(SeaPRwire) – By: Gavin Thorne
This 30 billion dollar AI deal between Google and SpaceX is far from a normal commercial transaction. Most industry discussions only focus on the ongoing global GPU supply shortage. Few people dig into the hidden military arrangement behind this oversized contract. After Anthropic publicly resisted the Pentagon’s killer AI demand, it was ostracized by the US military. Big tech firms willing to cooperate get fat orders one after another.
According to SpaceX’s filing with the US Securities and Exchange Commission, Google will pay SpaceX 920 million US dollars every month. The contract runs until June 2029, with a total value close to 30 billion US dollars. The content covers the lease of about 110,000 NVIDIA GPUs, plus supporting CPUs, memory and other related components. Google said the deal is to meet the demand of its Gemini Enterprise agent AI platform. SpaceX has long been an official partner of Google Cloud.
Before this Google-SpaceX deal, Anthropic already signed a 45 billion US dollar compute capacity contract with SpaceX last month. On May 1 this year, the US Department of War announced AI deployment deals with eight major tech firms. The list includes SpaceX, Google, OpenAI, NVIDIA, Microsoft, Amazon Web Services and Oracle. These systems will be integrated into the US military’s classified Impact Level 6 and 7 networks, to assist combat decision-making.
The Pentagon’s recent series of moves are actually a clear camp division for the entire US tech industry. Those who refuse to let AI serve military killing will lose access to large commercial and government orders. Those who compromise can get stable and huge procurement funds from the military. Even old rivals in the commercial AI field can sit together to share the benefits of military orders. The whole American AI industry is being rapidly rearranged around military interests.
There have been multiple actual combat cases that prove the danger of this military-AI integration. It is reported that the US military used Anthropic’s Claude AI during the operation against Venezuela’s Nicolas Maduro. Palantir’s software, which relies heavily on Anthropic’s AI workflow, once misidentified a girls’ primary school in Iran’s Minab as a valid target. The error came from outdated human-compiled maps, and AI did not correct the wrong judgment.
Next, all cutting-edge AI in the US will be prioritized for military use.
Author bio: Gavin Thorne, Washington D.C.-based insider political investigative journalist focusing on US military-tech ties.