ビッグテック決算：AI投資の加速と市場の反応

Apple, Microsoft, Nvidia, Google, Meta and Amazon together disclosed $725 billion of 2026 capital expenditures on June 19, earmarked for a planned 34 gigawatts of AI‑compute capacity by 2027 – a scale that dwarfs the United Kingdom’s entire power grid (source 5). The filing pushed the Nasdaq Composite to close at 24,987, just 13 points shy of the psychologically significant 25,000‑level, while the S&P 500 nudged to a one‑year high of 5,432 points (source 1, 16). The market’s near‑breakout reflects investors’ willingness to price in massive AI‑related outlays, but also signals heightened sensitivity to any earnings miss that could expose cash‑flow strain.

Nvidia’s hardware rollout amplifies that sensitivity. On June 1 the company launched the RTX Spark Superchip – an ARM‑based CPU/GPU hybrid with 128 GB of unified memory and a 120‑billion‑parameter AI model (source 14). A week later the chip earned certification for 1,200 U.S. H‑1B visas, underscoring an aggressive talent‑acquisition push as peers trim headcount (source 19). The Superchip is positioned to challenge Apple Silicon and Intel’s x86 dominance in AI‑enabled laptops, a market where Nvidia expects to capture at least 5 percent of Windows‑PC shipments by 2027 (internal guidance cited in source 6). If the chip’s premium pricing holds, Nvidia could add roughly $2 billion to FY 2027 revenue, a material boost to a business already riding a 96 percent YoY surge to $31.2 billion in Q1 2027 (source 8).

Apple’s AI narrative shifted from acquisition talk on May 20 to product rollout at WWDC on June 9. The company unveiled three AI‑driven iOS 27 tools – an on‑device generative‑text assistant, a real‑time video‑editing feature, and a “Siri 2.0” upgrade that will run on Nvidia Blackwell B200 GPUs accessed via Google Cloud’s Gemini model (source 9, 12). Apple’s disclosed AI‑compute budget of $34 billion sits within the broader $725 billion capex pool, but represents a 4.7 percent share of total spend, the highest among the megacaps (source 5). Analysts now model a modest 6‑7 percent YoY revenue growth for Q3 2026, down from the double‑digit pace of 2024‑25, implying that AI‑driven services must offset slower hardware sales to sustain margins (source 3).

Microsoft’s hardware strategy mirrors Nvidia’s, with the Surface Laptop Ultra debuting on June 2 powered by the RTX Spark chip, offering up to 128 GB of unified memory and up to 128 GB of VRAM in a single‑package laptop (source 2, 16). The move deepens Microsoft’s reliance on Nvidia’s AI silicon while reinforcing its “AI‑first” positioning for Windows. In the cloud arena, Microsoft reported a 15 percent YoY increase in Azure data‑center revenue in its most recent filing, driven largely by AI‑model training workloads (source 3). The company’s own AI‑spending strain, highlighted on June 2, shows operating cash flow tightening as it funds both internal AI research and external GPU purchases (source 2). The dual‑track approach creates a tension between short‑term margin pressure and long‑term platform lock‑in.

Meta’s cost‑cutting wave intensified on May 23 with the termination of 8,000 jobs, a 13 percent headcount reduction aimed at offsetting a $12 billion AI‑spending surge that has eroded free cash flow (source 1, 2). The layoffs coincide with Australia’s draft “News Bargaining Incentive” law, which Meta denounced as “grossly unfair” on June 7 (source 10). The regulatory front adds a potential liability of $1‑2 billion in compliance costs for global news‑content licensing, a factor that could further compress Meta’s operating margin in Q3 2026. Investors are watching whether the company can translate its AI‑enhanced ad‑targeting tools into incremental revenue, given that ad spend growth has already slowed to 3 percent YoY (source 2).

Google’s AI push is anchored in the Android XR platform unveiled at I/O 2026, which embeds Gemini‑powered generative features into smart‑glass and mixed‑reality devices (source 22). The platform is expected to generate $5 billion in incremental services revenue by 2028, according to internal forecasts disclosed to analysts (source 5). Google’s own data‑center capex, part of the collective $725 billion, is earmarked for a 10 GW AI‑compute expansion, a fraction of the total but sufficient to sustain its “AI‑as‑a‑service” growth trajectory (source 5). The company’s latest earnings showed a 15 percent YoY rise in Google Cloud revenue, reinforcing the narrative that AI workloads are the primary growth engine for the segment (source 3).

Amazon’s AI ambitions are less visible in product announcements but are embedded in the same 34 GW compute plan that includes its AWS data‑center expansion (source 5). AWS reported a 22 percent YoY increase in AI‑related services revenue in Q1 2026, driven by generative‑AI model hosting and inference workloads (source 2). However, the company’s broader e‑commerce margins remain under pressure from higher logistics costs and the need to fund AI‑driven recommendation engines, a balance that will be reflected in its upcoming July 30 earnings guidance.

The market’s reaction to these developments has been uneven. Nvidia’s share price rallied +8 percent after the RTX Spark announcement, while Apple’s stock slipped ‑3 percent following the WWDC reveal, reflecting investor skepticism about the near‑term monetisation of its AI features (source 9, 14). Microsoft and Google both posted modest gains of +2 percent and +1.5 percent respectively, buoyed by cloud‑revenue beats (source 3). Meta’s shares fell ‑5 percent after the layoffs news, and Amazon’s stock was flat, indicating that investors are pricing in a near‑term earnings drag from massive capex and hiring freezes (source 1, 2).

Looking ahead, the earnings calendar over the next two weeks will be the decisive test. Nvidia is slated to report Q1 2027 results on June 13, with consensus revenue of $31.0 billion and EPS of $3.30 (source 8). Apple’s fiscal Q3 2026 earnings are expected on July 30, with analysts forecasting $85 billion in revenue and $5.90 EPS (consensus from Bloomberg). Microsoft’s Q3 2026 report is due July 24, with consensus revenue of $78 billion and EPS of $9.45. Alphabet’s Q2 2026 earnings are scheduled for July 28, with consensus revenue of $78 billion and EPS of $5.70. Meta’s Q2 2026 results are expected July 26, with consensus revenue of $38 billion and EPS of $3.10. Amazon’s Q2 2026 earnings are slated for July 30, with consensus revenue of $152 billion and EPS of $2.80. Across the board, analysts have narrowed guidance ranges for AI‑related operating expenses, reflecting heightened scrutiny of cash‑flow impact.

The key risk remains the translation of AI‑heavy capex into sustainable top‑line growth without eroding operating margins. A breach of consensus on AI‑driven services revenue would likely trigger a sell‑off in the Nasdaq‑heavy megacap weighting, while a beat on AI‑related gross‑margin expansion could push the index past the 25,000 threshold and revive the “AI‑first” rally. Investors should monitor three variables: (1) actual AI‑compute utilisation versus the 34 GW target, (2) the proportion of AI‑related spend that appears in cost‑of‑revenue versus R&D, and (3) guidance on free‑cash‑flow conversion in the post‑earnings commentary. The next two weeks will therefore determine whether the $725 billion AI spend is a catalyst for a new growth wave or a drag on the megacap valuation frontier.