AI Developers Lag Behind Market Expectations

We’re officially in the era of “the AI ate my homework”—at least if you’re in the business of making AI, Andrew Nusca reported for Fortune Tech.

DeepSeek had to apologize for delays in releasing its much-anticipated R2 model.

Months after Meta said its latest Llama model wasn’t ready and OpenAI delayed the launch of GPT-5, Chinese rival DeepSeek finds itself in a similar position, apologizing for delays in releasing its much-anticipated R2 model.

The official reason relates to compatibility with domestic Huawei Ascend chips, rather than Nvidia’s more widely preferred—but trade-war-constrained—GPUs. Still, Nusca noted that all AI firms face the same challenge: racing to meet public expectations that consistently run a few steps ahead.