If I Were in My 20s, I'd Buy These 2 Trillion-Dollar Stocks and Hold Them Forever

The article recommends Amazon and Microsoft as long-term AI investments, citing AWS revenue growth of 37% in Q2 and Azure growth of 43% in the June quarter. Amazon reported a $496B AWS capacity backlog and Microsoft cited a $678B Azure AI backlog, supporting substantial future cloud expansion; both trade at forward P/E ratios near 25. Microsoft’s paid Copilot for 365 licenses reached 30M as of June 30, up 50% from March, while Amazon’s AWS AI coding assistant Kiro tripled usage sequentially in Q2.

Analysis

The actionable issue is not AI demand but conversion of booked demand into revenue without a further leg-up in capital intensity. AMZN and MSFT can sustain premium growth only if new capacity is monetized at improving utilization; otherwise depreciation, power contracts, and accelerated chip refresh cycles absorb the incremental gross profit. Over the next 1-3 months, cloud growth prints and capex guidance matter more than backlog headlines, which are not equivalent to contracted, margin-accretive revenue.

MSFT has the cleaner earnings bridge: enterprise distribution allows software attach revenue to scale with minimal incremental sales expense, and higher recurring mix supports valuation resilience if infrastructure returns remain credible. AMZN offers greater operating leverage if AWS growth remains elevated because retail/logistics profitability can dilute consolidated capex concerns, but it is also more exposed to cloud price competition and enterprise customers using lower-cost inference alternatives. The second-order beneficiary is NVDA only if hyperscaler capex remains GPU-heavy; proprietary silicon adoption by both platforms is a medium-term risk to NVDA unit economics, even while aggregate compute spending rises.

Consensus appears too comfortable treating similar headline multiples as evidence of equal value. MSFT's valuation requires paid-seat adoption to translate into measurable ARPU and retention, while AMZN's requires AWS margins to withstand a larger mix of lower-priced AI workloads. A deterioration in either cloud margin, a material increase in useful-life/depreciation assumptions, or capex guidance rising faster than cloud revenue would challenge the AI infrastructure rerating within one or two earnings cycles.

Key Decisions for Investors

• Prefer long MSFT / short AMZN over the next 1-3 months: MSFT has the more defensible software-monetization path, while AMZN carries greater cloud-margin and retail-capex noise. Target a 8-12% relative move; exit if AWS growth reaccelerates while its operating margin expands, or if paid Copilot monetization disappoints.
• Maintain NVDA exposure only as a capped satellite rather than the primary AI expression through the next earnings cycle. Add only if hyperscaler capex guidance rises sequentially and management commentary confirms GPU mix is stable; reduce on evidence that internally designed accelerators are taking meaningful inference share.
• For MSFT, use downside-defined exposure via 6-9 month call spreads rather than outright leverage ahead of cloud/capex disclosures. The thesis is falsified if commercial cloud growth decelerates materially while capital expenditures remain elevated and paid AI attach fails to offset depreciation.
• Set an earnings-monitor alert for AMZN and MSFT: cloud revenue growth, cloud operating margin, capex as a percentage of revenue, and management commentary on power availability/utilization. Do not add on backlog narratives alone absent evidence that revenue recognition and returns on invested capital are improving.