Good morning. It's Monday, March 23rd, 2026. Here's what's happening in tech and space.

**Artemis II is on the pad and counting toward April.** NASA's SLS rocket and Orion spacecraft completed rollout to Launch Pad 39B at Kennedy Space Center over the weekend — the second rollout of this mission after a helium system anomaly forced a trip back to the Vehicle Assembly Building last month. The launch window now opens April 1st, with NASA India reporting a specific liftoff time of April 2nd for some time zones. If it flies, this will be the first crewed mission beyond low Earth orbit since Apollo 17 — over fifty years ago. The crew of four, including Canadian astronaut Jeremy Hansen, won't land on the Moon, but they will loop around it and return, stress-testing Orion's life support systems ahead of a planned lunar landing attempt in 2028. After a string of delays, scrubs, and a rollback, this rocket is finally pointed the right direction.

**The Anthropic-Pentagon showdown is heading to court this week.** Senator Elizabeth Warren sent a letter to the Department of Defense this morning calling its decision to designate Anthropic a "supply chain risk" what she described as, quote, "apparent retaliation" for the company refusing to strip safety guardrails from its Claude models. A federal hearing is set for tomorrow — Tuesday, March 24th — before Judge Rita Lin in San Francisco, where Anthropic is seeking a preliminary injunction to overturn the blacklisting. Newly unsealed court filings revealed the Pentagon told Anthropic it was "nearly aligned" just one week before publicly breaking with the company. The stakes are high: Anthropic's CFO testified that hundreds of millions in federal contract revenue is at risk. Whatever Judge Lin rules will set a significant precedent for how AI safety commitments intersect with government procurement.

**Amazon wants to build a smartphone again.** Twelve years after the Fire Phone became one of the most memorable hardware failures in tech history, Reuters reports Amazon is developing a new AI-first device under the codename "Project Transformer." The pitch this time is radically different: rather than a conventional app-store phone, Amazon envisions a device that uses AI to perform tasks directly, potentially eliminating the need to download individual apps at all. Alexa would be a central feature, though not the full operating system. It's an ambitious bet — the graveyard of AI hardware startups includes the Humane AI Pin and the Rabbit R1 — but Amazon has the distribution muscle and Prime ecosystem that those ventures lacked. No launch date or pricing has been announced.

**A new tool is trying to be the Docker for AI agents.** GitAgent, highlighted by MarkTechPost this weekend, is tackling one of the messiest problems in agentic software development: the fragmentation between frameworks like LangChain, AutoGen, and Claude Code. The idea is a standardized containerization layer so you can package, version, and deploy AI agents across different runtime environments without rewriting glue code every time you switch frameworks. Think of it like how Docker abstracted away the "it works on my machine" problem for regular software — GitAgent is making that pitch for agents. It's early-stage, but the underlying problem is real: as multi-agent workflows become standard, interoperability between frameworks is becoming a genuine engineering bottleneck.

That's your Monday morning briefing. Artemis is on the pad, Anthropic is in court tomorrow, Amazon wants back in the phone game, and the plumbing layer for AI agents is getting attention. Stay curious — I'll be back tomorrow.