The DeepSeek news last month may not have derailed the artificial intelligence stock rally, but it undoubtedly shook up the narrative driving those stocks higher, and it’s helping make winners out of non-hyperscaler AI companies like Palantir (PLTR) and BigBear.ai (BBAI).
For the last two-plus years, investors have been approaching the AI theme with the assumption that the inevitable outcome is a growing number of data centers, powered by nuclear energy, supporting larger and larger artificial intelligence models learning and iterating on as many Nvidia (NVDA) chips as the mega-cap tech companies can get their hands on.
But the arrival and adoption of smaller, fine-tuned models (like DeepSeek) 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 first scenario is no doubt the best-case scenario for companies such as Meta Platforms (META), Microsoft (MSFT), Alphabet (GOOG), and Nvidia, as a company’s ability to spend a veritable fortune on GPUs and build out data centers becomes table stakes. After all, there are precious few companies that can afford to spend tens of billions of dollars a year simply building out the infrastructure needed to maybe turn a profit.
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), lowers the payout for the tech hyperscalers (they provide AI-as-a-Service to train smaller models instead of having a monopoly over every step in the process; I think Meta still ends up a winner here) while purpose-built AI models from smaller companies thrive.
Palantir is a notable winner here, at least based on share price (up 291% in the last six months, making it the best-performing mega-cap stock over that period) 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 592% in the last six months, making it the fourth-best-performing mid-cap stock) and business model (recently secured a government contract for its AI-powered Virtual Anticipation Network (VANE)).
So, if investors are betting that BigBear.ai will become the next Palantir, is it worth taking a position in the stock, or is it too late to buy BigBear.ai?
Let’s look at some numbers…
BigBear.ai vs. Palantir: By the Numbers
Palantir (PLTR) | BigBear.ai (BBA) | |
Market Cap | $260.5 billion | $2.1 billion |
Trailing P/E | 602 | N/A |
Trailing P/S | 94.81 | 11.65 |
Quarterly Revenue Growth (YoY) | 30% | 22.10% |
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 602x 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 95x sales while BigBear.ai is trading at almost 12x 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 the latest 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.
Sign up now!
Sign up now!
Sign up now!
Sign up now!
