\n\n
Google is rolling out its Gemini model to replace the long‑standing Assistant on Android devices, beginning September 4, 2026, with a phased rollout that varies by device, country, and language. The move signals a shift toward tighter integration of Google’s own multimodal AI and improved privacy features, marking the end of the Assistant brand for most users.
At the same time, Anthropic announced a new internal chip design team aimed at creating custom silicon for its Claude model, reducing its dependence on external GPU makers such as Nvidia and AMD. This hardware strategy is intended to cut costs and boost performance as the company scales its AI services.
In the business arena, OpenAI projects annualised revenue of $70 billion by year‑end, up from about $50 billion in September, highlighting the rapid monetisation of its suite of models.
Regulatory developments include a U.S. policy shift where open‑weight models like Nvidia’s Nemotron and Meta’s Llama will be exempt from voluntary safety testing, underscoring a more permissive stance toward certain public models.
Finally, Tencent has made its Hy3 model widely available through multiple Tencent platforms, achieving the top spot on OpenRouter’s leaderboard and showcasing the growing competition in the generative AI market.
These developments matter because they illustrate the consolidation of AI services under major tech firms, the drive for hardware‑level efficiency, and the expanding economic scale of AI deployments—all of which will shape user experiences and regulatory debates. Watching the performance of Gemini on Android, Anthropic’s custom silicon rollout, and the market response to OpenAI’s revenue trajectory will be key indicators of the sector’s direction in the coming months. [[1, 2]]