Palantir Technologies CEO Alex Karp was uncharacteristically wearing a sports coat as he stood on stage at his company's annual AIPCon gathering in March. As usual, though, he was ebullient.

"I'm wildly enthusiastic about where we are, wildly enthusiastic about the people we are partnering with, and wildly enthusiastic about our supporters in our investor base," he said to a packed event space in Palo Alto, Calif. In return, Palantir's customers and its committed corps of retail investors are wildly enthusiastic about Karp and his company.

But on Wall Street, it has been more on-again, off-again applause for Palantir and its stock. The business has become the subject of great debate, ranging from Karp's views on Silicon Valley and politics to the scale of opportunity that lies ahead for the company's unique approach to software.

Wall Street generally agrees on its near-term growth opportunity. When the company reports first-quarter results on Monday, analysts expect revenue to grow 36% to $862 million, with earnings surging 62%.

Analysts can't agree, though, on how to price the asset. Price targets among Wall Street analysts currently range from $40 to $130. Palantir recently traded at $118. The uncertainty shows up day to day, as well. Since going public in 2020, Palantir's daily moves have swung an average of 5.6 percentage points from the high to the low. The S&P 500 index has seen a daily average spread of 1.2 points for the same period.

The volatility and debate stem from the company's nontraditional roots: It's a new-age software firm that grew up in the staid world of national security and defense. A devoted base of retail investors unbridled from fundamentals further clouds the company's value.

Karp's sales pitch combines problem-solving engineers and artificial-intelligence-based analysis fed into what software people like to call a "single pane of glass."

Two trends launched Palantir around the turn of the millennium. The first was the proliferation of data, which began to explode as the internet took root.

"It is clear that we are all drowning in a sea of information," a pair of prominent University of California academics wrote in an influential 2000 study. "The challenge is to learn to swim in that sea, rather than drown in it," the professors, Peter Lyman and Hal Varian, noted. "Better understanding and better tools are desperately needed if we are to take full advantage of the ever-increasing supply of information."

One of those tools turned out to be software from Palantir, which was founded by Karp, Peter Thiel, and Stephen Cohen in 2003.

The other catalyst for Palantir's business was 9/11 and the global war on terror. U.S. intelligence and defense agencies were creating mountains of data, and trying to respond to threats in real time. But disparate information sources and formats were slowing down agencies unequipped to move at a faster pace.

Palantir offered to integrate the data using automation to aid human interpretation and decision-making. The company's software turned Lyman and Varian's "sea of information" into an aquarium, where data was visible, secure, and contained.

Even then, Palantir had trouble raising capital, according to Karp. "In the beginning, we had no money, we had no clients, no VC would back us," he said during a February forum that was livestreamed to retail investors. "No one inside the company thought I should be running the company, and every smart person, every expert, thought we were either crazy or stupid or clowns."

Karp, aside from that rare jacketed appearance, usually wears T-shirts and sports a haircut that would be the envy of a mad scientist. He holds a doctorate in neoclassical social theory from Goethe University in Frankfurt, and uses philosophical concepts to build Palantir's software. The company frequently talks about ontology, which is a branch of metaphysics that studies relationships between entities.

Palantir has used the ontology concept to map the nuts and bolts of an organization and draw links between them. Within a healthcare provider, for example, a nurse is an employee who is linked to locations, specialties, shift preferences, and more.

Drawing connections inside these "data lakes" is a key first step, according to reinsurance giant Swiss Re, a Palantir customer. "Many enterprise data problems are in fact data-integration problems," the company tells Barron's.

For years, the corporate answer to making sense of more data was hiring more analysts, but that approach gets expensive and difficult to coordinate. Palantir's ontology-based software offered a new approach. The idea was particularly appealing to intelligence agencies, which deal with troves of unlinked data.

In 2005, Palantir received funding from a nonprofit venture-capital firm established by the Central Intelligence Agency, today known as In-Q-Tel. Palantir began to win government contracts with U.S. intelligence, defense, and law enforcement agencies, and by 2009 had launched a similar product for commercial customers that were running into the same kinds of data issues.

A decade later, Palantir's revenue is roughly split between government and commercial customers.

Palantir approaches potential customers with an unusual sales pitch. "They don't come in superorganized. It's incredibly unconventional -- there's no sales force presentation, there are no pamphlets, there's none of that foolishness," Pete Suerken, the CEO of the co-op that supplies Wendy's restaurants, tells Barron's.

Suerken ultimately hired Palantir as Wendy's sought to adapt to changing habits during the pandemic. "As we looked across our network, we said to ourselves, 'The amount of inventory that we're carrying to generate a sale is outrageous. Why do we do it this way?' No one ever tried to optimize the entire network," Suerken says.

The co-op centralized inventory management using Palantir's predictive models. Suerken says that stores now have the right amount of inventory and have reduced spoilage on fresh beef, and store managers have more time to spend in the front of the restaurants. It's now figuring out how to make fewer deliveries with fewer trucks.

Palantir is also winning contracts with companies that have accrued years of data across various IT systems that never learned to talk to each other. It's particularly problematic for healthcare companies, where hospitals often use decades-old technology.

During the Covid pandemic, there was a nationwide shortage of nurses; Palantir offered a nontraditional software fix. "It's not like they just sent a computer science engineer," says Etter Hoang, the chief data and analytics officer for Tampa General Hospital. "They literally went to the floor, started observing workflows, started talking to our schedulers. They saw firsthand how they were scheduling nurses, and then they interviewed them. They wrote out the workflows and then they designed the tool to help them with their workflow."

