Reflections on Palantir

Always forward multiples, never trailing ones. Palantir likely trades on Enterprise Value / NTM Revenue (next 12 months).

Don't just take the average provided by something like Yahoo Finance. You need to look at which analysts are providing estimates, decide which of those analysts are reliable (e.g. a Bank of America analyst can be trusted, a Morningstar bot that writes research reports cannot), write down all their estimates, take either the mean or average

Because few analysts provide quarterly estimates, you need to use annual estimates instead. But the next twelve months are going to be made of some part of 2024 plus some part of 2025. Palantir's fiscal year is 12/31/2024 so it's a bit less annoying to calculate.

Their most recently reported quarter was Q2 2024, so the next 12 months = Q3 2024 + Q4 2024 + Q1 2025 + Q2 2025[1].

Then you have to calculate enterprise value, which is easier said than done. In a nutshell, it's the total equity value + debt - cash, but there are always minor things to adjust. Equity value is the number of diluted shares outstanding[2] multiplied by today's share price. To calculate diluted shares, you will need to know the options that are outstanding on the company and use the Treasury Stock Method to assume all of the in-the-money options are exercised, with the proceeds from those options being used to buy back shares. Debt you can get from financial statements, unless the company has publicly traded debt in which case you might need to adjust for its current value rather than its book value. Cash you can simply get from financial statements, but there can be issues there too depending on how complex the company is. Add all of that together (subtract cash!) and you get Enterprise Value.

Divide Enterprise Value by NTM Revenue and you'll get a revenue multiple for this company today. But if you want to calculate what the company _should_ be worth relative to competitors, you can do the same thing for all of its competitors, then take the mean/average EV/Revenue of those comps and say "PLTR should be worth this much"

Also separately you can build a DCF if you have sufficient visibility into the future cashflows of the company.[3]

You can take some shortcuts or go even deeper in all of the above. It comes down to how much scrutiny you need for the investment you're making. Are you SAP trying to acquire Palantir? You're going to do all of the above with more detail than I explained. Are you deciding whether to rebalance a bit of your portfolio out of Palantir as an individual trader? Maybe Yahoo Finance Pro estimates are serviceable enough (I wouldn't know).

OR just find an analyst whose views on the company you happen to like and who you think is generally right and look at their multiples so you don't have to do all that legwork yourself. But you'll need to be a client at their bank to get access to their research...

----

[1] Some people like to do (days left in 2024 / 365) * FY 2024 estimates and take the remaining days to make up a year * FY 2025, but that's totally wrong for many reasons, the most obvious being that investors aren't updating their models (and thus the valuation multiples those models output) on a daily basis. There's no new news about the company every single day, so estimates should be stable over the course of the quarter.

[2] NOT from the earnings report, as that "diluted shares" for EPS means something else: to simplify, it means diluted over the course of the year rather than today, which is what we want.

[3] For fast growing companies, this is harder because you need to extrapolate all the way until you get to a year with relatively low growth cash flows in order to get to a "terminal year" for a DCF analysis, but if you're projecting 10-20 years into the future, chances are you're wrong!

My pleasure! Wall Street likes to gatekeep this info (it's very simple math but banks charge millions for it) and there's a disheartening shortage of publicly available repositories with this knowledge (most of it can be automated, except for one-off adjustments you need to make for each company here and there for accounting reasons or out of the ordinary occurrences)

The bit I forgot to add is that you kinda have to do the reverse too, if you're valuing the company based on comparables: take their mean multiple, then apply that PLTR's forward revenue to get to some enterprise value, then subtract net debt (i.e. minus debt _plus_ cash now!) and get to equity value. Then divide by the diluted shares (you have to imply the Treasury Stock Method dilution in some somewhat circular Excel math) to get to a final dollar value per share

You can take this one step further and draw line charts over time with these multiples vs. comparables to see how the sentiment has changed for this stock (or for comparables) over time. And many other similar analyses...