[DAT] DataWalk: From Wrocław Frustration to Washington Disruption

The unlikely story of how Polish impatience with slow databases toppled a Pentagon contractor

Origins

Krystian Piećko was tired of waiting.

In his cramped office at Wrocław University of Technology, the computer science professor watched the progress bar crawl across his screen. Four hours to copy the dataset. Another two to restructure it. Only then could his analysis begin.

This was the dirty secret of big data in 2010: before you could find insights, you had to move mountains of information from one place to another.

"Knowledge does not occupy physical space," Piećko would tell his students, "but data does."

That observation would pit a Polish startup against Palantir, the secretive, multi-billion dollar company that helped hunt Osama bin Laden.

Wrocław in 2011 was a city caught between centuries. Medieval church spires. Communist apartment blocks. Google engineers eating pierogi at milk bars where Solidarity once planned strikes.

The city had become Poland's tech hub with one advantage Silicon Valley lacked: brilliant programmers at a fifth of California prices.

When Piećko founded PiLab S.A. that October with Paweł Wieczyński and Sergiusz Borysławski, they started small. Document management. Data archiving. The unglamorous plumbing of enterprise software.

But Piećko had built something more ambitious underneath: a hybrid graph-relational database engine.

Think of traditional databases as filing cabinets, great for storage, terrible for seeing connections. Graph databases work like detective boards with red string linking suspects and evidence. Piećko's innovation? Both systems in one, without copying data back and forth.

By 2014, this technical trick had massive implications. Data volumes were exploding 40-60% annually. Enterprises were burning $8 million on average trying to make sense of it all.

The founders faced a choice that would define everything.

Stay a comfortable Polish software house with regional clients. Safe. Profitable. Limited.

Or … bet everything on a global product.

They chose the impossible option. Of course they did. With just PLN 3.1 million (about $1 million), they decided to compete with Palantir, backed by the CIA's venture fund, valued in billions, with contracts running $10-100 million and implementations taking years.

But the PiLab team had spotted something others missed.

For every Fortune 500 using Palantir, thousands of smaller organisations needed the same capabilities. Police departments tracking drug networks. Regional banks spotting fraud rings. Government agencies unravelling financial crimes.

They had sophisticated problems but mid-market budgets.

"Link analysis", seeing hidden connections in data, wasn't just for hunting terrorists anymore. An insurance company needed to spot staged accident rings. A small-town police force needed to map distribution networks.

The transformation accelerated in 2016. They established DataWalk Inc. in Delaware with offices in Palo Alto. Not to abandon Poland, Wrocław remained the R&D heart, but to build a bridge.

The hire that turned heads was Gabe Gotthard. He'd sold his last company to HP for $2.5 billion. Now he was running DataWalk's U.S. operations.

It's quite obvious that when Silicon Valley veterans start betting on Polish startups, people notice.

More strategic recruits followed. Christopher Westphal, who'd written the book on intelligence analysis software. Bob Thomas from NetApp. They weren't just employees, they were translators, teaching Polish innovation to speak Washington.

Early wins came slowly. Polish Ministry of Finance in 2016. TUiR Warta, a major insurer, in 2017, fraud detection improved 60%, saving millions.

Each success became proof for the next pitch.

Then came the breakthrough nobody expected.

In 2023, the U.S. Department of Justice announced it was replacing Palantir in its Money Laundering and Asset Recovery Section.

The winner? DataWalk.

A DOJ official's explanation was blunt: "Having a monopoly is bad for us."

But price wasn't the whole story. While Palantir built walled gardens requiring massive commitment, DataWalk created something that played well with others. Start small. Grow gradually. Keep existing systems.

The professor's insight about data movement had evolved into a business philosophy. Why move mountains when you can analyse them where they stand? Why demand years when customers need answers now?

Staying in Poland kept costs low without sacrificing quality. Going public on the Warsaw Stock Exchange instead of chasing Silicon Valley VCs maintained control. Targeting the "middle market of sophistication", too small for Palantir, too complex for basic tools, revealed an ocean of opportunity.

