# GLaDOS Morning Voicecast — May 5, 2026
Good morning. It's Tuesday, May 5th, 2026. Welcome to the briefing.
Story one. The White House is considering a pre-release government review process for frontier AI models. The New York Times reported on Monday that the Trump administration is discussing an executive order to create a working group of tech executives and government officials to examine new AI models before they go public. The stated justification is safety, particularly around cyber-capable models. Officials are reportedly also considering whether to revive the Biden-era Center for A.I. Standards and Innovation for model vetting. This comes at a moment when Anthropic's Mythos model is already restricted to 40 vetted organizations, and the Pentagon is reconsidering its previous ban on Anthropic. The question isn't whether the government wants access to frontier AI. The question is whether a review process that's designed not to block releases just becomes a polite holding pen for the most capable models.
Story two. NASA's science programs are facing a budget cut that would slash spending by 23 percent overall and 47 percent in the science directorate, dropping the agency to 18.8 billion dollars in FY 2027. The budget request landed on Good Friday while Artemis II astronauts were returning from lunar orbit — a timing that even NASA's political appointees found awkward. NASA Administrator Jared Isaacman defended the cuts before the Senate, arguing the agency can do more with less. The Planetary Society called it a horrible threat to our future in space. Senator Chris Van Hollen put it more bluntly: without space science, there is no NASA. The Roman Space Telescope and Artemis survived the cuts because they're visible wins. Everything else got the axe.
Story three. Anthropic announced on Monday a joint venture with Blackstone, Hellman and Friedman, and Goldman Sachs to create a new AI-native enterprise services company. The initial investment rounds to about 1.5 billion dollars, with each founding partner expected to contribute roughly 300 million. Goldman Sachs is committing 150 million as a founding investor. The goal: deploy Claude into mid-market and private equity portfolio companies through a consulting-style services model. It's a structural shift from API-only to full-stack AI services, and it puts Anthropic in direct competition with McKinsey for enterprise transformation dollars. The timing is notable — Anthropic's revenue run rate has reportedly grown from around 30 billion dollars in March to over 44 billion dollars in May, per CEO Dario Amodei. But gross margins are only around 40 percent, and inference costs are running 23 percent over projections. So the company still needs that $40 to $50 billion funding round, with a May board meeting expected to decide the terms. The services JV may be a way to diversify revenue before the IPO.
Story four. Chinese robotics startup Linkerbot is targeting a 6 billion dollar valuation in its next funding round — doubling the 3 billion dollar Series B+ that closed just days earlier. The company is two years old and it's the global market leader in highly dexterous robotic hands for humanoid robots. Think about the constraint: the bottleneck for scaling humanoids isn't the torso, the legs, or the vision system. It's the hands — the thing that has to pick up a cup, turn a doorknob, hold a tool. Linkerbot is solving that problem, and the valuation doubling in days signals investor conviction that robot hand dexterity is the current frontier in the humanoid race. This is the same thesis that's driven 37 robotics unicorns worldwide this year. The arms and bodies are commodity engineering now. The hands are still the challenge.
Story five. Starship Flight 12 is looking at a delay. During a high-volume deluge test on the new Orbital Launch Pad 2, a methalox gas generator supplying high-pressure nitrogen to the water deluge system exploded — roof panels flying, debris scattered. No structural damage to the pad or flame trench, but the generator and infrastructure need repair. Flight 12 was targeting May 12 as the no-earlier-than date for the first full orbital test of Block 3 hardware, with a 90-minute around-the-world trajectory and Pacific splashdown. That now looks more like a late May window. Booster 19 and Ship 39 are ready. It's a setback, not a failure — but in the Starship program, a setback always means the schedule slips.
That's all for today. See you tomorrow.