Three AI infrastructure giants—NVIDIA, Broadcom, and Microsoft—have all pulled back sharply from spring highs even as their AI revenue growth accelerates, creating what investors see as a buying opportunity heading into the second half of 2026.

What happened: NVIDIA closed at $200.42 on June 10, down 26% from its 52-week high of $236.26, despite Q1 FY2027 Data Center revenue hitting $75.25 billion（約12兆円） (+92% year over year) and management guiding Q2 to $91.0 billion（約15兆円） in revenue. Broadcom plunged 22% in a single week to $372.10 (52-week high: $495) after reporting Q2 AI semiconductor revenue of $10.80 billion（約1.7兆円） (+143% year over year), but guidance missed investor expectations. Microsoft fell 17% year to date to $397.36 (52-week high: $551.05), even though Q3 FY2026 Intelligent Cloud grew 30% year over year to $34.68 billion（約5.5兆円） and its AI business reached an annual revenue run rate of $37 billion（約5.9兆円） (+123% year over year).

Why it matters: All three companies anchor the spending cycle for AI infrastructure—silicon, custom accelerators, and cloud platforms—yet are trading well below recent peaks. The article frames this as a potential entry point because their underlying revenue trajectories remain accelerating; for instance, NVIDIA's Data Center Networking revenue jumped 199% year over year, Broadcom guided Q3 AI semiconductor revenue to $16.0 billion（約2.6兆円） (more than 200% year over year growth), and Microsoft's Commercial RPO (a long-duration backlog) reached $627 billion（約100兆円） (+99% year over year). The pullback appears to reflect investor caution rather than deteriorating business fundamentals.

What to watch: NVIDIA trades at a forward P/E of 23 with 95% bullish analyst sentiment and $80 billion（約13兆円） in additional buyback authorization; Broadcom trades at a forward P/E of 34 with 92% bullish sentiment; Microsoft, at a forward P/E of 21, has the cheapest multiple in the group but faces scrutiny over capex of $30.88 billion（約4.9兆円） in the quarter (up 84% year over year). The next earnings reports, especially commentary on Azure growth and capex, will signal whether AI returns justify the buildout pace.