Today, DataWalk operates from two worlds. Wrocław engineers perfect algorithms. Palo Alto teams navigate federal procurement. Polish pragmatism meets Silicon Valley ambition.

The company that started because a professor was tired of waiting had discovered something crucial. Between data giants, there was room for something more nimble. More accessible. More human-scaled.

The question now wasn't whether a Polish company could compete globally, DataWalk had proven that. The question was how to turn capability into sustainable revenue. How does a company positioned between giants actually make money?

How do they make money

DataWalk makes money the same way a private investigator does: by finding what's hidden, except they do it across billions of digital records simultaneously.

They get paid three ways. Software licenses give customers the right to use DataWalk's platform, like buying a metal detector. Maintenance fees keep that detector updated and working, annual subscriptions that ensure the software evolves with new criminal tactics. Implementation services are the training and setup, teaching organisations how to actually find gold with their new tool.

But these aren't line items on an invoice. They're three acts in a complex drama.

Until recently, DataWalk sold perpetual licenses. Pay once, use forever. Like buying Microsoft Office in 2005. Simple, clean, done.

Except that model was killing them. A perpetual license meant one paycheck, then years of support with no additional revenue. It undervalued what DataWalk actually delivers. When your software prevents a terrorist attack or stops a nine-figure fraud scheme, charging $300,000 once seems almost insulting.

So they ripped up the playbook. The new model: annual subscriptions starting at $560,000.

Price inflation? Not really. It's philosophical alignment with the SaaS dogma. Graph databases, DataWalk's core technology, work by mapping relationships between data points, like LinkedIn for criminal networks. These relationships constantly evolve. New bad actors materialise. Schemes become more sophisticated. A one-time software purchase can't keep pace. Subscriptions fund continuous innovation.

Then there's maintenance, the unsung hero. Every DataWalk customer pays annual maintenance fees, typically 20-25% of their license cost. This isn't tech support answering "Have you tried turning it off and on again?" This is specialised assistance for organisations tracking international crime syndicates.

The brilliant part? Maintenance revenue is sticky. Once an organisation builds investigations around DataWalk, switching to another platform means retraining every analyst, rebuilding every workflow, risking missed connections during transition. The cost isn't monetary, it's existential.

This creates negative churn, a SaaS holy grail where you extract more revenue from existing customers over time. Price increases at renewal. Expanded usage. Additional modules. The customer base becomes an appreciating asset.

Implementation services complete the trinity. What civilians don't understand about enterprise software: buying it is arguably 30% of the total cost. The real expense is making it work with your Byzantine IT infrastructure.

DataWalk implementations are archaeological expeditions through layers of legacy systems. That police department has crime data in a 1990s database that speaks only COBOL. The bank stores transaction records across seventeen different systems that have never talked to each other. Someone has to build bridges between these digital islands.

This is skilled, painful work. DataWalk's implementation team doesn't install software, they perform organisational surgery, rewiring how institutions think about data.

But implementation revenue is treacherous. It's project revenue, labour-intensive, low-margin, non-recurring. Implementations are necessary evils that enable the real business: recurring license and maintenance streams.

Too much implementation revenue means you're a consulting company. Too little means customers can't use what they bought. DataWalk walks this tightrope monthly.

The sector dynamics make everything harder. Government clients, police, intelligence agencies, regulators, move at statutory speed. Budget cycles measured in years. Procurement processes designed to prevent corruption that also prevent efficiency. A deal might take 36 months from first meeting to signed contract.

But government clients, once landed, rarely leave. They pay slowly but surely. They renew religiously. They're the tortoises in a race where most vendors want hares.

Banks are the opposite animal. They move faster, for banks. Six to twelve months from pitch to purchase if you're lucky. But they demand enterprise-grade everything. Security audits that would make the NSA proud. Service level agreements with penalty clauses. Integration with their proprietary risk systems.

The compensation? Banks have money and they're terrified of losing it. Anti-money laundering fines can reach hundreds of millions. One prevented breach, one caught fraud ring, and DataWalk pays for itself forever.

