Since January 2025, US equity markets have been driven by a central conviction: generative artificial intelligence will transform the business models of major digital platforms and generate an unprecedented wave of productivity gains. The S&P 500 is up +14% year-to-date as of July 18, 2026, largely carried by the seven mega-cap technology stocks that lend their name to this group: Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta and Tesla.

Following the FOMC decision on July 17 — rates on hold, mildly restrictive bias maintained — and the semiconductor sector correction of 7.5% observed in mid-July following disappointing chip order data, market attention is now focused on GAFA publications this week. These results will be a real-world test of the AI thesis: are the billions invested in cloud infrastructure and GPU accelerators finally translating into concrete revenues and margin improvement?

The answer conditions not only the performance of these individual stocks, but the entire equity market trajectory for the remainder of the second half of 2026.

The week of July 21 concentrates the majority of Magnificent 7 publications. Each announcement carries its own catalyst and its own risks. Here is the analytical framework we use for each.

Tesla has already set the tone by reporting below-consensus results on July 16: deliveries in retreat, gross margin eroded by price cuts on US and Chinese markets. The stock lost 5.2% on the session. This warning illustrates a systemic risk specific to the Magnificent 7: their index weighting is such that each individual disappointment generates shockwaves across passively managed portfolios.

Beyond revenue and earnings-per-share figures, three indicators are being particularly closely monitored by institutional managers this season.

1. Cloud revenues (Azure, AWS, Google Cloud) — Cloud is the primary vehicle for monetizing AI for these platforms. Markets expect growth acceleration: Azure at +28%, AWS at +22%, Google Cloud at +26%. A figure more than 3 percentage points below expectations would likely trigger a significant correction in the relevant stock.

2. Capex and capex guidance — Cumulative investments by the five hyperscalers are set to exceed $340 billion in 2026. The question is no longer whether this investment is enormous — it is — but whether guidance for 2027 confirms the trajectory or whether executives begin to moderate spending. A slowdown in capex would be interpreted as a sign of AI cycle maturity — not necessarily negative if margins improve as a result.

3. Operating margins — The bulls’ central argument is that generative AI improves margins by automating high-value tasks (customer service, software development, targeted advertising). If operating margins compress despite revenue growth, the « AI pays » narrative becomes difficult to sustain.

The concentration phenomenon in US equity indices is unprecedented in modern financial history. The Magnificent 7 collectively represent 33.4% of the S&P 500 market capitalization as of July 18, 2026. For an investor holding a standard « world » or « S&P 500 » ETF, one third of their equity exposure is concentrated in seven US technology companies.

This concentration has direct consequences for risk management. A synchronized correction across all seven stocks — possible in the event of a generalized Q2 earnings disappointment — would mechanically translate into a 10-15% index decline, even if the remaining 493 components remained stable. This is the scenario we are monitoring most closely this week.

Given this configuration — stretched valuations, high concentration, decisive earnings week — our investment committee recommends a pragmatic rather than directional posture.

Do not add exposure before publications. Buying a technology stock at 38 times earnings on the eve of its results means taking a binary directional option with an unfavorable risk-reward profile. The upside potential in the event of a positive surprise is limited (consensus already prices in an optimistic scenario), while the downside in the event of a disappointment remains wide open. We recommend waiting for publications before any rebalancing.

Partial profit-taking on the most concentrated positions. For portfolios where the tech allocation exceeds 30% of equity exposure, a partial reduction — in the order of 5 to 8% of the relevant allocation — creates room to re-enter after publications if numbers are strong, or to limit losses in the event of a disappointment.

Maintain the defensive buffer. Consistent with our positioning since the FOMC on July 17: risk gauge at 4.5/10, overweight US Treasuries 3-7 years and physical gold. Both assets play their cushioning role and allow investors to benefit from a risk-off shift if tech results disappoint.

Identify tech defensive alternatives — technology companies with reasonable valuations that benefit from AI spending without directly carrying it on their balance sheets. Enterprise software publishers (ServiceNow, Salesforce), cybersecurity companies (CrowdStrike, Palo Alto), and certain automotive semiconductor players offer more balanced risk-reward profiles at this stage of the cycle.

Related Articles

MARKETS

FOMC July 2026: Hold Decision and Market Scenarios

The Fed keeps rates unchanged — analysis of the implications for bonds and equities.

MARKETS

Semiconductors: AI Correction and Sector Repositioning

Understanding the mid-July correction and identifying opportunities in the chip ecosystem.