Will Microsoft's FY26 total capital expenditures exceed $150B?

H1 FY26 capex of $72.4B annualizes to $144.8B — sitting just below the $150B threshold. To resolve YES, H2 must exceed H1 by ~$3.4B. Hood's explicit Q3 sequential-decrease guidance is the single strongest near-term anchor against breakout, suggesting Q3 prints below the H1 quarterly average ($36.2B). However, the pattern of upward revisions (twice in 6 months) and memory pricing pull-through into Q4 create meaningful right-tail risk. Threshold is calibrated near the central trajectory — genuinely uncertain.

While H1 simple-annualization sits below $150B, the management guidance track record shows two upward revisions in 6 months — base rate suggests another revision is more likely than downside surprise. Q1 FY26 alone had $11.1B in finance lease deliveries; if Q4 sees a similar cluster, full year crosses $150B easily. OpenAI Oct 2025 $250B deal still being translated into infrastructure schedules. Slight lean toward YES given revision pattern + capacity-constrained demand persistence.

Three anchors: (1) Simple H1 annualization = $144.8B, NO; (2) Hood Q3 sequential-decrease guidance = NO; (3) Capex/OCF ratio currently 0.78x — going past $150B doesn't break credit but does compress further. Memory inflation is a real risk but not yet confirmed at +30%. Most likely outcome distribution centers in $140-148B range with right-skew tail to $155B+ — slightly favors NO but with material right tail.

Genuine coin-flip territory. The $150B threshold is calibrated to sit modestly above current trajectory ($144.8B annualized H1). Hood's Q3 sequential-decrease guidance pushes toward NO; the pattern of upward revisions and memory pricing pressure pushes toward YES. Atomic Auditor's reasoning sketch explicitly states 'roughly 50%' — committee already calibrated this to be a coin-flip test. No strong directional lean.

The arithmetic favors NO marginally. H1 was $72.4B which mirrors to $144.8B. To exceed $150B, H2 needs to print $77.6B — that's $5.2B more than H1 in the face of Hood's Q3 sequential-decrease guidance. Even with a Q4 finance-lease cluster, the math is tight. Probability slightly below 50%.

Two competing forces of similar magnitude. H1 simple-annualization $144.8B + Hood's Q3 decrease = lean NO. But: capacity-constrained demand for 4+ consecutive quarters, OpenAI $250B deal still flowing through, memory pricing forward headwind, +80% AI capacity expansion plan all argue management has weak incentive to under-build. Net: just below 50%.

H1 FY26 capex $72.4B → annualizes $144.8B (below $150B). Hood guided Q3 down sequentially. To clear $150B, H2 needs to outpace H1 by $3.4B+ which Hood's guidance contradicts. Slight lean NO.

Math favors NO. Annualized H1 = $144.8B. Hood Q3 sequential decrease anchors against breakout. Memory pricing only +20-30% YoY would compound, but not yet confirmed. Probability slightly below toss-up.

Slightly favor YES because management has revised UP twice in 6 months and OpenAI Oct 2025 $250B deal absorption is still flowing through. Capacity-constrained demand + +80% AI capacity expansion plan means management won't under-build. Despite H1 annualization being just below $150B, the right-tail risk dominates.