Fed Rate Cut Hopes Build as Crude Surges to 5-Week High and Tech Faces AI Capex Scrutiny

The week of July 21, 2026 compressed the entire H2 macro debate into a single trading frame: US CPI cooling to 2.6% (with core MoM at 0.16% — the lowest sequential reading in 14 months) drove September rate cut probability to 84% and knocked the 10-year Treasury 14bps lower to 4.12%. But Brent crude simultaneously surged to $87.50/bbl (WTI to $83.40) on fresh Hormuz disruption fears and a 3.8M barrel EIA inventory draw — a price level that, if sustained, would add approximately +0.17% to US headline CPI by Q4 via energy pass-through. And Big Tech's Q2 earnings revealed a structural tension: hyperscaler capex totalled $45B+ in the quarter (up 34% YoY) with cloud revenue growing 21.4% YoY — the AI infrastructure buildout is real and accelerating, but the monetisation gap between "AI infrastructure spend" and "AI-attributable revenue" is widening, creating the capex scrutiny that drove semiconductor indices down 2.8% mid-week.

---

📈 US CPI at 2.6% — The September Rate Cut Mechanics

Breaking Down the June 2026 Inflation Report

June 2026 US CPI — detailed breakdown:

The two drivers of the June beat:

1. New vehicles −0.8% MoM: Automaker pricing discipline collapsed as used car supply normalised post-pandemic. This is structural (not temporary) — EV price competition and inventory glut are suppressing new car prices
2. Energy −2.8% YoY / −0.40% MoM: Reflects the June oil price decline (Brent $84.80→$71.70) — this is partially temporary if Brent re-escalates to $87.50

The core 0.16% MoM — why it's so significant: If core CPI runs at 0.16% MoM for 12 months → annualised core = 1.92% → below the Fed's 2% target. Even though one month is not a trend, the lowest monthly core print in 14 months signals that the disinflationary trajectory is intact:

A 3-month annualised core at 2.3% is the number that gives FOMC doves their strongest case for September action.

Federal Funds futures — the 84% September cut probability:

The Treasury yield reaction:

Parallel shift down across the curve = consistent with front-end rate cut expectations transmitting to long end. DXY softening provides USD-denominated commodity tailwind and emerging market debt relief.

The ECB-Fed convergence: European sovereign bonds tracked: German Bund 10Y fell 9bps to 2.34%. The ECB-Fed convergence (10Y spread: Fed 4.12% minus Bund 2.34% = 178bps) is at its narrowest since Q1 2026 — narrowing spread = EUR/USD appreciation potential.

---

🛢️ Brent at $87.50 — The 5-Week High and Inventory Draw

The Double Supply Shock

WTI and Brent trajectory — the 5-week high context:

The near-+23% rally in 6 weeks for Brent is one of the fastest recoveries in the post-ceasefire period — driven by the double supply shock of geopolitical risk premium reinstatement + fundamental inventory draw.

The EIA inventory draw — mechanics:

A 3.8M barrel draw when consensus expected 1.5M = the physical market is tighter than financial models suggest. This often happens in summer driving season peak (June–August) when US gasoline demand is seasonally highest.

Energy sector market impact:

The energy-inflation feedback loop for the Fed: The CPI gave the Fed hope; oil threatens to take it back:

• June CPI: −0.40% MoM from energy (Brent at $71.70 during June)
• If Brent stays at $87.50 through July: July energy contribution = +0.55% MoM (reversal)
• July CPI projection: If energy reverses from −0.40% to +0.55% = +0.95pp energy swing in July CPI
• September CPI risk: If oil holds, August and September CPI readings could re-accelerate headline above 3.0%

This is the Fed's "one step forward, one step back" problem — the exact reason Warsh uses "patient and data-dependent" language.

