Good morning. It's Tuesday, May 26th, 2026. Here's what's happening in tech and space.
The SpaceX S-1 prospectus reveals something investors weren't expecting: the company warns that current AI chip supply chains may be insufficient for its orbital compute ambitions. The filing lays out a $119 billion chip manufacturing risk — that's Terafab, the joint Intel facility in Austin that Tesla, SpaceX, and xAI are banking on. During Tesla's Q1 earnings, Musk put it bluntly: "Either build Terafab, or lose the future."
But here's the context: Starlink alone generated over eleven point four billion dollars in revenue last year, with more than a billion dollars in operating profit in Q1 2026 — and zero churn among major enterprise customers since 2023. The satellite communications business is printing money. The launch business is strategic. And the AI division, which includes xAI and Grok, is burning through six point four billion annually.
So the IPO filing isn't just about rockets. It's about building the entire compute stack — from silicon fabs to orbital data centers. The filing explicitly warns that if Terafab can't deliver, the orbital AI compute plans scheduled for 2028 could stall. With a targeted IPO valuation of one point five to two trillion dollars, that risk is very much on the table.
Just a few hours ago, NASA held a news conference at 2 PM Eastern to share details on plans for a sustained lunar surface presence. This follows the Artemis II crewed flyby in April and the recent delay of Artemis III to late 2027.
The key shift: Artemis IV, now targeted for 2028, is being designated as the first crewed lunar landing — essentially performing test missions with the Starship HLS and Blue Moon landers in Earth orbit before committing to a lunar descent.
NASA is also looking at a major organizational restructuring under administrator Jared Isaacman. The agency has been under pressure on both budget — facing proposed cuts that would hit science programs hard — and on execution credibility after the Artemis delays. Today's moon base briefing appears to be laying out a clearer roadmap to reassure stakeholders that there's a real plan for staying on the Moon, not just visiting.
Kore.ai released Artemis on May 21st, and it's generating serious enterprise buzz. It's an AI-native platform for building, governing, and optimizing multi-agent systems at scale — launching initially on Microsoft Azure, with Google Cloud and AWS versions coming in Q4.
What makes it interesting for our space: Kore.ai is positioning this not as another chatbot wrapper, but as compiled, governable multi-agent orchestration with guardrails baked in before agents hit production. That's the enterprise gap — everyone can deploy agents, but nobody trusts them unsupervised. Artemis tries to solve that with governance as a first-class feature.
The company is already listed as a leader in Forrester's Agentic AI PEAK Matrix assessment. Competition is heating up fast: ServiceNow has Otto and its Action Fabric, and Microsoft's Agent 365 went general availability earlier this month. The enterprise agentic AI platform war is officially a multi-player race.
That's all for today. Have a great Tuesday.