Last night, Claude Opus 4.5 was released, boasting programming capabilities surpassing Gemini 3 Pro and GPT-5.1. In just one week, the AI community has completed a closed-loop iteration. The global coding throne has changed hands again.
Just a couple of days ago, I discussed Google's recent surge in popularity, the overwhelmingly positive reviews for the Gemini 3 release, and many's belief that Google's rise inevitably leads to Nvidia's decline and the inadequacy of other large-scale models. This misunderstands the significance of the competitive landscape: Firstly, I have high hopes for Google; its comprehensive AI industry chain layout promises excellent future momentum. However, the better Google's AI performs, the more anxious they are. It's Microsoft, Amazon, Meta, Apple, and OpenAI. Until the final competitive landscape is established, they will continue to increase investment to catch up with and even surpass Google.
Today, new media reported that Meta is planning to buy Google's TPUs, but it's unlikely that Microsoft, Amazon (a cloud computing provider), and OpenAI will purchase TPUs, as they are direct competitors, or they might make small purchases as supplements.
Looking ahead from the current point in time, the potential for growth is still enormous; the market penetration rate is far from exhausted (only 10%), and we are far from reaching a stage of zero-sum game. At the same time, technology iterates rapidly, and a single company's technological lead may only be maintained for months or weeks at a time.
twitter.com/qinbafrank/status/...
