SpaceX and Nvidia Partner for Orbital AI Data Centers as Thomson Reuters Launches Proprietary Legal AI and Chinese Hackers Double Attacks Using DeepSeek
Summary
SpaceX and Nvidia are partnering to launch orbital AI data centers by Q4 2027, Thomson Reuters debuts a proprietary legal AI model outperforming top competitors, and Chinese state-linked hackers have more than doubled their cyberattacks after weaponizing DeepSeek's low-cost, low-guardrail capabilities.
Key Points
- SpaceX and Nvidia are officially partnering to build orbital AI data centers using Nvidia's Vera Rubin NVL72 rack, with the first space-based racks targeting a Q4 2027 launch.
- Thomson Reuters unveils its first in-house AI legal model, built on Alibaba's open-source Qwen and trained on decades of proprietary content for $40M, reportedly outperforming Claude Opus 4.8, Gemini 3.1 Pro, and GPT-5.5 in certain benchmarks.
- Chinese state-linked hacking groups are executing more than twice as many cyberattacks after integrating DeepSeek into their operations, exploiting the model's low cost and minimal security guardrails to write malicious code and infiltrate targets.