AI race intensifies as Alibaba faces distillation accusations and Microsoft launches its own models

Anthropic has accused a group linked to Alibaba of large‑scale model distillation, using up to 25,000 fake accounts and VPNs to query its Claude system from April to June 2026. The operation generated roughly 28.8 million interactions, focusing on high‑value tasks such as software programming and AI reasoning, and is described as one of the biggest detected distillation campaigns.

Meanwhile, Microsoft is moving away from reliance on external AI partners by introducing in‑house models aimed at enterprise customers at lower costs than offerings from OpenAI, Anthropic and Google. CEO Satya Nadella downplayed extreme AI risk warnings while the company plans a $190 billion investment in data‑center infrastructure, including a 20‑year power‑supply agreement with Chevron for a Texas facility expected to operate by 2028. Microsoft’s strategy seeks to revive investor confidence after a significant share‑price decline and to boost Azure’s growth.

Both developments highlight escalating competition in the artificial‑intelligence sector, as firms pursue greater control over model development and seek to mitigate risks of technology leakage.