Palantir and the great revenue mystery – TechCrunch
As I write to you on Friday afternoon, the Palantir S-1 has yet to drop, but TechCrunch did break some news regarding the impending filing and just how big the company actually is. Please forgive the block quote, but here’s our reporting:
In screenshots of a draft S-1 statement dated yesterday (August 20), Palantir is listed as generating revenues of roughly $742 million in 2019 (Palantir’s fiscal year is a calendar year). That revenue was up from $595 million in 2018, a gain of roughly 25%. […] Palantir lists a net loss of roughly $580 million for 2019, which is almost identical to its loss in 2018. The company listed a net loss percentage of 97% for 2018, improving to a loss of 78% for last year.
A few notes from this. First, those losses are flat icky. Palantir was founded in 2003 or 2004 depending on who you read, which means that it’s an old company. And it was running an effective -100% net margin in 2018? Yowza.
Second, what the flocking frack is that revenue number? Did you expect to see Palantir come in with revenues of less than $1 billion? If you did, well done. After a deluge of articles over the years discussing just how big Palantir had become, I was anticipating a bit more (more here for context). Here are two examples:
- Reporting from TechCrunch that Palantir expected “more than $1 billion in contracts” in 2014
- Reporting from Bloomberg that Palantir had “booked deals totaling $1.7 billion in 2015”
Notably, Palantir’s real revenue result, or one very close to it, made it into Business Insider this April. The reporting makes the company’s S-1 less of a climax and more of a denouement. But, hey, we’re still glad to have the filing.
The Exchange will have a full breakdown of Palantir’s numbers Monday morning, but I think what Palantir coverage over the years shows is that when companies decline to share specific revenue figures that are clear, just presume that what they do share is misleading. (ARR is fine, trailing revenue is fine, “contract” metrics are useless.)