Rohit Chandra, Cleveland Clinic's chief digital officer, says staffing at his health network had previously been a "massive spreadsheet exercise."

Palantir helped replace the spreadsheets with a unified dashboard: "Now we have the ability, in a single enterprisewide way, to have a place where we can look at all patient assignments, all bed assignments, and all nurse assignments," Chandra says.

Palantir went public in September 2020, amid the market's post-Covid lockdown euphoria. Shares closed their first day of trading at $9.50, giving the company a market value of roughly $16 billion. They have swung wildly ever since, but trended in one direction -- higher. Shares are up 1,145% from the first-day close price, a 73% annualized return. Total return for the S&P 500 index in that period was 11% a year.

Palantir's 2024 financials are everything investors want to see from a midstage software company. Revenue grew by 29%, and Wall Street analysts are expecting 31% growth in 2025.

Plenty of software companies grow revenue at that rate; the trick is doing it profitably. Palantir had an 11% operating margin in 2024 and a 40% free-cash-flow margin.

While Palantir's roots are in government contracts, the commercial opportunity ahead of it may be larger. In 2024, 55% of revenue came from governments, a customer base the company restricts to the U.S. and its allies.

To be sure, Palantir faces plenty of risks. Sluggish growth outside of the U.S. and United Kingdom is slowing the company's expansion, a risk likely to mount amid the continuing trade war. Meanwhile, Palantir's top three unnamed customers -- probably branches of the U.S. government -- comprised 17% of total revenue last year.

But for investors, the biggest risk of all is the price of the stock. At a recent $118, the stock trades at 69 times 2025 estimated sales. The next closest stock in the S&P 500, Texas Pacific Land, trades at less than half that ratio.

"I think the company is incredibly well run. It's a fundamentally sound story," Jefferies analyst Brent Thill told Barron's when the share price was around $91. "But the valuation doesn't make any sense."

"If it went to the second-most expensive software multiple today, that'd be a $40 stock," Thill says, pointing out the troubling history for pricey shares. "When Snowflake and Datadog and others had gotten to these multiples, they didn't sustain; they fell out of the sky. And history is a good guidebook. We've seen this movie before."

Thill has a $60 price target and an Underperform rating on Palantir shares. "Even if you assume they were able to grow at 40% to 50% for the next five years straight, five years from now it would still be the most expensive name in software," he says.

Karp has long shown disdain for analysts who are bearish on his stock. In a February interview he said, "I love the idea of getting a drone and having light fentanyl-laced urine spraying on analysts who've tried to screw us."

Thill believes Karp is taking it too personally. "It's a great company. I'm not disputing the fundamentals. We're talking about the stock. The stock and the company are two different things, and Alex Karp doesn't seem to understand that."

Morningstar's Mark Giarelli thinks the wide range of price targets comes from confusion around Palantir's commercial opportunity. "Analysts cannot get a good handle on what the total addressable market is," he told Barron's when the stock was at $82. "Theoretically, Palantir can help almost any company in the world that utilizes data, but what will actually materialize is ambiguous and anyone's best guess." Giarelli has a $90 price target on Palantir shares.

The discounted cash flow models that are the bread-and-butter of Wall Street analysts don't mean much to Palantir's base of devoted retail investors. Institutions and insiders own 59% of Palantir's shares; across the S&P 500, that average is 88%.

The retail-heavy breakdown changes the dynamics for Palantir's stock and its valuation. "That shareholder base is willing to pay a significant premium for the magnitude of opportunity ahead for Palantir," analyst Gil Luria of D.A. Davidson says. "We believe this is equivalent to the significant premium Tesla traded at for many years, as core investors saw an opportunity much greater than the market perception. In acknowledgment of that, we maintain a multiple that is well above other companies we cover."

Luria has a $100 price target, recently lowered from $105, and a Hold rating.

Palantir has catered to its retail investors, with Karp taking questions from them in informal online forums and, even during earnings calls, always heaping praise on them for their endurance. During AIPCon, Amit Kukreja, a full-time financial blogger, hosted intermission programming on Palantir's livestream.

In the past two weeks, Kukreja has posted three new Palantir videos to his YouTube channel, which has 70,000 subscribers. "People can look at whatever type of valuation metrics they want to look at and they can say, yeah, it's expensive," he said when the share price was around $90.

Kukreja has been covering the company for four years, and says that the community of retail investors were ahead of Wall Street because they ignored the confines of cash flow modeling, instead focusing on more "qualitative elements."

In Kukreja's telling, the Palantir investor community began to gel online in 2022 as the stock was getting pummeled, bottoming at $6 late that year. Many in the group stuck with the stock to see it touch $124 this February. By early April it was nearly cut in half, before the rally began anew.

Kukreja says the high was "an aggressive valuation."

Davidson's Thill sees the base of retail investors as a potential weakness for the stock because they have a "lower tolerance for pain" than institutional investors, leading to extreme volatility.

But those who owned the stock at $6 believe analysts continue to miss the bigger long-term picture with Palantir. "I think the problem with a lot of the analysts is that they were saying it was expensive at $6, at $10, at $15, at $20, at $30, at $35," Kukreja says. "And so, obviously the stock has outrun them."