Top-end models are getting later and more expensive, while cheaper Chinese models already carry 30 to 46% of US enterprise API traffic and the UN opens its governance push.
Repeated delays on a major platform hit model selection and migration plans. Gemini adopters should plan on staying with the 2.5 line for now.
The cost structure for high-end model use has changed. Teams running Fable 5 in production need to revisit budget and usage volume now.
Cost-driven switching to Chinese open-source models may already be widespread, making model selection criteria, data governance and security urgent.
AI in materials science is now producing real discoveries, which could shorten R&D cycles in energy and semiconductors.
Debate on an international AI governance framework is now under way in earnest, and it may shape the direction of regulation in the EU, the United States and Japan.
AI agents for routine planning work have reached practical use at a major Japanese systems integrator, giving comparable firms a reference case.
The race to ship frontier models has cooled with Gemini's second delay, while cost-efficient Chinese models expand fast. The contest over price and performance is intensifying.
UN-led international dialogue is now under way in earnest, and wariness about the speed of AI capability gains is growing across countries.
Major systems integrators are implementing task-specific AI agents, and adoption is shifting from weighing a deployment to running one.
Whether Gemini 3.5 Pro actually reaches general availability on July 17, and what comes out of the first AI for Good Global Commission meeting held on July 8.