Fair value is a slippery concept.

For my own use, I define fair value of any financial investment as the price I would have to pay today to get inflation + 6.5%/year in the next 5-10 years. Historically in the US that was the "monkey with a dartboard" rate: roughly what the average stock returned in the average year.

Let's imagine Berkshire's intrinsic value per share grows at that same rate in the next decade or so. That's a bit lower rate than history which has been very close to inflation + 8% for ages, but Berkshire is a bit bigger than historically, and the market is more richly valued giving fewer opportunities, so it might be a pretty good guess. Another factor is that it's pretty certain Mr Buffett won't be the one allocating the capital by then.

To the extent that one is OK with all the reasoning above, the fair value of a share today is exactly equal to the valuation multiple that you expect to see as typical 5-10 years from now. If the next 10-20 years resembles the last 10-20 years, and book per share remains as good as ever as a quick yardstick of value, then fair value today is around 1.4 times book today.

If they grow book at 10% year, we could still do well if the stock stays at a 1.6x ish valuation.

Certainly true, but I personally would not bank on valuation multiples being that high. The average valuation multiple in the last 3 months is the highest 3-month average since 2008. I think it's rather more likely that we've been seeing a transient high spot. No matter, we'll all know soon enough.

Jim