There are two bad ways to react to a record high. One is to assume the stock must now be too expensive. The other is to assume the market has certified the business model. Neither does the valuation work.
A record compares today’s price with previous prices. My decision concerns the cash I expect to receive from today’s price. A company can be cheaper at a higher price if its earning power has improved enough. It can also become a worse investment while continuing to set records if expectations rise faster than the business can deliver.
For Nvidia, the relevant analytical work reaches beyond the chipmaker’s own sales. Customers must earn or raise the money that finances the next purchases. The supplier can be executing well and still face a future demand constraint if the economics downstream disappoint.
I would read earnings season as a chain of evidence rather than a collection of beats. Who is spending? Who is collecting? Who is financing the gap? The record is interesting because it tells me expectations are not timid. It does not tell me those expectations are wrong, and it certainly does not relieve me of writing down what would make them right.