1. Repeated Middle East geopolitical clashes; global market sentiment swings
The U.S.-Iran conflict in the Middle East has kept intensifying. Since a temporary ceasefire was reached in early April, the region has seen the largest armed clashes since the latest round of fighting began. On June 2, U.S. forces struck an oil tanker tied to Iranian capital, saying it had illegally transported crude oil and violated U.S. sanctions laws.
This immediately triggered a retaliatory response from Iran, which used ballistic missiles and suicide drone swarms to attack U.S. bases in Kuwait and Bahrain. The strike on Kuwait International Airport caused casualties, and tensions in the Persian Gulf surged at once.
Just as the market feared a full escalation, Israel and Lebanon signed a temporary ceasefire agreement. Hezbollah withdrew from southern Lebanon and both sides stopped cross-border attacks. Iran's channel for expanding the war through proxies was cut off, and Middle East geopolitical risk cooled quickly.
At the same time, a new twist emerged in U.S. politics: the House passed a bill limiting the president's unilateral authority to use force against Iran. Domestic political maneuvering in the U.S. also makes the outlook for U.S.-Iran talks highly uncertain, continuing to unsettle global risk assets.
2. Strong U.S. jobs resilience cools Fed rate-cut expectations sharply
The U.S. Labor Department released the latest weekly unemployment data. Initial jobless claims rose to 225,000, the highest in nearly three months, which the market initially read as a sign of labor-market weakening. But continuing claims fell to 1.77 million, hovering near the 1.8 million mark and remaining in the low range of the past two years.
Overall, the U.S. labor market has not shown any material slowdown. Job data proved more resilient than expected, directly narrowing the Fed's room to cut rates this year. Cooling rate-cut expectations sparked a broad pullback in high-flying tech stocks, and Broadcom, right after its earnings release, became the main target of concentrated selling.
3. Broadcom's 15% plunge: emotional overreaction or market top?
3.1 Fundamentals improve across the board; growth leads global semiconductors
After earnings were released, Broadcom's intraday drop reached nearly 15%, making it the core name in this round of U.S. tech-sector correction. The market broadly interpreted the move as Broadcom's AI chip business falling short of expectations and the AI trade peaking, but a closer look at the numbers shows the company's fundamentals are far above the industry average.
- Revenue and profit: Second-quarter revenue reached $22.19 billion, up 48% year over year, setting a nine-year high for quarterly growth. Non-GAAP EPS was $2.44, up 54% year over year; both figures beat expectations.
- AI core business: AI-chip-related revenue reached $10.8 billion, up 143% year over year, accounting for nearly 70% of semiconductor revenue and serving as the first growth engine.
- Cash-flow business: Software revenue was $7.18 billion, maintaining steady growth and providing a solid cash-flow base.
- Order backlog: Total backlog topped $30 billion, order visibility extended to the end of 2028, and capacity for the next two years has already been fully locked in by top global customers, keeping supply and demand tight.
On hard operating metrics, this earnings report puts Broadcom in the top tier of U.S. tech earnings season and is by no means negative news.
3.2 Core trigger for the plunge: expectations were badly overextended, and a small miss sparked a stampede
Broadcom's stock had already enjoyed a long one-way run, and expectations were pushed to the extreme. That is the core reason for this selloff:
- Third-quarter revenue guidance was $29.4 billion, up 84% year over year and leading the global chip industry, but the market had already hyped expectations close to $30 billion, so a small gap was magnified into a miss;
- Full-year AI revenue guidance was $56 billion, more than 180% higher year over year, slightly below the $57.6 billion target cited by some investment banks;
- Third-quarter AI semiconductor revenue guidance was $16 billion, $200 million below the market average, giving short-term funds a chance to amplify bearish claims such as an AI demand inflection and major-customer order cuts.
With expectations running so high, ordinary guidance differences were enough to trigger concentrated profit-taking and a selloff.
3.3 Google's supply-chain diversification is overstated; customer mix keeps improving
Another major source of panic was Broadcom's CEO confirming that top customer Google is pursuing supply-chain diversification for chips. The market worried that Google would cut orders and lose a core customer. But when you factor in long-term contracts and the industry backdrop, this bearish view is badly overstated:
Google and Broadcom have a legally binding multi-year TPU procurement agreement that locks in more than 10 GW of computing capacity, with contract value reaching tens of billions of dollars. Global AI compute demand is growing more than 200% a year, and Google's in-house chip capacity cannot cover its own needs. Diversification simply adds backup suppliers; it will not reduce Broadcom's total orders.
