Apple vs. Microsoft: What Revenue Trajectories Tell Us About Two Different Megacaps

San Francisco, May 25, 2026. The two largest companies in the S&P 500 by market capitalization are once again trading places at the top, and the gap between them is now small enough that a single earnings cycle can flip the ranking. Microsoft (MSFT) closed Friday at roughly $3.7 trillion in market cap, Apple (AAPL) at $3.5 trillion, with a combined weight of about 13.8% of the index. A piece in Motley Fool over the weekend looked at the revenue trajectories of both names and concluded, with appropriate caution, that one is accelerating while the other is grinding. That framing is correct in direction and almost useless in magnitude, which is the part worth unpacking.

The reason it matters is mechanical. When two stocks together account for roughly one euro of every seven invested in a global cap-weighted index, the difference between a 6% topline growth rate and a 14% topline growth rate, compounded over three years, decides whether a passive European saver via a S&P 500 UCITS ETF is buying Apple at a forward multiple that has quietly drifted into the 95th percentile of its own history, or buying Microsoft at one that is merely expensive. ARIA surfaces both, side by side, with the calibration data attached.

Two Revenue Engines, Two Distributions

Apple's trailing-twelve-month revenue sits near $395 billion, growing in the low to mid single digits depending on which quarter you anchor on. The Services segment, now around $100 billion annualized, is the only line item growing convincingly above 10%, and it has been doing the heavy lifting for gross margin expansion for six consecutive quarters. iPhone revenue, the cyclical heart of the model, has oscillated in a narrow band between $190 and $205 billion for three fiscal years. Greater China has stopped getting worse, which the sell-side is treating as a positive, a framing that should be examined rather than absorbed.

Microsoft's trailing-twelve-month revenue is approximately $270 billion, growing at a blended 14 to 16% pace. Intelligent Cloud, anchored by Azure, is compounding north of 28% in constant currency. Productivity and Business Processes, where Microsoft 365 Copilot lives, is now contributing measurable incremental ARPU rather than promised ARPU, with attach rates in the enterprise base running ahead of where the consensus model had them eighteen months ago.

The distinction that matters for an investor is not which growth rate is higher. It is which distribution of future revenue is wider. Apple's next-twelve-month revenue distribution is tight: the interquartile range of sell-side estimates is roughly 3.5 percentage points of growth. Microsoft's is wider, around 5.5 points, because Azure capacity, GPU supply, and enterprise AI seat adoption are still being discovered in real time. ARIA's regime detector flags Microsoft as operating in a high-dispersion regime and Apple in a low-dispersion one. The same point estimate of "growth" carries very different information content in each case.

What the Multiples Are Actually Pricing

Apple trades at roughly 31x forward earnings against a ten-year median of 22x. Microsoft trades at roughly 33x forward earnings against a ten-year median of 27x. On a PEG basis, using consensus three-year EPS CAGRs, Apple is at approximately 3.6 and Microsoft at approximately 2.0. Those are not equivalent valuations. They reflect a market that is willing to pay a quality and capital-return premium for Apple, and a growth premium for Microsoft, and the question is whether each premium is anchored to something the cash flows can defend.

Apple's buyback program retired about 2.5% of shares outstanding in the last twelve months. That is real, it is durable, and it shifts the per-share denominator in a way that flatters the earnings line even when topline grows 3%. Microsoft's buyback is smaller in proportional terms, around 0.7%, because the company is redirecting roughly $80 billion of annualized capex toward data center infrastructure. One firm is returning capital because the marginal project is no longer attractive enough. The other is consuming capital because the marginal project still clears its hurdle rate, or at least the company believes it does. Both can be correct at the same time. Neither narrative survives if it turns out wrong.

The 10-year Treasury yield this morning is at 4.41%, real yields on the 10-year TIPS at 2.05%. At those discount rates, the present value of a dollar of earnings in 2031 is roughly 30% lower than it was at the 2021 peak of multiple expansion. Both names have absorbed that compression and re-rated higher anyway. The Russell 1000 Growth index trades at a 12% premium to its ten-year average forward multiple. Apple and Microsoft are jointly responsible for a meaningful share of that premium.

The China Question and the Capex Question

Apple's Greater China revenue is approximately $67 billion annualized, roughly 17% of the total. The Motley Fool piece notes that the segment "stabilized" in the most recent quarter, meaning it stopped declining double-digits and now declines low single-digits. Stabilization is a verbal trick that compresses two very different statistical claims into one. The point estimate has improved. The distribution of outcomes, conditional on the trade and tariff environment that exists in May 2026, has not necessarily narrowed. The yuan is trading around 7.21 against the dollar, USDCNH options are pricing one-month realized volatility at the high end of the post-2022 range, and the BTP-Bund spread widening over the last fortnight is a reminder that political risk is not a US-only variable. ARIA's scenario engine outputs a calibrated 22% probability that Apple's China revenue is more than 8% lower year-over-year by Q4 fiscal 2026. Consensus implicitly prices something closer to 8%.

Microsoft's exposure is different and arguably more concentrated, though in a less visible form. The capex line is the exposure. If Azure growth decelerates from 28% to 18% over four quarters, the GPU procurement schedule already signed becomes a fixed cost against a slower-growing revenue base, and operating margin compresses by roughly 250 basis points on ARIA's central scenario. The market is currently pricing a deceleration to about 24%. Below 20% would require a re-rating. Above 30% would justify the current multiple comfortably. The probability mass between 20% and 30% is wide and not symmetric.

What the Deflated Sharpe Says About Each Trade

A naive backtest of "long Microsoft, short Apple" over the last three years produces a Sharpe of roughly 0.95. After deflating for the number of pair-trade hypotheses an investor could have tested on the Mag 7 over that window, the Deflated Sharpe drops to approximately 0.28. The point estimate looks compelling. The calibrated estimate, accounting for the multiple-testing problem that any pairwise large-cap trade inherits, looks ordinary. This is the kind of correction that distinguishes a backtest from an edge, and it is the reason ARIA reports both numbers rather than the flattering one.

The same exercise on "long the equal-weighted Mag 7, short the cap-weighted S&P 500" produces a Deflated Sharpe close to zero over the last five years. The concentration trade has worked because two or three names did the work, not because megacap as a category has been systematically mispriced. That distinction is invisible in a point estimate and obvious in a distribution.

A Question Worth Answering Before Anything Else

If Apple's iPhone revenue is flat for the next eight quarters and Services grows at 12%, what forward multiple is consistent with a 7% long-run equity return at a 2% real discount rate? Run the number before deciding whether the current 31x is cheap, fair, or expensive. The answer is not in any headline this week, and it is the only one that matters for a buy-and-hold position sized at index weight.

The Motley Fool framing of "what revenue trends reveal" is a fine starting point. It is not a finishing point. Revenue is the most observed and least informative line on the income statement when the cash conversion, the capital intensity, and the discount rate are all moving. ARIA's job is to show all four together, with the dispersion attached, and to flag when the consensus point estimate is sitting in a thin part of the distribution. For Apple, the consensus is currently sitting in a fat part of the distribution and the surprise potential is mostly to the downside. For Microsoft, the consensus is sitting in a thinner part and the surprise potential is roughly symmetric. Those are different risk profiles dressed in the same megacap clothing.

Research, not advice. Always verify before acting.