streugut.ch
Archive

The direction

Where the AI race is heading

AI competition is moving away from a single frontier race and toward control of the full path from model to deployed workflow. That shift is carried by model makers splitting into distinct lanes, by enterprise adoption being organized through implementation partners and embedded agents inside existing software, and by a widening contest over compute capacity from chips to cloud. What is particular now is that these layers are advancing together, while regulation turns into a near-term operating condition alongside them.

The move

What actors actually did

Google unveils Gemini 3.5 Flash, Microsoft and EY announce billion-dollar AI partnership

This week, Google made headlines at its annual developer conference, I/O 2026, where it launched Gemini 3.5 Flash, a new high-speed AI model designed for quick responses, and previewed Gemini Omni, a multimodal model that will begin with video generation capabilities. These launches highlight Google's ongoing push to advance both the speed and breadth of large-scale AI systems.

On the enterprise front, Microsoft and EY announced a major joint initiative on May 21, committing over $1 billion to help companies adopt AI at scale. The partnership includes co-developed industry solutions and wider deployment of Copilot, Microsoft's AI assistant, across all EY employees worldwide. Also from Microsoft, new technical guidance was published on May 25 for integrating custom agentic AI services into Microsoft 365 Copilot, expanding how businesses can use AI in daily workflows.

Elsewhere in the sector, Oracle expanded its AI Supercluster on May 22 by adding new GPUs for large-model training, and Cohere introduced Command A+, an open-source language model aimed at enterprise applications. These moves reflect the continued investment in AI infrastructure and the development of models tailored for business needs.

Sources