Microsoft and Tesla Enter September With Momentum but Which Is Better?

Microsoft closed a record fiscal year with Q4 revenue of $90B (+18%) and Azure crossing $100B annually, alongside Copilot reaching 30M paid seats; CFO commentary highlighted demand exceeding supply and a $678B commercial RPO. Tesla posted record Q2 deliveries of 480,126 vehicles but missed EPS ($0.33 vs $0.54 est.) as operating margin collapsed to 1.4%, with free cash flow at -$1.09B amid capex up 142% YoY. The article frames September positioning as favoring MSFT on durable monetization (Azure efficiency gains) while TSLA’s upside hinges on near-term ramp milestones (Cybercab/Optimus) and margin stabilization.

Analysis

The market is still treating “AI exposure” as one trade, but the economics are diverging sharply. MSFT is monetizing compute demand now, which means its capex intensity is self-funding and should support multiple stability; the cleaner second-order winners are the infrastructure bottlenecks around it, especially VRT, ANET, and ETN, where order books can re-rate before the software names do. TSLA, by contrast, is asking the market to underwrite a long-duration option with near-term margin dilution, so any disappointment in rollout timing will hit both earnings and the multiple.

In the next 1-3 months, the key question is not whether both companies can grow, but whether investors keep paying for growth that has different financing profiles. MSFT’s risk is mostly execution on capacity deployment: if GPU supply eases and capex-to-revenue normalizes, the backlog converts and the earnings revision cycle extends. For TSLA, the nearer catalyst is margin stabilization; without that, each delivery beat risks being read as lower-quality volume, which is a bad setup for a 300x+ stock.

The contrarian miss is that TSLA’s optionality is already very expensive while MSFT’s durability is still being discounted as “boring.” That sets up a relative-value trade more than a directional one: the spread should favor the name with visible cash conversion unless TSLA can prove that FSD attach is turning into recurring profit, not just feature adoption. Falsifier for the bearish TSLA view: a clean two-quarter inflection in automotive gross margin and FCF, not just strong rhetoric around robotaxi.

Over 6-18 months, if MSFT capex keeps pulling through, the real upside may migrate to the picks-and-shovels basket rather than the mega-cap software holder. If TSLA’s AI story remains unmonetized, the stock becomes increasingly dependent on sentiment and macro liquidity, which is a poor setup if rates stop falling or risk appetite cools.

Key Decisions for Investors

• Long MSFT / short TSLA as a 3-6 month relative-value pair; target is multiple compression in TSLA vs steady earnings revisions in MSFT, with the trade invalidated if TSLA delivers two straight quarters of margin/Fcf improvement.
• Buy MSFT call spreads into pullbacks rather than chasing spot; the thesis is backlog conversion and cash generation, not further multiple expansion, so structure for moderate upside with limited theta bleed.
• Watch VRT and ANET as cleaner beneficiaries of MSFT capex than MSFT itself; if Azure supply constraints persist into the next guidance cycle, these infrastructure names should lead the next leg higher.
• Avoid outright shorting TSLA on a one-week horizon; use put spreads or a call overwrite against any rally tied to robotaxi headlines, since narrative-driven squeezes remain the main near-term risk.
• Set an alert on TSLA auto gross margin and free cash flow in the next two prints; if margin fails to stabilize while capex stays elevated, the equity should de-rate despite delivery growth.