If valuations hadn't soared for high-growth software firms the way they have this year, I think there's a good chance that Palantir Technologies would've waited longer to file for an IPO.
Though there are some clear positives within Palantir's IPO prospectus, the story told by it is a messier one than what's delivered by, say, Unity Software or Snowflake's recent IPO filings. And even in a market environment like this one, this story might temper investor enthusiasm some.
Palantir, which was founded 17 years ago, has long sold an advanced data-mining and analytics software platform (it's known as Gotham) that's used by various law enforcement, military and intelligence agencies. And a few years ago, the company rolled out a data-integration and analytics platform that's used by enterprises and to a lesser extent government agencies (it's known as Foundry).
Palantir provides its software to a relatively small base of large enterprises and government agencies: It claimed just 125 customers for the first half of 2020. On the other hand, customer loyalty and product stickiness are generally high -- Palantir notes that its top 20 customers in 2019 have been with it for an average of 6.6 years -- and so is Palantir's average revenue per customer ($5.6 million in 2019).
Revenue rose 25% in 2019 to $743 million, and a more impressive 49% in the first half of 2020 to $481 million. But it's far from clear whether Palantir's first-half growth rate is sustainable for an extended period of time.
Though Gotham has been around much longer than Foundry, Palantir's strong first-half growth was fueled primarily by a 76% annual increase in government revenue to $257.7 million. Commercial revenue rose a more modest 27% to $223.5 million.
The government revenue growth appears to have much to do with a 70% jump in Palantir's government contract backlog in 2019, to $1.14 billion (a big U.S. Army contract looks like a major contributor here). In the first half of 2020, the backlog rose just slightly to $1.16 billion.
Along with questions about how its government growth will trend going forward, expect Palantir to field its share of questions about its cost structure.
On a GAAP basis, Palantir generated a $576.4 million operating loss in 2019, and a $169.3 million op. loss in the first half of 2020. Excluding stock comp, Palantir generated a $325 million op. loss last year, and (with the help of that 49% revenue growth) a modest $17 million op. profit during the first half of 2020.
These bottom-line pressures existed even though Palantir's GAAP operating expenses rose just 2% in both 2019 and the first half of 2020. In spite of such cost controls, not to mention how long Palantir has been in business, sales/marketing spend equaled 61% of 2019 revenue. R&D spend equaled 41% and G&A spend somehow equaled 43%.
The sales/marketing spend can partly be explained by the fact that (unlike, say, Zoom (ZM) and Atlassian (TEAM) , which have more viral, bottoms-up, sales models) the adoption of Palantir's software at a company or agency depends heavily on convincing high-level decision-makers to sign off on large contracts. But overall, Palantir's opex is something that's likely to give many institutional investors pause.
Throw in a three-class share structure that guarantees Palantir's founders will maintain control of the company, and investors have a lot of reasons to tread carefully here. Palantir has a lot of the markings of a company that's eager to go public because it hopes the current investor fervor for enterprise software firms, together with a recent revenue growth inflection that might not be sustainable, will yield a much less skeptical investor reaction than what it might otherwise get.