For investors, distinguishing between "structural decline" and "phase rotation" of capital is crucial at this stage. Clarifying which declines are valuation reversion and which reflect deeper industry shifts may be the key to navigating current market divergences.