This creates a business paradox. DataWalk's technology is cutting-edge, graph algorithms that would make Google engineers sweat. But their business model is almost medieval. Long courtships. Careful negotiations. Revenues that come in chunks rather than streams.

They're trying to modernise with subscriptions, but changing how enterprise software gets sold is like turning an oil tanker. The entire ecosystem, buyers, procurement, budgets, was built for perpetual licenses. DataWalk gambled they can drag this ecosystem into the SaaS age.

The transition maths are brutal. A $300,000 perpetual license hits revenue immediately. A $560,000 annual subscription recognises $140,000 per quarter. Better long-term, painful short-term.

It's like switching from harvesting mature trees to planting saplings, wise for the forest, tough on quarterly lumber quotas. This is why DataWalk burns cash despite growing revenue. They're financing their customers' transition to subscriptions. Every deal signed under the new model is an investment in future revenue streams.

The real question isn't how DataWalk makes money. It's whether they can survive long enough for the subscription model to compound. Whether they can sign enough Rabobanks before the cash runs out. Whether the promise of recurring revenue can sustain them through the valley of death between business models.

They're selling tomorrow's revenue streams to today's investors, hoping the maths work out before the money doesn't.

Numbers

When a Polish software company grows 109% year-over-year to 13.5 million PLN, your first instinct might be to assume they've cracked some viral growth formula. But DataWalk's spike tells a different story entirely. This is what happens when enterprise sales cycles stretch 24-36 months and suddenly convert. One massive Rabobank contract, years in the making, finally landed. Triple-digit growth here means old seeds sprouting, not new momentum building.

Imagine running a business where 71% of quarterly revenue comes from a single customer. That's DataWalk's reality with Rabobank. Most healthy software companies religiously keep their largest customer below 10% of revenue, it's basic risk management. But when you're transitioning from perpetual licenses to subscriptions, sometimes you take what you can get. The terrifying part? One tough renewal negotiation in 2026 could transform today's 109% growth into tomorrow's 71% collapse.

And here’s the truth about 2.7 quarters of cash runway: it sounds like imminent death but tells you something else entirely. DataWalk burns 4.5 million PLN quarterly with 12 million PLN in the bank. Yes, that's 8 months until the lights go out. But they just raised 58 million PLN (total Series R+S 83.9 million PLN), extending runway to 15 quarters. The real story? They're burning cash not from incompetence but from financing a business model transition. Every new subscription deal delays cash collection compared to the old upfront perpetual licenses. It's expensive being right about the future.

The least understood number on DataWalk's books might be their 77% contract assets ratio, with 10.3 million PLN in delivered-but-unbillable software against 13.5 million PLN in quarterly revenue. Normal software companies show maybe 5-10% here. So why is DataWalk's so high? Enterprise procurement. When Rabobank needs six departments to sign off before cutting a check, DataWalk delivers the software but waits for payment. They're essentially acting as their customers' bank, betting that blue-chip clients are good for the money even if it takes 12 months to collect.

You'd think losing customers while growing revenue is impossible, but DataWalk proves otherwise. Revenue per maintenance customer jumped 45% to 86,000 PLN even as customer count fell from 40 to 36. How? The nine customers who left were paying peanuts, averaging 103,000 PLN each. The ones who stayed expanded by 559,000 PLN total through increased usage and price hikes. It's the software equivalent of a restaurant losing tourists while regulars order better wine. Sometimes a smaller, richer customer base beats a larger, poorer one.

Negative 4.7 million PLN in equity would normally scream insolvency. But DataWalk's case reveals how modern tech accounting can terrify investors unnecessarily. The culprit? A 50 million PLN liability for employee stock compensation (RSUs) that will probably never require cash. These RSUs only cost money if employees exercise options and sell shares they don't yet own. Strip out this paper liability and real equity is positive 35.7 million PLN. Still, try explaining that nuance when investors see negative equity and immediately assume bankruptcy.