---

💻 Tech Q2 Earnings — $45B Hyperscaler Capex and the Monetisation Gap

The AI Infrastructure Spending vs Revenue Accountability Debate

Hyperscaler Q2 2026 capex — the $45B combined figure:

The capex-to-revenue ratio problem:

The monetisation gap — what the numbers reveal: Cloud revenue is growing at 21.4% YoY. AI capex is growing at 34% YoY. The gap = 12.6 percentage points of capex growth that is not yet generating proportional revenue growth. This is the "monetisation gap" that institutional investors are now pricing as a risk:

1. Depreciation catch-up: H100/H200 GPUs depreciate over 5 years → rising quarterly depreciation charges erode operating margins
2. AI workload ramp time: New AI services (Azure OpenAI, Google Vertex AI) typically take 12–18 months post-deployment to achieve full utilisation
3. Customer concentration risk: ~40% of AI cloud revenue comes from <50 enterprise customers — still early-stage enterprise diffusion

Semiconductor index (SOX) −2.8% mid-week — what caused it: The SOX drawdown mid-week was triggered by:

1. TSMC Q2 guidance: "Customer inventory digestion" in smartphone and PC segments offsetting AI demand strength
2. Intel Gaudi 3 delay: Intel confirmed 3-month delay in Gaudi 3 AI accelerator ramp, citing yield issues at Intel Foundry Services 18A node
3. Analyst multiple compression: Morgan Stanley trimmed Nvidia P/E target from 38× to 34× on valuation normalisation, citing AI monetisation timeline risk

Cloud infrastructure 21.4% — the resilience story: Despite the capex scrutiny, 21.4% cloud revenue growth for hyperscalers represents genuine fundamental strength:

• Azure AI revenue (Copilot, Azure OpenAI Service): +65% YoY (within Azure's 26% total growth)
• Google Vertex AI API calls: +120% QoQ
• AWS Bedrock (managed AI service): New enterprise customers +38% QoQ

The market is not questioning whether AI cloud revenue is growing — it's questioning whether the pace justifies $45B/quarter in capex, or whether the build-out will eventually create excess capacity.

---

📌 The Bottom Line

• us-cpi-2-6pct-core-0-16pct-mom-september-84pct-cut-probability: June CPI 2.6% headline / 2.6% core (−0.2pp MoM); lowest core MoM: 0.16% (14-month low); 2 drivers: new vehicles −0.8% MoM (structural EV competition) + energy −2.8% YoY; 3M annualised core: 2.9%→2.7%→2.3% (accelerating disinflation); September cut probability 62%→84% (+22pp); 10Y Treasury −14bps to 4.12%; DXY −0.5% to 104.1 (EM debt relief); ECB-Fed 10Y spread 178bps = narrowest since Q1 2026 (EUR/USD upside potential); November 71%, December total 75bps = 46% probability.
• brent-87-50-wti-83-40-5week-high-hormuz-3-8m-bbl-eia-draw: Brent +22.9% in 6 weeks (June low $71.20 → July 21 $87.50); EIA surprise: crude −3.8M (vs −1.5M expected), Cushing −0.9M; double supply shock: Hormuz risk premium reinstatement + physical inventory draw; energy sector rally: XOM +3.8%, CVX +4.1%, Shell +3.2%; airlines −3.2% (jet fuel cost); maritime insurance +18%; distillate crack spread $24.50 (vs normal $18-20 = refiner windfall); energy-CPI feedback: June energy −0.40% MoM → if $87.50 holds, July energy flips to +0.55% MoM = 0.95pp swing → September headline potentially >3.0%.
• tech-q2-earnings-45b-hyperscaler-capex-21-4pct-cloud-growth-semiconductor-pullback: Combined hyperscaler Q2 capex $45.1B (+34% YoY): Microsoft $16.5B, Alphabet $13.4B, Amazon $12.0B, Meta $3.2B; cloud revenue +21.4% YoY vs capex +34% = 12.6pp monetisation gap; capex as % cloud revenue: Alphabet 111% (Google Cloud), Microsoft 43%, Amazon 43%; monetisation gap: 5-year GPU depreciation + 12-18 month workload ramp time + top-50 enterprise customer concentration; SOX −2.8%: TSMC inventory digestion + Intel Gaudi 3 delay + Morgan Stanley multiple cut (38×→34×); resilience: Azure AI +65% within 26% total, Google Vertex +120% QoQ, AWS Bedrock enterprise customers +38% QoQ.

---

📬 Stay Updated

Get global markets analysis delivered to your inbox every Friday. Subscribe to our free newsletter →

---

Disclosure: This post contains affiliate links. If you purchase through our links, we earn a small commission at no extra cost to you. We only recommend products we believe in.