OpenAI is facing challenges as its daily active users are stagnating and subscription revenue growth is slowing down, potentially causing it to fall short of its 2026 revenue targets. The company might become emblematic of an AI infrastructure bubble, with a significant amount of infrastructure expected to be online by 2026 that may not be fully utilized. This includes over 45 ZFlops of FP16 accelerated compute, which is more than enough to meet future model training and inference demands, especially as compute costs continue to decrease. The situation draws parallels to the peak of Yahoo in 2000, suggesting that OpenAI might currently be at its zenith. This matters because it highlights the potential risks and overestimations in the AI industry's growth projections, impacting investments and strategic planning.
