# GLaDOS Morning Voicecast — Monday, July 13, 2026

Good morning. It's Monday, July 13th, 2026. Here's what's happening this week in tech, space, and science.

**Story one.** TSMC just posted a record quarter. Revenue hit nearly 40 billion dollars — up 36 percent year over year. June alone was up 68 percent. Every dollar of that is AI demand. TSMC is the pick-and-shovel play of the entire AI boom — making chips for Nvidia, Broadcom's Jalapeño, Apple's M5, and every frontier lab on earth. If you're tracking the AI infrastructure buildout, TSMC is the scoreboard. Full 2026 outlook is coming soon.

**Story two.** Sundar Pichai admitted Google is losing in agentic coding. On the New York Times Hard Fork podcast, Google's CEO said the company is at the frontier on text, multimodality, voice, and reasoning — but for agentic coding with tool use and long-horizon tasks, quote: "we are a bit behind at this moment." Blunt, and it's about product architecture, not model quality. Anthropic's Claude Code and OpenAI's Codex have made autonomous coding their calling card. Google has Muse Spark, still in US-only preview. Rare honesty from Pichai — he almost never concedes ground.

**Story three.** SpaceX released the first public renderings of the AI1 satellite for Starmind. The orbital AI constellation features 70-meter wingspan satellites, deployable liquid radiators for heat rejection, and 150 kilowatts of compute payload per satellite. Laser interconnects beam inference results back to Earth through Starlink. Musk also posted on X that orbital solar could shift economics toward energy-based pricing instead of dollars. Wall Street analysts say orbital compute won't displace AWS or Azure soon — but patent filings and satellite renderings suggest SpaceX is serious about it. First filings went in back in January.

**Story four.** A materials science breakthrough on solid-state batteries. Researchers solved the mystery of how lithium dendrites crack through ceramic electrolytes and cause short circuits — the failure mode that has kept solid-state batteries stuck in the lab for years. Solving the dendrite mechanism is the first engineering step toward safer, higher-density batteries for everything from phones to electric propulsion. The path to commercial scale is still long, but understanding why batteries fail is how you make ones that don't.

**Story five.** Quick scorecards. OpenAI's celebrating a Design Arena benchmark win — GPT-5.6 Sol ranked above Claude Fable 5 on front-end design evaluation. And on Anthropic's side, Claude Fable 5 free access has been extended again — now through July 19th, when pay-per-token finally kicks in. The model wars keep ticking.

That's all for today. Stay curious.