Why would any tech company slash R&D from 48% of revenue to just 14%? Survival. DataWalk cut research spending from 12.3 million PLN to 6 million PLN annually while Palantir, their main rival, spends $400 million. It's like bringing a knife to a gunfight, but what choice do they have? When you're burning cash and need to reach profitability, you harvest today's technology advantage and pray it lasts. They're gambling their current product stays competitive without major upgrades. In enterprise software, that bet sometimes pays off, features matter less than switching costs.

36 maintenance customers doesn't sound impressive until you understand enterprise software economics. DataWalk lost 9 customers but maintenance revenue still grew 8% to 10 million PLN. How? The customers who left averaged 103,000 PLN each. The ones who stayed now average 233,000 PLN. With 96% revenue retention but only 80% logo retention, we see the classic enterprise pattern: proof-of-concepts fail, production deployments expand. In this business, keeping the right customers matters more than keeping all customers.

When management announces a $50 million sales pipeline, seasoned investors immediately divide by three. Not because executives lie, but because enterprise pipeline math is optimistic by design. With 24-36 month government sales cycles and typical conversion rates of 20-30%, this pipeline might yield $10-15 million over two years. But here's the twist: DataWalk changed how they count. Under the new subscription model, a five-year, $2 million deal shows as $2 million in pipeline, not $400,000 annually. So that $50 million might be both real and inflated, depending on your perspective.

Would you pay 87% more for software just to rent instead of own? That's what DataWalk asks with their shift from $300,000 perpetual licenses to $560,000 annual subscriptions. The gamble is breathtaking: customers who previously paid once now pay nearly double that amount every single year. Over five years, that's $2.8 million versus $300,000, a 9x increase. But if DataWalk embeds deeply enough in customer workflows, if switching costs become prohibitive, if the software truly prevents financial crimes worth millions, then maybe, just maybe, customers accept paying forever. Rabobank already has.

People

Three Polish guys own 18% of DataWalk but control 27% of the votes.

That's the ownership story in one sentence. Paweł Wieczyński (CEO), Krystian Piećko (CTO), and Sergiusz Borysławski each own exactly one-third of a company called FGP Venture. FGP Venture owns 1.175 million DataWalk shares.

Do the maths: 18.41% of shares, but those shares carry double voting rights, so 26.73% control.

Startups some times have one dominant founder who takes 60-70%. DataWalk split everything three ways. The double-voting shares are Series A, issued back when DataWalk was PiLab. Everyone else gets regular shares with regular votes. It's a control mechanism as old as capitalism: founders keep the steering wheel even as they sell seats in the car.

Wieczyński is the true believer CEO who "invested over half a decade working 7 days a week and all [his] money" into DataWalk. That's not PR speak. That's what he actually wrote to shareholders. He owns just 58,161 shares directly (0.91%), keeping his real stake hidden in FGP Venture.

Piećko is the technical brain with 18 patents and what management calls "knowledge that is unique on a global scale." DataWalk explicitly admits losing him would be catastrophic. The company's survival depends on one man's brain.

Then there's Gabe Gotthard, the American CEO who sold his last company to HP for $2.35 billion. He runs US operations from California.

But there's a contradiction: DataWalk Inc. generates just 12% of revenue while holding 92% of the employee stock obligations. In other words, it's an expensive sales office.

DataWalk employs 38 people. Full stop. That's fewer people than a typical McDonald's.

Yet they issued stock options to 113 people. Where are the other 75? They're contractors, advisors, former employees who kept their equity. DataWalk runs like a distributed network, not a traditional company.

Revenue per employee? $162,000 annually for 2024. Q1 2025 numbers project $389,000 per employee annually at current exchange rates. That's a 140% jump, driven almost entirely by the Rabobank deal.

What's remarkable is that they support Fortune 500 clients and government agencies with this skeleton crew.

The quirk worth noting: 91.6% of stock compensation costs sit with the US subsidiary, which generates 12% of revenue. Expensive American talent gets paid in equity to sell Polish technology. The engineers in Wrocław building the actual product? They're on salary.

After the founders, the biggest shareholders are Dutch pension funds (Nationale-Nederlanden with 11%) and a French investment firm (RAISE SAS with 10%).

