Forget Buying All Seven: The "Magnificent Seven" Stock Most Likely to Double by 2028

The article argues Amazon could nearly double its EBIT by 2028, supported by AWS revenue growth of 37% year over year to a $148.4B trailing-12-month run rate and potential expansion toward $300B in revenue. AWS generated a 37% trailing profit margin, while Amazon's North America and International retail segments produced $627B in revenue and are projected to reach $829B after two years of 15% growth, with a potential 10% margin generating $83B of profit. At a $2.7T market capitalization and roughly 27x trailing EBIT, the upside case also depends on Amazon's estimated 15% Anthropic stake potentially being worth hundreds of billions of dollars if Anthropic IPOs at a projected $2T valuation.

Analysis

The core underwriting error is treating AWS AI demand and retail-margin expansion as independently additive. Both require sustained capex: data-center depreciation, power procurement, and logistics automation can delay consolidated free-cash-flow conversion even if segment EBIT rises. Near-term AMZN upside therefore depends less on another bullish long-range EBIT bridge than on evidence that incremental AWS revenue is arriving with stable contribution margins and that North America fulfillment cost per unit continues falling despite faster-delivery investment.

AWS is leveraged to AI workload growth, but its competitive position is more nuanced than a simple cloud-share story. MSFT can bundle Azure capacity with enterprise software commitments, while GOOG can use TPU economics to protect pricing on inference workloads; this raises the risk that industry capacity additions translate into lower unit pricing in 2027 rather than a clean margin expansion. The more direct public beneficiaries of continued AWS buildout are NVDA and networking/power suppliers, but AMZN's return on invested capital will be the key differentiator as depreciation catches up with the current capex cycle.

The embedded value assigned to Anthropic is highly path-dependent: a private-market valuation is neither liquid nor equivalent to operating earnings, and any IPO could introduce lockup, concentration, and mark-to-market volatility rather than immediate distributable value. Consensus also underestimates the possibility that a successful Anthropic listing increases scrutiny of related-party cloud commitments, making the economic terms of AWS usage more important than the headline stake valuation. This is not a fresh catalyst by itself; the tradable events are AWS backlog/conversion, capex guidance, depreciation growth, and disclosed Anthropic monetization terms over the next 1-3 quarters.

Contrarian view: AMZN may be comparatively attractive versus software-heavy AI beneficiaries if AI demand remains robust but inference pricing compresses, because advertising and retail efficiency offer separate earnings levers. Conversely, a consumer slowdown would expose the fragility of a 10% retail-margin endpoint: wage, shipping, and fulfillment deleverage can erase several years of incremental automation savings. A break in AWS growth versus Azure or a material step-up in capex without corresponding operating-income guidance would falsify the bullish operating-leverage thesis.

Key Decisions for Investors

• Maintain a measured long AMZN versus short GOOG basket over the next 1-3 quarters only if AWS growth reaccelerates while consolidated capex-to-sales stabilizes; AMZN has more retail/advertising diversification, while GOOG faces greater AI-search monetization uncertainty. Exit on two consecutive quarters of AWS growth deceleration or a material reduction in AWS operating margin.
• Do not underwrite Anthropic at a headline private valuation. Create an event watch for IPO filing, Amazon ownership dilution, lockup terms, and disclosed commercial commitments; add AMZN only if the filing establishes realizable value materially above the market-implied stake value rather than relying on promotional estimates.
• Use NVDA as a hedge against an upside AWS capex surprise rather than assuming AMZN captures all AI economics: long AMZN / long NVDA is preferable to an unhedged AMZN AI thesis during the next earnings cycle. Reduce the hedge if hyperscaler capex guidance rolls over or NVDA supply lead times normalize sharply.
• Avoid adding to AMZN on broad AI enthusiasm before earnings. Scale exposure after evidence that fulfillment cost per unit and North America margin improve alongside delivery-speed investment; a miss driven by depreciation, labor, or transportation costs would offer a better entry only if retail demand remains intact.