Tesla (TSLA) CEO Elon Musk says his approx guess is that the Terfab AI compute output would be about 25% for Tesla Optimus and about 75% for AI spacecraft

Offhand allocation figures from this CEO have a long record of being directional rather than binding: past remarks on production targets, timelines and capital split have frequently been revised, and the market has learned to treat approximate guesses as signalling intent rather than guidance. The read here is strategic weighting, with the bulk of dedicated compute earmarked for a space-linked AI effort and a minority for the humanoid robot programme, which speaks to how management internally ranks its option set beyond the core vehicle business. Episodes of this kind have tended to matter less for the immediate print than for how they feed the standing debate on whether the equity is being valued as a carmaker or as an AI and robotics platform, since that framing drives the peer set and the multiples applied. Worth noting is the governance dimension: compute is a scarce, expensive input, and commentary about splitting it across ventures touches recurring questions about resource allocation between the listed entity and adjacent Musk-controlled projects, a friction point that has surfaced before. The follow-ons are whether any of this is formalised in filings, capex disclosure or earnings-call detail, where hard numbers would replace guesswork. As a single informal remark, the signal is soft.