At the same time, Broadcom continues to deepen ties with OpenAI, Meta, Amazon, and other top global clients. Meta and OpenAI have both locked in long-term compute orders, and the company has also partnered with institutions to set up a $35 billion AI compute fund. Dependence on any single customer keeps falling, and business risk is further diversified.
3.4 Gross margin swings are structural; profitability keeps strengthening
The slight quarter-over-quarter decline in non-GAAP gross margin was interpreted by bears as a deterioration in profitability. In reality, the reason is that higher-value AI-chip revenue is taking a larger share. This category requires higher foundry spending and fixed-asset investment in the early stages, temporarily diluting overall gross margin.
As capacity scales up, gross margin will gradually recover. The company's non-GAAP operating profit has now reached a record high, revenue growth is far outpacing expense growth, scale effects keep being released, and its true earnings power keeps improving.
Several top investment banks have raised Broadcom's target price against the trend. Deutsche Bank raised its target from $430 to $515, while Citi kept its $500 target and Buy rating, clearly stating that the current drop is a non-rational sentiment-driven false selloff and that long-term positioning value is clear.
Overall, Broadcom's plunge was driven by front-loaded expectations, short-term selling pressure, and amplified negative rumors. Revenue, orders, customers, and long-term contracts are all improving. This is not a market top; after the short-term sentiment washout, valuation is likely to recover.
4. CrowdStrike earnings: good news priced in, triggering a sell-the-news move
Also in the tech sector, cybersecurity leader CrowdStrike posted strong earnings but still fell more than 10% premarket, delivering a classic sell-the-news move. Its logic is completely different from Broadcom's: this is a reasonable pullback caused by a mismatch between valuation bubbles and earnings growth.
The company's core earnings figures were strong: revenue was $1.386 billion, up 25.7% year over year; non-GAAP net income turned profitable; annual recurring revenue rose 24% year over year; the company also raised full-year revenue guidance and announced a 4-for-1 stock split. Despite all these positives, the stock fell instead of rising.
4.1 Stock doubles in six months; valuation already prices in future growth
Year to date, CrowdStrike's stock has risen more than 64%, nearly doubling in six months. The market has already fully priced the long-term upside of AI-empowered cybersecurity into today's share price, creating a severe valuation bubble.
The company's 25.7% revenue growth is excellent for the SaaS industry, but it cannot justify a stock price that nearly doubled. Market sentiment has swung from extreme pessimism that AI would replace traditional cybersecurity to extreme enthusiasm that AI would double cybersecurity results, pushing expectations far away from fundamentals.
AI will indeed expand the cybersecurity market, but industry adoption will happen gradually. Overall industry growth is concentrated in the 15%-20% range, far below the high-growth fantasy the market imagined. The sector has now entered a bubble-deflation phase, and leader CrowdStrike is the first to correct.
4.2 Hidden report flaws intensified profit-taking
The earnings details also contain several hidden issues: on a GAAP basis the company is still loss-making, stock-based compensation equals 22.9% of total revenue, and high equity dilution is eroding shareholder returns. Sales and marketing expenses account for 35.3%, so the high-spend-for-growth model has not changed. In a high-valuation environment, bears magnify these details and use them as an additional reason for funds to leave.
In the near term, the cybersecurity sector is likely to enter a volatile valuation digestion phase. Only continued earnings beats and steadily improving growth can absorb the elevated valuation and restart the rally.
5. Conclusion: short term follows sentiment; long term anchors on fundamentals
This round of divergence in U.S. tech stocks clearly confirms a core market rule: in the short term, stock prices are driven by expectations and sentiment; in the long term, they always depend on real fundamentals.
Middle East conflict swings and cooling rate-cut expectations shape the external market backdrop, while the divergence in individual stocks comes down to the fit between valuation and earnings. Broadcom is a quality company that was punished by sentiment; its short-term drop creates a buying window. CrowdStrike, by contrast, is facing a valuation bubble that is too large for its earnings to support, so it has entered a rational correction phase.
For investors, no matter how strong the theme or how impressive the earnings, once a stock price has already priced in several years of future growth, valuation mean reversion will eventually arrive. On the other hand, when quality companies are sold in panic, that is often the best time to build positions rationally. U.S. sector rotation is still at an early stage, and AI-related stocks differ sharply in valuation, expectations, and fundamentals, so investors need to distinguish between names and judge rationally.
