Why aren't Gemini models more competitive despite Google having data, compute, and being an early AI pioneer?

Despite Google's early AI pioneering, extensive data, and computing power, the competitiveness of Gemini models is questioned. The core issue might not be the base model itself, but rather deficiencies in product execution, including the API, tooling, reliability, user experience, pricing, and developer ecosystem. This raises questions about whether Google lacks in research capability, training strategy, organizational focus, or other critical areas compared to leading AI labs.