Valid point on the immediate cash burn, but you could argue this is the cost of a paradigm shift.
It's a land-grab scenario. If they achieve AGI or even a dominant utility layer (like an OS for AI), that $200B might look like a cheap acquisition cost for a trillion-dollar market. The question isn't if the business model works today, but if the product utility will eventually force the market to pay for it at scale. High-risk, high-reward hyper-scaling.