A Cheaper AI Alternative to Palantir (PLTR)

The first two years of this AI-driven bull market have minted a handful of clear winners (like Nvidia (NVDA)) based on a consensus narrative.

Namely, that growth in and widespread adoption of artificial intelligence tools would require a large number of data centers, which would guzzle (nuclear-generated) electricity, and which would be running on innumerable servers containing dozens and dozens of high-priced graphic processors (GPUs).

But the arrival and adoption of smaller, fine-tuned models (like DeepSeek, which drew countless headlines last month) present an alternative future, where large language models are increasingly commoditized and the size of the model matters less than what a company can do with it.

The biggest beneficiaries of the former narrative were the tech hyperscalers like Meta Platforms (META), Microsoft (MSFT), and Alphabet (GOOG), all of whom could afford to spend a veritable fortune on GPUs.

After all, when table stakes for the AI game are spending billions of dollars a year on capital expenditures, you won’t find too many companies that can afford to ante up.

The second scenario, which is more consistent with how tech has traditionally operated (a smaller, smarter company devises a better mousetrap and shakes up the preexisting hierarchy), opens the table up to a lot more competitors.

After all, if fine-tuned models are now the name of the game, you don’t need to buy your own data center, all you need to do is pay Meta to let you use their models and server capacity to build your own smaller, smarter model.

Palantir is a notable winner of the downstream pivot, at least based on share price (up 228% in the last six months, even after taking a 10% haircut on rumors of defense spending cuts) and public perception. The company is a provider of advanced software and analytical/security platforms for government entities (especially U.S. and friendly militaries).

And BigBear.ai is increasingly drawing comparisons to its much larger peer—both for its share price performance (up 465% in the last six months) and business model (recently secured a government contract for its AI-powered Virtual Anticipation Network (VANE)).

So, if investors are now betting on the downstream beneficiaries of the AI revolution, is it worth taking a position in this cheaper Palantir alternative?

Let’s look at some numbers…

BigBear.ai vs. Palantir

As you can see, Palantir dwarfs BigBear.ai in terms of market cap (100x the size) and is actually (if nominally) turning a profit, trading at 559x trailing earnings.

It’s also growing revenue a little bit quicker (30% revenue growth over last year’s quarter vs. 22.1% for BigBear.ai).

But what really stands out for both these companies is their price-to-sales multiple (P/S).

Palantir is trading at almost 96x sales while BigBear.ai is trading at almost 11x sales, a far cry from Palantir’s figures but still elevated over the 5x average ratio for software stocks as a group.

Cabot Growth Investor’s Chief Analyst Mike Cintolo spent some time discussing Palantir and its prospects in a recent episode of the Street Check podcast, and my assessment was that Palantir is in the “land grab phase,” where fundamentals are largely ignored as competitors try and gobble up bigger and bigger slices of an emerging pie.

That doesn’t mean the elevated metrics don’t give me pause—they do—but if BigBear.ai can replicate even a fraction of Palantir’s success, it seems worth taking a chance on this high-risk/high-reward stock.

The AI build-out phase (NVDA, hyperscalers, picks and shovels, etc.) seems to have reached peak perception, but the companies whose businesses will utilize that build-out are just getting started.

When that next stage gets more mature and begins printing winners and losers, more traditional metrics like P/S will carry more weight.

But, for the time being, I think we’re better off trusting these movers early while controlling risk via position sizing.