9 results on this page
Several Chinese listed companies that announced large AI compute service agreements are now facing cancellations, weak execution, and regulatory scrutiny. The article explains why framework agreements often fail to become real revenue, highlighting customer verification, GPU delivery, payment terms, and termination clauses.
AI is increasingly used across the sales cycle to draft proposals, recommend discounts, and suggest next steps. The article argues that these recommendations should not become business decisions without governance, controls, and human accountability.
Google reportedly bid $10 million for Spirit Airlines’ internal enterprise data, including approximately 100 million emails and 500 million Microsoft Teams messages, in a bankruptcy asset auction. The deal signals the emergence of a market for acquiring sanitized corporate communications and work records to train AI systems and enterprise agents.
British AI-chip startup Fractile is reportedly in advanced talks to raise funds at a $6.5 billion pre-money valuation. The proposed valuation is more than six times its level three months ago, following a chip-supply deal with Anthropic.
Aegis is a runtime governance system that treats agent outputs as action proposals, then evaluates and authorizes them through a trusted policy layer before tool execution. In a sandbox evaluation, it recorded zero governed mock-tool applications and zero governed risky side-effect completions, though the authors caution that this does not establish general agent safety.
Zhongbo Juli is increasing its investment in Moonshot AI once again. The move underscores the continuing competition around AGI, computing limits, and engineering execution.
SpaceX approached AI coding startup Cognition AI about a potential acquisition, according to people familiar with the matter. The deal would have marked SpaceX’s second major acquisition in recent months as it seeks to strengthen its position in AI.
SkillEffect is a checked-lowering runtime that converts agent-generated tool programs into audited, memory-bounded implementations before execution. Its shared runtime supports six operator families and rejects adversarial proposals while improving completion under fixed memory caps.
Tencent is reportedly reshaping its multimodal AI direction from content generation toward contextual understanding and action in the real world. The shift appears closer to Liang Wenfeng’s emphasis on reasoning and execution, and farther from Fei-Fei Li’s world-model-oriented research approach.