Should You Buy Microsoft Stock Now That It's Back Within 6% of Its Record?

Microsoft shares rebounded about 46% from their June 25 low to roughly $516 after Azure growth accelerated to 43% in the June quarter, prompting a 26% gain over four trading days following its July 29 earnings release. Management expects Azure growth of 44%-45% in the September quarter, but AI infrastructure spending remains substantial at roughly $175 billion for calendar 2026 and is expected to rise again in fiscal 2027. While fiscal 2026 EPS rose 32% and the valuation declined to about 29x earnings from 40x at the prior peak, free cash flow fell to about $67 billion from $72 billion, leaving the stock viewed as fairly priced rather than a clear buying opportunity.

Analysis

MSFT has shifted from an AI-capacity skepticism trade to an execution-proof trade, but the valuation rerating is largely complete after the sharp post-results move. The next incremental earnings debate is not Azure growth alone; it is whether incremental AI revenue converts into free cash flow fast enough to offset a capital-intensity regime change. If capex remains elevated into FY27, MSFT could retain premium earnings growth while its FCF yield and return-on-invested-capital profile compress—limiting multiple expansion versus asset-light software peers.

The more actionable second-order read is positive for AI infrastructure suppliers: faster deployment and utilization imply demand visibility for NVDA, networking, power/cooling, and data-center equipment remains intact beyond a single hyperscaler quarter. However, MSFT’s ability to monetize new capacity quickly weakens the near-term bear case for excess GPU supply, while simultaneously raising the risk that cloud price competition eventually transfers AI economics from hyperscalers to customers. That margin-pressure risk is more relevant over 6-18 months than over the next quarter.

Near term, consensus appears likely to reward a second consecutive acceleration in Azure growth, but the stock is exposed to asymmetry: revenue can meet expectations while cash conversion disappoints. The key falsifier for a cautious stance is a clear sequential improvement in FCF after lease-adjusted capex, alongside sustained Azure growth above the mid-40s; that combination would support renewed multiple expansion. Conversely, any deceleration toward 40% or indication that capacity is being built ahead of contracted demand would reopen the capital-discipline discount within 1-3 months.

OpenAI concentration is less important as a direct revenue-risk headline than as a counterparty-quality and utilization-risk variable. A funding disruption or materially slower OpenAI consumption would leave MSFT with capacity that may be economically useful but lower-return, particularly if enterprise AI workloads ramp more slowly than frontier-model demand.

Key Decisions for Investors

• Maintain MSFT at benchmark/neutral; do not chase within 5% of prior highs. Add only after the next report if Azure growth sustains at 45%+ and lease-adjusted FCF turns positive year over year. Upside is renewed premium-multiple support; downside is a 10-15% de-rating if capex rises without cash conversion.
• Express near-term AI infrastructure durability through long NVDA versus short MSFT in equal beta-adjusted dollars for 1-3 months. The relative trade benefits if hyperscaler capex remains firm but MSFT’s FCF debate caps its multiple; exit if MSFT demonstrates clear FCF inflection or NVDA faces order-cancellation evidence.
• Use MSFT downside hedges rather than outright short exposure ahead of the next earnings print: buy 3-6 month put spreads financed selectively with upside calls only if implied volatility is below post-earnings realized volatility. The thesis is cash-flow guidance risk, not an Azure demand collapse.
• Monitor Microsoft disclosed remaining performance obligations, data-center lease commitments, and capex-to-revenue trend. A widening commitment base without corresponding commercial backlog growth is the alert that converts the neutral view into an underweight.