Musk: The launch date for the first batch of AI satellites equipped with Nvidia chips has been set.

On local time Monday (24th), SpaceX CEO Elon Musk previewed on social media platform X that SpaceX's first batch of AI satellites equipped with NVIDIA chips will be launched for the first time in the fourth quarter of next year, with large-scale deployment expected by 2028. SpaceX had stated in its prospectus that the goal is to complete deployment by 2028. Prior to this, NVIDIA announced that SpaceX will deploy its new Vera CPU to power next-generation AI workloads for intelligent agents, thus extending the chip manufacturer's architecture from ground data centers to orbital computing infrastructure. Musk directly responded to this news on X. He stated that SpaceX collaborated with NVIDIA to design a Vera Rubin NVL72 system optimized for the space environment. He described Vera as the first CPU specifically designed for intelligent agents, stating that it will accelerate the orchestration, code execution, and data processing necessary for SpaceX's AI agents to act swiftly. Our design is simpler, cheaper, denser, and lighter than traditional racks, Musk added on X. The Vera itself is equipped with 88 NVIDIA-designed Olympus cores and supports memory bandwidth of up to 1.2 TB/s, claiming to complete tasks 1.8 times faster than comparable x86 processors. According to NVIDIA, SpaceX also plans to leverage NVIDIAs Vera Rubin platform to expand its AI infrastructure for Grok chat, Siasun Robot & Automation, and related services. Earlier this month, Musk told investors that SpaceX will acquire a "very large proportion" of NVIDIA's graphics processing units by 2027. Musk's frequent interactions with NVIDIA this year indicate a commitment to a "deep binding" between the two tech giants. Previously, SpaceX successfully went public on Nasdaq on June 12, with its stock price soaring on the first day and its market value surpassing $2 trillion, significantly increasing Musk's wealth. At this historic moment, NVIDIA congratulated SpaceX on social media, to which Musk responded that he looks forward to taking their "exciting collaboration to new heights." Earlier this month, Musk also committed during the company's first earnings call that SpaceX will build its AI services entirely on NVIDIAs systems, due to the superior performance of these products compared to their peers. We believe the Vera Rubin architecture is the best architecture. We believe it is the best AI computer, and we greatly value our close collaboration and partnership with NVIDIA on multiple levels, Musk said. So we are only partnering with NVIDIA. The market reaction to space data centers is currently divided. Supporters argue that the advantage of space data centers is that businesses can avoid complicated regulatory processes and strong community opposition. Critics point out that for the coming years, and possibly even longer, the operating costs of ground data centers will remain far lower than those of space data centers. Research firm Wood Mackenzie estimates that a 1-gigawatt orbital data center will cost $170 billion, more than three times the cost of traditional data centers. However, as rocket launch costs decrease, it is expected that costs will gradually decline, with SpaceX leveraging its Starship to promote this goal. This article is reproduced from "Financial Associated Press," author: Huang Junzhi; edited by GMTEight: Xu Wenqiang.