News
What's happening in robotics and AI — curated by the wujec.ai editors.
SpaceX spends more on AI than on rockets: 15.8 of 18.4 billion dollars in one quarter
SpaceX published its second-quarter results on 4 August 2026 and filed them with the US Securities and Exchange Commission on form 8-K. The headline number is not the revenue but the capital expenditure: the company spent 18.369 billion dollars in three months, and 15.828 billion of that — 86 percent — went into artificial intelligence. A launch company now puts more money into compute than into everything else it does combined. The rest of the release supports the same reading. Revenue reached 7.8 billion dollars, up 92 percent from 4.1 billion a year earlier, and the AI segment alone brought in 2.561 billion — behind connectivity at 4.291 billion, but nearly three times the 962 million from the space segment that the company was built on. Adjusted EBITDA rose 191 percent to 3.5 billion. The net loss narrowed to 541 million dollars from just over one billion, an improvement of 467 million. SpaceX reported 100 billion dollars in cash and marketable securities, a backlog of 47.5 billion, and nameplate compute of 1.4 gigawatts at the end of the quarter. On the earnings call the same day, Elon Musk is reported to have said the company will build its AI infrastructure exclusively on Nvidia hardware and to have given targets of more than two gigawatts of capacity by the end of 2026 and roughly ten gigawatts by the end of 2027 — though the figures quoted for 2027 vary between outlets. Those statements come from press coverage of the call, not from the filed release, and no transcript has been posted; they should be read accordingly. The market reaction split across the day: the shares rose during the session and gave the gain back after hours, closing the after-market at 114.60 dollars. What this catalogue takes from the filing is a single figure with a long shadow. A quarterly AI capital expenditure of 15.8 billion dollars is spent by a company whose core business is launching things into orbit — a measure of how far the compute build-out has spread beyond the firms that sell models.
AMD goes after Nvidia's grip on robot brains
AMD used its Advancing AI 2026 event on 23 July to enter physical AI in earnest, announcing Kria AI system-on-modules built around the new Ryzen AI Embedded X100 processors, together with a robotics developer platform that pairs the module with a carrier board and evaluation kit. The hardware is aimed squarely at the compute box inside a humanoid or an autonomous machine: up to sixteen Zen 5 cores for real-time control, an RDNA 3.5 integrated GPU, an NPU for inference and, in the top configurations, up to 128 GB of unified memory. AMD quotes more than 8,000 control decisions per second and vision-language-action reasoning below 100 milliseconds, and benchmarks the platform against Nvidia's Jetson line — claiming 3.4x better real-time reliability, 1.6x more free CPU cores and support for 2.3x more concurrent agents than the Jetson T5000. Independent measurements are not yet available; these are vendor numbers. The strategic point is less about any single figure than about choice. Nvidia's Jetson Thor has been close to the default brain for humanoid developers, and AMD is selling an open stack — Linux, ROCm, a hypervisor, PyTorch and ONNX — to robot makers who would rather not be locked to one vendor. The modules are sampling with early customers now, with general availability through ODM partners expected in the fourth quarter of 2026.