The headline
Amazon reported its first $200 billion quarter. Net income tripled. AWS growth hit a four-and-a-half-year high. Every headline called it a record. Every number Amazon released is true.
Apply the same three tests that belong in every account review—baseline, calendar, market rate—and the record dissolves.
Not because Amazon misled. Because a scoreboard read without context tells you whatever you want to hear.
Test one: the baseline
Amazon reported net income of $62.6 billion for Q2 2025. The year-ago quarter delivered $15.3 billion. That is the headline that tripled.
$53.4 billion of the $62.6 billion is a non-cash paper gain on Amazon's stake in Anthropic. It appeared because accounting rules required Amazon to mark the investment to market. No customer bought more. No operation improved. The gain is real in the sense that the asset appreciated, but it tells you nothing about whether the retail or cloud business grew profitably.
Strip the windfall and net income was $9.2 billion, down from $15.3 billion in the prior year. The headline inverts.
This is not an accounting trick. It is the same discipline applied to any account review: when a stockout distorts a month, you strip the month before you read the trend. When a one-time event lands in the quarter, you net it out before you measure the baseline. The Anthropic gain is a windfall. It belongs in the balance sheet footnote, not in the operating assessment.
The earnings release itself separates operating income—the profit the business generates from its activities—from net income. Operating income was $17.0 billion, up from $14.7 billion a year earlier. That is the number that measures the quarter's operational performance. It grew 16%, not 309%.
Test two: the calendar
Amazon's North America segment revenue grew 11% year-over-year. International revenue grew 11%. Consolidated revenue grew 20%, the highest rate in nine quarters.
The 20% growth borrowed an event. Prime Day moved from July into June this year. One of Amazon's two largest shopping events shifted from Q3 into Q2, flattering Q2 and pre-loading a shortfall into Q3.
Amazon's Q3 guidance anticipates revenue growth of 8% to 11%. The midpoint is 9.5%. In the earnings call, Amazon's CFO noted that without the Prime Day timing shift, Q3 guidance would imply growth roughly 400 basis points higher—call it 13.5% at the midpoint.
Run the two-quarter average and the picture steadies: if Q2 grew 20% and Q3 is guided to 9.5%, the half-year growth rate is around 14.75%. That is the trend. The 20% quarter is the timing shift made visible.
Every account review watches for this. A Lightning Deal in week two inflates the week and drains week three. A coupon that runs five days into the next month splits the lift across the boundary. The calendar is never neutral. You measure across it or you measure the calendar itself.
The market read the 20% growth as acceleration. The two-quarter trend says it held steady, and the event that moved explained the gap.
Test three: the market rate
AWS revenue grew 37% year-over-year, the fastest growth in eighteen quarters. The release called it "continued momentum." Analysts called it a resurgence. Amazon's share price rose.
In the same quarter, Microsoft Azure grew 43%. Google Cloud grew 82%. AWS's fastest growth in four and a half years was still the slowest growth among the three largest cloud platforms.
This is the test that matters. Growth is only real relative to the market. If the category grows 40% and you grow 20%, you went backwards. If you grew 15% last quarter on a marketplace that grew 16%, you had Amazon's AWS quarter—a record that lost ground.
AWS remains the largest cloud platform by revenue. Its $32.8 billion in quarterly revenue is more than Google Cloud's $11.4 billion. But market share is a trailing indicator. Growth rate is the leading indicator, and AWS is growing slower than the market it still leads.
The explanation offered in the earnings call centered on new AI workloads and long-term enterprise deals. Both are real. Neither changes the relative picture. The competitors are signing the same deals and growing faster doing it.
Growth is only real relative to the market. If the category grows 40% and you grow 20%, you went backwards.
The pattern
Three tests. Three inversions.
The net income that tripled becomes a decline when the windfall is stripped. The revenue growth that accelerated becomes steady when the calendar event is smoothed. The AWS resurgence becomes share loss when the competitors' growth is placed beside it.
None of this requires skepticism about Amazon's disclosure. Every number is accurate. The inversions appear when the numbers are tested against the three measures that belong in every account review: baseline, calendar, market rate.
The same tests apply at every scale. When a product's revenue jumps 50% in a week, the first question is whether a stockout resolved or a deal ran. When a category grows 30% year-over-year, the second question is whether the event calendar matched. When growth looks strong, the third question is what the closest substitute products did in the same window.
The scoreboard is not the performance. The scoreboard is the input to the test.
What this means for sellers
Amazon's results do not set your results, but they do set the expectations that flow downstream. If the market believes Amazon's North America segment is accelerating, the assumption will be that strong sellers grew faster. If the internal account review shows you grew 12% in Q2, the question will be why you missed the 20% wave—even though the wave was a calendar event and your 12% may have been above the smoothed trend.
The work is to apply the same three tests to your own account before someone else applies a different set.
Strip the windfalls. A product that went viral from an external mention, a stockout at a competitor, a one-time bulk order—those belong in the narrative, not in the baseline. The baseline is what repeats.
Mark the calendar. If you ran a Lightning Deal in the month, if Prime Day fell in the period, if you shifted launch timing or inventory allocation, the month's growth rate is not the trend. Average across the event or measure the event separately.
Measure against the market. If your top product grew 18% and the category grew 25%, you lost seven points of relative growth. If the category grew 10% and you grew 18%, you gained eight. The absolute number is not the performance.
Amazon's record quarter fails the tests Amazon's own consulting teams would apply to a seller account under review. The work is to pass them first.