GOVERNANCE WATCH
By InsidEntity Editorial Desk · Jul 31, 2026 · 8 min read
Company Risk Rating: 3.12, Good
Auditor Independence pillar: 0.00
Amazon reported its second quarter on 30 July 2026. Total net sales passed $200 billion in a single quarter for the first time in the company’s history, and Amazon Web Services grew at its fastest rate in eighteen quarters. Both facts are real, and both were earned by the operating business.
Neither is what produced the headline. Amazon reported net income of $62.6 billion and diluted earnings per share of $5.75 against a consensus estimate of $1.82. Inside that number sits $53.4 billion of non-operating pre-tax income, disclosed by the company as arising primarily from its investments in Anthropic, a private company Amazon does not control and does not consolidate.
This is the third instance this series has documented in a single earnings season. Tesla’s second quarter carried a $1.005 billion mark on its SpaceX holding. Alphabet’s carried roughly $99 billion in gains on equity securities, led by Anthropic. Amazon’s carries $53.4 billion, primarily from the same private company. Two public companies, one quarter, one privately held asset, and more than $150 billion of reported profit between them.
1. What was reported
| Q2 2026 | Q2 2025 | Q2 2026 | Change |
|---|---|---|---|
| Net sales | $167.7bn | $200.6bn | +20% |
| AWS revenue | $30.9bn | $42.2bn | +36.7% |
| AWS operating income | $10.2bn | $16.6bn | +63% |
| Total operating income | $19.2bn | $27.5bn | +43% |
| Net income | $18.2bn | $62.6bn | +244% |
| Diluted EPS | $1.68 | $5.75 | +242% |
| Advertising services | $15.7bn | $19.8bn | +26% |
Source: Amazon.com, Inc. second quarter 2026 results announcement, 30 July 2026.
The operating story is strong and it does not need the mark to stand up. AWS accelerated from 28% growth in the first quarter to 36.7% in the second, its fastest since 2021, and beat the consensus estimate of 31%. AWS operating income of $16.6 billion came in well above the $13.6 billion expected, at a margin near 39%, and AWS now produces close to 61% of Amazon’s total operating profit. The company’s AI business and its in-house chips business each passed a $25 billion annualised run rate. Advertising grew 26%. Andy Jassy called AWS “booming,” and the segment numbers support the word.
Two figures on the other side of the ledger deserve equal billing. Capital expenditure reached $54.2 billion in the quarter, up 68%, and Amazon raised its full-year 2026 capital spending guidance to approximately $220 billion. On a trailing twelve month basis, free cash flow swung to an outflow of $7.6 billion, from an inflow of $18.2 billion a year earlier.
So the company that reported $62.6 billion of quarterly net income is, on a trailing twelve month basis, consuming cash rather than generating it. Both statements are accurate, disclosed, and derived from the same set of accounts. They are answering different questions.
2. The $53.4 billion line
Amazon disclosed that second quarter net income included non-operating pre-tax income of $53.4 billion, primarily from its investments in Anthropic. That figure is roughly double the entire operating income of the business for the quarter.
A fair value gain on a private holding is not revenue, not cash, and not repeatable by management effort. It is an accounting recognition that an asset Amazon already owned has been revalued upward, most commonly following a funding round that reprices the equity. The gain is legitimate, it is required under the applicable accounting standard, and Amazon disclosed it plainly in the release rather than burying it. Nothing in this section suggests otherwise.
What it does mean is that the single largest contributor to Amazon’s reported profit this quarter was a valuation judgment about a company whose accounts Amazon does not consolidate, whose revenue Amazon does not report, and whose board Amazon does not control. The operating business earned $27.5 billion. The mark contributed $53.4 billion before tax.
Strip the mark and Amazon had a very good quarter. Include it and Amazon had a historic one. The difference between those two descriptions is an estimate.
3. The relationship that sits on both sides of the ledger
The governance feature specific to Amazon is that Anthropic is not a passive holding. It is a customer, a partner, and a counterparty in the same business line whose growth the market is buying.
Amazon supplies Anthropic with cloud infrastructure and with its own Trainium chips. Anthropic’s models, including Claude Opus 5, are distributed on Amazon Bedrock, the managed model service Amazon cited in the same results announcement as growing quickly, with customers spending more in the quarter than in all prior quarters combined. AWS revenue and the Anthropic equity mark are therefore driven, in part, by the same relationship.
That is the same circularity this series flagged at AMD, where a customer holds a warrant over roughly 10% of the company it buys from, and at Alphabet, where the company supplies cloud and TPU capacity to a company whose valuation it also marks through earnings. It is disclosed at all three. It is novel at all three. And it means the pillar most people would reach for, Shareholder Influence, does not capture it, because the entanglement runs through the commercial relationship rather than the share register.
4. The auditor, and the point almost nobody has made
Ernst & Young is Amazon’s independent auditor, an appointment ratified by shareholders each year. The relationship dates to the company’s earliest years as a public company, well past the ten year threshold at which InsidEntity’s methodology stops crediting auditor tenure. On the ladder this series has applied at Nike, Boeing, Shopify, Tesla, FirstRand and Investec, the Auditor Independence pillar scores nil.
