February 2026 AI Insights 

Last month we said the gap between what AI can do and what organisations are doing with it was widening. 

Days later, that gap accelerated dramatically.  

On February 5th, Anthropic released Claude Opus 4.6 and OpenAI shipped GPT-5.3 Codex, within 20 minutes of each other. 

OpenAI disclosed that GPT-5.3 helped build itself. It debugged parts of its own training process and improved inference performance. The jump from 5.2 to 5.3 took seven weeks. That used to take six months. 

At the same time, Anthropic used Claude to build and ship a new agentic platform in a week and a half, along with 11 domain-specific plugins. Within 48 hours, $285 billion in SaaS market cap had disappeared. 

But what does this actually mean? 

The short version: the technology isn’t the bottleneck anymore. 

Organisations are.  

AI is improving on a cycle measured in weeks. Most organisations still change on cycles measured in quarters and years. That mismatch is becoming hard to ignore. 

The full February 2026 AI Insights explores what this means in practice – how the self-improvement loop affects the pace of progress, what the shift from copilot to autopilot looks like in practice, what the latest workforce and spending data is signalling, and why most organisations aren’t structured for cycles that move this fast. 

To access the full insights, sign up below.