Semiconductors: The June 23 Shock and the Great AI Market Reckoning
How a single Broadcom guidance miss erased $1.3 trillion in two weeks
- On June 23, 2026, a below-consensus Broadcom guidance triggered a 7.9% drop in the Philadelphia Semiconductor Index and a 3.3% decline in the Nasdaq 100
- More than $1.3 trillion in semiconductor market capitalisation was erased over two weeks
- The fundamental question: the four major hyperscalers (Amazon, Google, Meta, Microsoft) have committed $750 billion in AI capex for 2026 — revenues must now follow
- For European investors, technology concentration risk in passive ETFs warrants reassessment ahead of July earnings season
- Major tech earnings (Amazon, Google, Meta, Microsoft, AMD) from July 22 onwards will be the next decisive test
In the space of a single session, markets delivered a reminder that even the most powerful trends can falter when actual results fall short of expectations. On June 23, 2026, a Broadcom guidance deemed insufficient by analysts triggered a cascade of selling across the entire semiconductor ecosystem, wiping out more than $1.3 trillion in market capitalisation over two weeks. This is not the end of artificial intelligence as an investment theme — but it marks the moment when markets began demanding proof of profitability, not just promises.
The Trigger: A Broadcom Guidance That Shook the Sector
Broadcom, a leading provider of AI networking chips and custom processors, published its quarterly results on June 23 with a Q3 AI revenue guidance of $16 billion. While representing 44% year-on-year growth, the figure fell 7% below analyst consensus expectations of $17.2 billion. This apparent “miss” was sufficient to trigger a chain reaction across the entire sector.
Within hours, the entire sector turned sharply negative. Micron Technology fell 13.2% during the session, Sandisk dropped 11.2%, Qualcomm lost 8%, AMD declined 5.8% and Nvidia retreated 4.2%. The Philadelphia Semiconductor Index — the global benchmark for the sector — fell 7.9%, one of its worst single sessions since 2022. The Nasdaq 100 declined 3.3% and even the S&P 500 ended 1.4% lower, in what constitutes an unprecedented warning about the fragility of the technology rally.
This move is part of a broader reassessment: in the two preceding weeks, more than $1.3 trillion in semiconductor market capitalisation had already been erased as investors began questioning the sustainability of the AI investment cycle. June 23 transformed that underlying anxiety into an outright correction.
The AI Equation: $750 Billion in Capex — For What Return?
The real question raised by this market shock is not cyclical. It is structural. In 2026, the four major American hyperscalers — Amazon, Google, Meta and Microsoft — collectively committed close to $750 billion in infrastructure spending: data centres, AI chips, fibre networks, cooling systems. This represents an unprecedented level of investment in the history of the technology sector.
For more than a year, markets accepted this scale-up without demanding immediate returns. The prevailing thesis was straightforward: whoever builds the most AI infrastructure today will dominate the market of tomorrow. Nvidia was valued at $5.2 trillion in May 2026. Every quarterly beat reinforced the narrative.
But Broadcom’s $16 billion guidance — up 44% year-on-year — was sufficient to break the spell, precisely because it fell below expectations. This gap reveals a growing tension: hyperscaler demand for AI chips is growing well, but not as fast as analyst models anticipated. The market is now beginning to wonder whether the $750 billion in capex is generating temporary overcapacity, much as the fibre optic buildout did at the turn of the 2000s.
“Markets tolerate infrastructure spending for a long time. Until the day they demand a return on investment. That day has arrived for artificial intelligence.”Morgan Stanley Research — Technology Sector Note, June 2026 (adapted)
What This Means for Your Portfolio: The Concentration Trap
For a European investor holding a MSCI World ETF, the June 23 correction is not a distant event. Information technology now represents approximately 25% of the MSCI World index, with a significant share concentrated in five to seven American companies. In practical terms, a 10% correction in the Nasdaq 100 translates to a loss of approximately 2-3% on a globally diversified portfolio — and a full 10% loss on a pure Nasdaq 100 ETF.
This concentration is not necessarily a problem if it is conscious and calibrated to your risk profile. It becomes one if it results from passive accumulation, through automatic index rebalancing, without an explicit allocation decision. The question to ask today is this: what is your effective portfolio exposure to AI semiconductors and hyperscalers, and is it consistent with your risk tolerance?
It is important to note that the S&P 500 retains a positive YTD performance of approximately 10% as of this shock. The June 23 correction does not invalidate the long-term AI trend — it places it in a more sober perspective. Those with balanced exposure across technology, European value, fixed income and real assets absorbed the session significantly better.
July: The Month of Truth for Hyperscalers
The weeks ahead will constitute a decisive test. Major technology earnings are expected from July 22 onwards: AMD and Qualcomm open the season, followed by Microsoft (July 24), Alphabet and Meta (late July), then Amazon and Apple. These publications will answer one question: do hyperscaler AI revenues justify the scale of their investments?
Three indicators deserve close attention. First, cloud AI revenue growth at Microsoft Azure, Google Cloud and AWS — any slowdown below 25% annualised growth would send an alert signal. Second, second-half capex guidance: if any hyperscaler reduces its investment programme, this would amplify pressure on equipment makers. Third, the Philadelphia Semiconductor Index trajectory following each publication. This index remains the most direct barometer of market sentiment on the AI cycle.
The monetary context adds a further layer of complexity. Kevin Warsh, at the helm of the Fed since January 2026, has adopted a deliberately hawkish tone. Markets now price close to 70% probability of a rate hike in September. Higher policy rates raise the cost of capital, which mechanically compresses the valuations of high-growth companies with long-duration cash flows — precisely the category of semiconductors and hyperscalers.
- The June 23 correction is not the end of AI as an investment theme: it marks the transition from a “growth at any price” regime to a “show me the return on investment” regime
- $1.3 trillion wiped from market cap in two weeks illustrates the concentration risk that semiconductors can represent in an unmonitored passive portfolio
- The key question for your investments: what is your effective exposure to the 5-7 major AI technology names, and is it calibrated to your risk profile?
- The hawkish Fed context (higher policy rates potentially in September) adds structural pressure on long-duration technology valuations
- July is the test month: Amazon, Google, Meta and Microsoft earnings will determine whether the correction is temporary or the beginning of a deeper rebalancing
This document is provided for informational purposes only and does not constitute investment advice, a personalised recommendation, or an offer to buy or sell financial products. Past performance is not indicative of future results. All investments carry risk, including the risk of capital loss. The information contained in this article reflects the analysis of Riviera Wealth Management as of the date of publication and is subject to change. Riviera Wealth Management is an independent financial investment adviser (CIF), registered with ORIAS and a member of the CNCGP.
