Tech Earnings Quadkill, Red Button vs Blue Button | Diet TBPN
TL;DR
Big Tech earnings were framed as an AI capex stress test — with Google, Amazon, Meta, and Microsoft reporting in one afternoon and representing just under 20% of the S&P 500, the real question was whether $100B+ annual spend per company is turning into durable AI revenue before depreciation catches up.
Microsoft had the cleanest enterprise AI monetization story going in — the hosts highlighted Azure at 39% growth, Microsoft Cloud at 26%, and a startling $625 billion in remaining performance obligations, with roughly 45% tied to OpenAI, as proof that huge compute deals are already getting locked in.
Meta looked least vulnerable to the 'diffusion problem' — unlike enterprise software rollouts, Meta can push better models straight into ad targeting across 3.58 billion DAUs, where even slightly better ad placement can translate into 'another $20 billion in profit.'
Amazon’s underappreciated ad machine helps fund the AI buildout — alongside AWS growing 24%, Amazon’s ads business hit $21.3 billion for the quarter, putting it on a near-$100 billion annual run rate and giving the company real cash flow cover as capex heads toward $200 billion in 2026.
The show used a viral red-button/blue-button thought experiment to expose audience psychology — while Tim Urban’s 24 million-view post and MrBeast’s 10 million-view version both saw blue win with 55-58%, the TBPN chat skewed red, turning the debate into a live argument about rationality, trust, and moral framing.
OpenAI jitters were hanging over the market even as AI names bounced around — the hosts cited Wall Street Journal concerns that AI investments may not produce blockbuster profits, noting Oracle, CoreWeave, and SoftBank sold off on Tuesday before CoreWeave rebounded 8.4% the next day.
The Breakdown
The AI Earnings 'Quadkill' Setup
The episode opens in full market-circus mode: Fed decision at 2 p.m., Powell at 2:30, then Google, Amazon, Meta, and Microsoft earnings at 4 p.m. The hosts call it the "Super Bowl for big tech," noting these companies make up just under 20% of the S&P 500 and that the central question is simple but brutal: is the AI buildout actually paying off?
Why Google’s Search Economics Matter More Than Ever
Google gets framed as the most vertically integrated AI company: DeepMind, TPUs, Search, YouTube, Android, Cloud, Workspace — the whole flywheel. But investors aren’t just looking for cool product demos; they want to know whether AI Overviews and Gemini expand search usage and ad ROI fast enough to avoid compressing the economics of the core search business.
Microsoft as the Best Read on Enterprise AI Reality
The hosts zero in on Microsoft because it offers the clearest window into how AI moves through real businesses, not just hype cycles. The headline numbers are strong — revenue up 17%, Microsoft Cloud up 26%, Azure up 39% — but the eye-popper is $625 billion in RPO, up 110%, with about 45% tied to OpenAI, which they treat as evidence that giant compute demand is already contractually real.
Copilots, Seats, and the Future of SaaS
From there the discussion gets more nuanced: are companies actually paying more for Copilot, GitHub Copilot, and M365, or does AI eventually reduce seat counts instead of expanding them? One host jokes about a future where one M365 seat controls 20 agents, while another compares GitHub Copilot’s quieter momentum to Microsoft Teams beating the more "hyped" Slack — not sexiest product wins, distribution often does.
Amazon’s Real Test: AWS Reacceleration
Amazon’s setup is all about whether AWS growth can accelerate enough to justify the giant capex wave. The hosts cite Q4 numbers — net sales up 14%, AWS up 24%, ads up 23% to $21.3 billion, operating income at $25 billion — and stress that the market really wants AWS growth to start with a 3, because then Amazon looks less like it overbuilt and more like it bought scarce compute capacity ahead of demand.
Meta’s Advantage: No Diffusion Problem
Meta gets the most confident treatment because, in their view, AI is already visibly working in the business that matters: ads. With 3.58 billion DAUs, revenue up 24% to nearly $60 billion, and Family of Apps operating income at $30.8 billion, the hosts argue Meta can turn model improvements into cash almost instantly through ad ranking and placement — unlike enterprise AI, there’s no long rollout, training, or procurement cycle.
Wall Street’s Nerves and the OpenAI Overhang
Even with all that excitement, they bring in a Wall Street Journal note about renewed AI worries after OpenAI-related headlines rattled names like Oracle, CoreWeave, and SoftBank. The vibe is that the market leash is "very tight": any fresh doubt around OpenAI, Anthropic, or demand durability could trigger a sell-off, even if companies like IBM, Intel, and Texas Instruments have recently posted solid results.
The Red Button vs. Blue Button Fight Takes Over the Show
In the back half, the episode swerves into a viral Tim Urban / MrBeast thought experiment: if over 50% press blue, everyone lives; if under 50% press blue, only red-button voters survive. The fun is in the framing — is blue the heroic faith-in-humanity choice or is red the only rational move? — and the hosts milk the tension when TBPN chat tilts red even though Tim Urban’s audience went 58% blue and MrBeast’s went 55.7% blue.
Rationality, Morality, and a Weirdly Revealing Poll
What starts as a joke becomes a miniature philosophy seminar about trolley problems, genocide framing, and whether "hyperrational" behavior is actually moral. They close on survey data from 14,000 Americans showing blue won by a 3-to-1 margin, with the most predictive personality statement being "I tell the truth" — a perfect oddball ending for a show that started with trillion-dollar AI economics and ended with a referendum on human nature.