The direction

Where the AI race is heading

AI competition is moving away from frontier models alone toward control of the stack that turns models into governed execution. That shift is being carried by agent environments tied to enterprise identity, data, workflow, security, and everyday work surfaces, and by model access that is increasingly routed through managed clouds, compatibility layers, owned product surfaces, and access controls. The infrastructure race underneath is widening into financing, datacenter capacity, packaging, custom silicon, and power. What is particular now is that controlled model access, cloud gatekeeper pressure in Europe, and gigawatt-scale buildout are tightening together, making control of execution a market-structure question as much as a technical one.

The move

What actors actually did

OpenAI’s GPT-5 Pro aids scientific breakthroughs while Google restricts Gemini access for Meta

OpenAI published a case study on June 23 showing how GPT-5 Pro was applied by immunologist Derya Unutmaz to resolve a longstanding scientific problem in immunology. This instance demonstrates the use of advanced AI models in complex research workflows by experts outside traditional AI fields.

Google, on June 28, imposed limits on Meta's access to its Gemini AI model, citing computing capacity constraints, and signed an agreement to rent additional cloud resources from SpaceX’s Starlink network. Meanwhile, Kunlunxin Technology, Baidu’s AI chip subsidiary, announced plans to go public in Hong Kong at a target $50 billion valuation and called on IPO investors to pre-purchase its semiconductors.

Microsoft expanded the infrastructure landscape by completing its first fully operational datacenter facility in Mount Pleasant, Wisconsin, reported on June 23. Mistral released OCR 4, a document intelligence model that supports 170 languages and advanced extraction features for enterprise deployment.

Sources