The arithmetic consequence is fixed. With one pillar at zero and the overall rating calculated as the simple average of four equally weighted pillars, Amazon’s Company Risk Rating cannot exceed 3.75 even if the other three pillars all score Benchmark. That ceiling is a property of the tenure, not of the board. Amazon’s actual rating is 3.12.
| Component | Amazon | Alphabet |
|---|---|---|
| Director Independence | 5.0 | 5.0 |
| Shareholder Influence | 5.0 | 5.0 |
| Director Capacity | 2.46 | 3.60 |
| Auditor Independence | 0.0 | 0.0 |
| Overall Rating | 3.12 | 3.40 |
Amazon (5.0 + 2.46 + 0.0 + 5.0) ÷ 4 = 3.12. Alphabet (5.0 + 3.60 + 0.0 + 5.0) ÷ 4 = 3.40. Ratings update regularly; check each company’s live report for the current view.
Read that table across. Ernst & Young has audited Alphabet since before its 2004 initial public offering, a relationship this series has already scored at nil. Both companies hold stakes in Anthropic. Both marked those stakes upward in the same quarter. Both marks flow through reported profit. The same audit firm is testing the valuation of the same private asset on both sets of books.
And the two companies are identical on three of four pillars. Both score 5.0 on Director Independence, in both cases the platform’s default, because neither has completed the independence questionnaires. Both score 5.0 on Shareholder Influence. Both carry a nil auditor pillar. The entire 0.28 point difference between Amazon’s 3.12 and Alphabet’s 3.40 is Director Capacity, where Alphabet scores 3.60 against Amazon’s 2.46. On the framework’s measure of whether directors carry the concurrent mandates and board load that stretch their attention, Alphabet’s board has more room than Amazon’s, at the moment both boards are overseeing capital programmes of roughly $200 billion a year and valuation estimates on the same private asset.
Nothing about that is improper. A firm of Ernst & Young’s size audits a large share of the S&P 500, and separate engagement teams, independent valuation specialists and partner rotation are the standard controls. But the framework’s question is not whether the arrangement is permitted. It is who provides independent challenge, and how far removed they are from the thing being challenged. On a valuation judgment worth more than $150 billion of combined reported profit across two companies in one quarter, the answer is one audit firm, in both seats, both past the point at which this framework stops crediting the relationship.
5. Confirmed and contested
Confirmed, from the company’s own results announcement and the reporting of it: net sales of $200.6 billion, up 20%; AWS revenue of $42.2 billion, up 36.7%; AWS operating income of $16.6 billion; total operating income of $27.5 billion, up 43%; net income of $62.6 billion and diluted EPS of $5.75; non-operating pre-tax income of $53.4 billion primarily from investments in Anthropic; second quarter capital expenditure of $54.2 billion; full-year 2026 capital expenditure guidance of approximately $220 billion; trailing twelve month free cash flow outflow of $7.6 billion.
Requiring verification before publication: the exact appointment date of Ernst & Young as Amazon’s auditor, and separately as Alphabet’s, should be taken from the platform’s Company Auditors record rather than secondary sources. This series has already found that external tenure figures and platform records can differ materially. The pillar score of nil is confirmed on the platform and holds under any appointment date before 2016. The Director Capacity scores shown for both companies, 2.46 for Amazon and 3.60 for Alphabet, are carried from prior reporting in this series and have not been re-confirmed against the live platform in this pass.
Not alleged: no wrongdoing, misstatement, or breach of any standard by Amazon, Alphabet, Anthropic, Ernst & Young, or any individual is suggested by anything in this article. Every figure discussed was disclosed by the companies themselves. The fair value accounting described is required, not elective.
6. What to watch
Three things. Whether Amazon’s trailing free cash flow returns to positive as the $220 billion capital programme runs, or whether reported profit and cash generation continue to diverge. Whether the Anthropic mark reverses in a future quarter, which would run through earnings with the same force in the opposite direction and would test how clearly investors distinguished the operating result from the estimate. And whether either Amazon or Alphabet moves on its audit relationship, which is the one pillar in each case that is capping the rating and the one that a board can change by decision.
Amazon’s full Company Risk Rating of 3.12, with the pillar breakdown and auditor appointment record behind it, is on InsidEntity, alongside Alphabet’s 3.40 for the side-by-side above. A record quarter tells you what the business sold. The governance layer tells you who checked the number that made it a record.
About the InsidEntity Company Risk Rating
The InsidEntity Company Risk Rating (CRR) scores companies on a 1 to 5 scale across four equally weighted governance pillars: Director Independence, Director Capacity, Auditor Independence, and Shareholder Influence. A 5 is Benchmark, a 1 is Risk, and the overall rating is the average of the four, so a weakness in any single pillar shows up in the headline number. Ratings update regularly as corporate changes occur, so the platform is always the current view. Explore ratings for companies across 145 stock exchanges at InsidEntity.
Know your entity. This is independent research, not financial advice.