Amazon's AWS prints a 37% quarter, Anthropic stake delivers a $53.4bn windfall, and Zoox gets the keys to charge

Amazon's cloud business is doing something it has not done in a year and a half. AWS revenue grew 37% year-on-year to $42.2bn in the quarter, the fastest pace in 18 quarters, according to Polymarket's wire of the company's 30 July 2026 earnings release.

The print lands on a day Amazon is collecting wins on three different fronts: a cloud number that reasserts its position inside the AI compute boom, a $53.4bn mark-to-market gain on its Anthropic investment that lifted quarterly profit to $62.6bn, and a federal green light for Zoox to charge passengers for driverless rides in vehicles that have no steering wheel and no pedals. Read together, the three pieces form a coherent thesis: Amazon is converting its AI capex into income, its AI equity stakes into bookable gains, and its autonomous-vehicle subsidiary into a commercial operator. None of this is hypothetical anymore.

The cloud engine spools up

The 37% figure matters less for the headline than for what it implies about the AI capex cycle. Hyperscaler growth at this speed is the financial signature of a compute shortage easing into a compute flood. Microsoft's Azure and Google's GCP have spent the past four quarters signalling that the supply of GPU clusters is finally catching up with the demand for training runs and inference. AWS now joins them in showing that the demand did not vanish as the supply landed; it expanded.

Two readings are live. The bullish one, which is the wire consensus, is that AI workload migration is real and durable, and that the spend Amazon committed to data-centre build-out in 2024 and 2025 is now monetising. The reading from the desk's analysis is that the 37% print is also a denominator effect: the year-ago quarter was an unusually soft AWS comparison, and the 18-quarter record sits inside a noisy distribution. The available source items do not specify the segment's operating margin, the capex run-rate, or the customer concentration behind the number.

The Anthropic gain, marked to market

Net quarterly profit of $62.6bn, including a $53.4bn gain tied to Amazon's Anthropic stake, is not a cash number. It is a revaluation of the equity Amazon holds in the AI lab, written up to reflect Anthropic's most recent funding round. The mechanics are familiar from the SoftBank–OpenAI mark-ups of the past 18 months: investors mark their positions to the price of the latest round, the income statement captures the change, and the headlines pick up the headline number.

That is the counter-frame. The $53.4bn is not a finance event; it is an accounting event. The cash does not exist. What it does exist is a price discovery: someone, in a recent Anthropic round, paid enough to make Amazon's stake worth that much more on paper. The size of the mark-up is itself the news. It tells you that the private market for foundation-model equity is still pricing in an outcome that the public market has not yet repriced.

The desk's read is that the gain will be quietly ignored by serious analysts and loudly celebrated by retail commentary. The more lasting number in the release is the cloud revenue figure. The Anthropic figure is the asterisk.

Zoox gets the keys

Hours before the earnings, the National Highway Traffic Safety Administration cleared Zoox to operate a paid robotaxi service using purpose-built vehicles with no human controls, Reuters and TechCrunch reported. The exemption is temporary, which is the part the regulator will want stress-tested; the political part is the precedent. No US federal safety framework has previously countenanced a vehicle without a steering wheel or pedals as a paid product. Zoox now owns that precedent.

The structural frame is straightforward. Two of the three West Coast autonomy programmes, Waymo and Zoox, have crossed the commercial-deployment threshold. Cruise is back in service at smaller scale. Tesla's robotaxi effort remains in a separate regulatory lane, owing to the use of passenger vehicles with controls. The interesting question is no longer whether driverless rides can be sold. It is whether the unit economics work, and which city gets the first large fleet.

What to watch next

Three dates matter. First, the next Anthropic funding round, which will determine whether Amazon's $53.4bn mark-up holds or reverses, and which will inform the valuation of Microsoft's OpenAI stake and Google's DeepMind internal book. Second, the AWS capex disclosure in the next 10-Q, which will show whether the 37% growth was bought at a price the income statement can carry. Third, the first Zoox fare data, which will tell investors whether the regulatory green light translates into a unit that moves.

The sources do not specify any of these timelines. The consensus view is that the AWS print is bullish, the Anthropic gain is noise, and the Zoox approval is structurally real. The plausible alternative reading is that the AWS print is partly a denominator effect, the Anthropic gain is a leading indicator of a private-market AI bubble, and the Zoox approval is a regulatory dead-letter if the public does not get into the vehicles. The evidence to discriminate between these readings is not yet on the page. That is the next story.

Monexus framed the AWS print as a capex-monetisation story rather than a pure AI-demand story, and treated the Anthropic gain as an accounting event rather than a cash event, in line with the approach analysts took to the 2024–2025 wave of chip-investment mark-ups.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://x.com/Polymarket/status/2082924692365516908
• https://x.com/Polymarket/status/2082923183183692212
• https://x.com/Reuters/status/2082921695430467816
• https://techcrunch.com/2026/07/30/zoox-clears-final-federal-hurdle-to-launch-paid-robotaxi-service/
• https://x.com/Polymarket/status/2082825301642346515
• http://reut.rs/4wsmyfT