Key insights
- The author argues Palantir's valuation is unsustainable due to high stock-based compensation and reliance on expensive engineers. They compare Palantir to Accenture, suggesting it's more of a high-tech consulting firm than a pure SaaS company. The author raises concerns about dilution and the potential impact of AI on Palantir's business model, ultimately questioning its long-term viability and suggesting a bearish outlook.
I approach investing from a value perspective, so looking at PLTR’s 10-K with its 230+ P/E ratio gave me financial vertigo. But because I don’t trust my own judgments, I tried to strip away the "AI Revolution" narrative and just look at the physical plumbing of how the business actually makes cash.
When you look at their "forward-deployed engineer" model and the fact that 54% of their revenue comes from the government, it looks less like a pure SaaS monopoly and more like a high-tech defense contractor.
I ran the math comparing PLTR to a legacy body-shop like Accenture (ACN) and pure software like Adobe (ADBE).
Two things stood out:
-
The Dilution Cover-Up: Palantir's revenue per head is elite, but they achieve that by paying their engineers massive amounts of Stock-Based Compensation. Last year they issued $473M in SBC. More than 1/3 of their True Free Cash Flow is just adding that SBC back to net income. That is terrifying dilution.
-
The AI Paradox: If AI becomes so good that Palantir doesn't need expensive forward-deployed engineers anymore... doesn't that mean AI is so good that corporate clients can just write their own software? You can't have it both ways.
I am a big believer in Charles Goodhart’s demographic theories—companies reliant on expensive human labor are going to get squeezed. It feels like PLTR is masking its massive human labor costs by passing the bill directly to the retail shareholder via dilution.
I threw the historical FCF and SBC dilution charts into a post for my own sanity if anyone wants to see the visual math: https://open.substack.com/pub/cavemanscreener/p/the-349-billion-body-shop-why-palantir?r=29p94e&utm\_medium=ios
Am I completely missing a piece of the puzzle here, or is this just a massive momentum trap?