“Despite developing its own chip, OpenAI does not plan to reduce its compute orders from Nvidia.”