Alphabet sheds half a trillion as cloud backlog hits $561B despite earnings miss

Google's worst trading day despite a cloud boom

Alphabet's stock dropped sharply after an earnings report the host called a "rare" miss, down 12 and a half percent. Revenue came in at $127.6 billion against a consensus of $128 billion, a minor miss, but EPS fell well short at $2.12 versus an expected $3.01, partly due to an $8.7 billion writedown on SpaceX equity. Even excluding that writedown, EPS would have been $2.67, still below estimates.

The host framed cloud as "a monster business," with revenue doubling to $30.7 billion and operating profit tripling to $11.1 billion, at a 36.2% operating margin. The cloud backlog stands at $561 billion, and the host noted that 55% of that is expected to be recognized as revenue within two years, calling it fast growth for a unit that currently does $30.7 billion in quarterly revenue.

Search advertising growth slowed to 12% from 17% the prior quarter, which the host attributed partly to AI Mode taking share from traditional search while monetizing less effectively. Capex guidance for 2026 sits at $195–205 billion, with the host saying Google is "signaling" a further significant increase in 2027. A viral thread from a self-described Google ads buyer, Max Anderson, argued search revenue growth is "artificial and extremely unhealthy" long-term, a claim the host said has been disputed but resurfaced around the earnings reaction.

Games as a path to AGI

The hosts discussed reporting suggesting internal reasoning gains at some labs come from aggressive reinforcement learning in games and simulated environments, with transfer effects showing up in research and mathematics. They noted Anthropic's internal "Claude" work moving in a similar direction and cited a claim that Elon Musk's AI hiring explicitly seeks transfer skills from games into general computer use. The host raised whether this validates Demis Hassabis's original vision for reaching AGI, contrasting Google's continued investment in its Genie project with Anthropic's narrower coding focus and OpenAI's mixed record on side projects like Sora and computer-use tools.

A Chinese spy plot allegedly targeting Jensen Huang

Citing a New York Times report, the hosts described allegations that a Chinese businesswoman, Alice Wong, tied to a Singapore-based data center company, developed a close relationship with Jensen Huang while allegedly funneling intelligence to Beijing, with scrutiny focused on whether advanced Nvidia chips were being steered to China. The report says President Trump was briefed roughly a year ago and "chose not to do anything with the information."

"It's very, very tricky... it seems like at the very least it's like she is being accused of being deployed to figure out how to gain access to advanced AI chips." — the host

The hosts referenced their own earlier coverage of the company, Megaspeed, and its CEO, noting her background in Chinese state media. They flagged Malaysia's rising compute capacity as a possible warning sign, comparing chip smuggling incentives to drone arbitrage reported during the Russia-Ukraine war, though they stressed Jensen Huang has not publicly responded and the underlying intelligence may be roughly a year old.

A viral "lost the house" trading video is actually an ad funnel

The hosts dissected a viral video titled "losing our house because of day trader," noting the odd phrasing and a dramatic, "saxophone getting louder" soundtrack designed to provoke outrage. Following the same account's subsequent videos, they found it leads to a "trading mastermind" funnel selling a free course, then a $49 course, then a $100 course.

"I don't know if they lost the house or not... it could be that they're doing better than ever because they're not actually trading at all." — the host

The hosts said this pattern reflects a broader category of apps offering simulated trading subscriptions, often around $100 a month, that promise real capital to users who perform well. They said some of these firms have been exposed as Ponzi-like schemes paying rewards from other subscribers' fees, and noted one such firm reportedly generated "hundreds of millions" in fees despite running what is effectively a fake trading app with no real market exposure.

Broader market turbulence

The hosts closed by noting wider market weakness alongside Alphabet's drop, with the Nasdaq down 1.3% and the Dow down 0.25%. They remarked that "best day ever" and "worst day ever" headlines are becoming increasingly common as market swings grow larger.