Broadcom’s quarterly print came in broadly as expected on revenue of $22.19bn versus a range of about $22.13bn–$22.27bn, while adjusted EPS of $2.44 topped roughly $2.39–$2.40. AI revenue was $10.8bn against a $10.7bn guide, and the company guided Q3 revenue to $29.4bn versus around $28.3bn. The pressure point was Q3 AI revenue guidance of $16.0bn, below some “whisper” expectations near $16.36bn. Technically, the shares are near the top of a rising weekly channel, with RSI divergence flagged, yet the structure remains supported by the 50-week EMA band.
The AI trade is diverging between chip and infrastructure “sellers” and large-cap “buyers”. While Micron, TSMC, Marvell, AMD, SMH and SOXX are near highs, Nvidia and Broadcom have cooled, as Meta moves sideways and Alphabet and Amazon weaken from May peaks. MAGS is testing a potential double-top neckline near 68; a break would put 65 in view, although it is still riding its 50-day EMA. Separately, BTIG data show the S5INFT stretch signal triggered 10 times from 1995–2024, with average forward S&P 500 returns of -1.18% (5 days), -1.54% (10), -1.04% (20), -1.54% (30) and 0.32% (40), alongside an average RSI of 79.14 and a 31.58% premium to the 200-day moving average. After a nine-week rally beyond 7,300, Fibonacci markers sit near 7,395 and 7,650, with 7,300 and 7,000 framed as downside reference levels.
Shifts in Market Sentiment and Strategy Implications
We are seeing that the market’s mood on AI is shifting, where strong results are no longer enough. Broadcom’s sell-off shows that investors now demand perfection and blockbuster guidance. This suggests selling out-of-the-money call spreads on the most extended AI names might be a prudent strategy for the weeks ahead.
The AI trade is clearly splitting into two camps: the sellers of AI infrastructure and the buyers using it. This divergence means we should be cautious about broad index plays on the semiconductor sector and instead focus on individual names. The easy money from buying any stock with an AI story appears to be over.
On the seller side, implied volatility for downside puts on names like Broadcom and Nvidia has ticked up, with the 25-delta skew moving from -12% to -15% over the last week. While the charts are not broken, the rising cost of puts relative to calls signals that larger players may be bracing for a pullback. We see this as a warning that momentum is weakening even among the leaders.
AI Buyer Fatigue, Big Tech Concerns, and Broader Market Risks
The more significant weakness is appearing among the big AI spenders like Alphabet and Amazon, where questions about monetization are growing. Recent analyst reports show a consensus that Big Tech’s AI-related capital expenditures will grow by 45% this year, while related revenue is only projected to increase by 18%. This widening gap makes us consider long put positions on select names in this group.
The MAGS ETF is our main indicator for this AI buyer fatigue, as it is now testing the critical 68 support level. A decisive break below this level on daily closing volume would be a signal for us to add to bearish positions, perhaps by buying July puts. For now, the level is holding, but the repeated tests show underlying weakness.
This is all happening while the broader S&P 500 looks historically overextended. Last Friday’s Non-Farm Payrolls report came in hotter than expected, raising concerns the Fed might delay its anticipated rate cut. The VIX, while still low at 14.5, has crept up from its May lows near 12, suggesting some traders are quietly buying protection.
Historically, when the S5INFT technical signal has flashed, the S&P 500 has seen average negative returns over the following month. This historical precedent makes buying short-term index puts or VIX call options an attractive way to hedge long portfolios. A pullback toward the 7,300 level seems increasingly plausible in the coming weeks.
The key is to become more selective rather than abandoning the AI theme entirely. We will be watching the MAGS 68 level closely, as its failure would confirm that the fatigue among AI buyers is real. If that support breaks, it would reinforce our view that a broader market consolidation is likely.
