AI & Tech Daily Briefing — 2026-09-03
1. Nvidia agrees to buy Hugging Face for nearly $13 billion
Nvidia agreed to acquire Hugging Face for $12.93 billion, one of the chipmaker’s largest deals, giving it control of a major platform for sharing and deploying open AI models. Nvidia says Hugging Face will remain open to other hardware and cloud providers even as the deal deepens Nvidia’s reach beyond chips into the developer ecosystem.
Why it matters: The acquisition gives Nvidia a strategic foothold in the software layer used by millions of AI developers, helping defend its position as hyperscalers build competing chips and open-weight models gain ground against proprietary systems.
Sources: Reuters · Associated Press
2. Broadcom projects another surge in custom AI-chip demand
Broadcom forecast strong AI semiconductor sales over the next two years after its quarterly AI-chip revenue jumped sharply, reinforcing evidence that major technology companies are continuing to spend heavily on custom accelerators and networking infrastructure despite scrutiny over returns on AI investment.
Why it matters: Custom chips are becoming a significant alternative to general-purpose Nvidia GPUs for the largest AI buyers, potentially reshaping the economics and competitive structure of the AI infrastructure market.
Source: Reuters
3. China expands support for AI, robotics and quantum startups
China unveiled a five-year plan to channel more funding and policy support to specialized small and midsize technology companies, including startups working on embodied AI, robotics, quantum technology and brain-computer interfaces. Beijing aims to increase its group of nationally recognized “little giant” companies to 22,000 by 2030 while boosting R&D spending and access to long-term capital.
Why it matters: The plan shows China broadening its technology self-sufficiency drive beyond large national champions, building a deeper domestic supplier and startup base as U.S.-China technology restrictions intensify.
Source: Reuters
4. Abu Dhabi institute releases unusually open family of AI models
Abu Dhabi-based IFM released six K2 Horizon AI models together with their training data, code, methodologies and development checkpoints. The models span smaller systems designed for constrained devices through models with as many as 375 billion parameters, with the institute positioning full reproducibility as an alternative to both closed models and open-weight releases that withhold training details.
Why it matters: Full disclosure of training data and development artifacts could raise expectations for what “open-source AI” means, while strengthening the UAE’s effort to establish itself as a global AI research hub alongside the U.S., China and Europe.
Source: Reuters
5. London launches its first robotaxi service with Uber and Wayve
Uber began offering autonomous rides in London using self-driving technology from British AI company Wayve, making London the second European city where Uber offers robotaxis. The launch moves Wayve’s AI-based driving system from testing toward commercial deployment in one of Europe’s largest and most complex urban transport markets.
Why it matters: Commercial service in London is an important test of whether autonomous-driving AI can expand beyond tightly controlled deployments and navigate European regulatory requirements and dense city environments at scale.
Source: Reuters