The Dutch makes sense, Rabobank is Dutch, and pension funds love boring B2B software companies. The French jumped in during the 2025 Series S round at "a premium to market value."

It's worth noting the absence: zero US institutional investors despite having US operations since 2016. When American VCs won't touch your American subsidiary, it reveals a lot.

DataWalk has 36 paying customers. I'll repeat that: thirty-six.

One of them, Rabobank, represents 71% of Q1 2025 revenue. If Rabobank sneezes, DataWalk gets pneumonia.

The customer roster reads like a spy novel: Department of Defense, Department of Justice, "unnamed domestic security services," Morgan Stanley, Barclays. These aren't companies buying software on a credit card. These are multi-year procurement cycles with security clearances.

The typical DataWalk buyer is a Chief Analytics Officer or Head of Financial Crime at either a bank terrified of money laundering fines or a government agency hunting bad guys. They have million-dollar problems and six-figure budgets.

What's telling is who DataWalk is losing: they went from 45 to 43 customers in 2024, but revenue per customer jumped 45%. They're firing small clients to focus on whales?

Reading between the lines of their 94/100 user satisfaction score, customers see DataWalk as the "good enough" alternative to Palantir. Not revolutionary, but 70% cheaper and without the consultant army.

One customer quote reveals everything: "Initially, the project will cover less than 5% of users in the given department." Even Rabobank, their biggest win, is testing the waters carefully.

These customers are pragmatists with budget constraints. They need Palantir-like capabilities but can't afford Palantir prices. So they gamble on 38 Poles and hope the company survives long enough to support their deployment.

DataWalk is owned by three Polish founders who split everything equally and haven't killed each other in 13 years. Run by a true believer CEO and an irreplaceable CTO. Funded by European institutions while being ignored by Silicon Valley. Serving 36 of the world's most demanding customers with fewer employees than a Starbucks.

It's a company that shouldn't work but does.

Which raises the question: in a world where Palantir has 3,500 employees and charges 10x more, how does a 38-person Polish company compete? What's their edge, their moat, their reason to exist?

Competition & the moat(?)

DataWalk competes in a market that shouldn't exist.

Think about it. If you need to find hidden connections in massive datasets, you have two options. Hire Palantir for $10-100 million and wait two years. Or build it yourself with open-source tools and a team of PhDs.

DataWalk found the gap between these extremes. But finding a gap and defending it are different things entirely.

The obvious competitor is Palantir. DataWalk offers 70% of Palantir's capabilities at 20-30% of the price. But Palantir isn't really DataWalk's competition, not in the way that matters. It's quite obvious that they serve different markets. If you're considering DataWalk, you've already decided Palantir's $141,000 per server core is too rich for your blood.

The real competition comes from three less obvious places.

First, there's IBM i2 Analyst's Notebook, the Microsoft Word of investigation software. Every police department has it. Every analyst knows it. It's been around since the 1990s. IBM i2 does link analysis too, arguably worse, but worse plus familiar often beats better plus unknown.

Then there's Quantexa, the British upstart going after financial crime. They've raised $370 million and signed HSBC and Standard Chartered. With their Microsoft partnership and 60 billion+ record processing capability, Quantexa represents what DataWalk could have been with Silicon Valley funding: slicker marketing, bigger contracts, deeper pockets.

The third competitor? ChatGPT plus a determined analyst. This is 2025. Why buy specialised software when you can upload CSVs to Claude and ask "find me the hidden connections"? DataWalk's management even acknowledges this threat, positioning their platform as providing "grounded facts" to AI models to "minimise hallucinations." But if AI keeps improving, how long before it makes graph databases feel like fax machines?

So what's DataWalk's moat? What stops these competitors from crushing 38 Poles with a clever database?

Switching costs are their strongest defence. As we mentioned earlier, once those 36 customers build investigations around DataWalk's Universe Viewer, they're stuck. Every saved query, every trained analyst becomes a hook. The 96% maintenance renewal rate proves it. Ripping out DataWalk means retraining everyone and rebuilding years of work.

The Goldilocks positioning creates another moat. Too sophisticated for AI prompting, too cheap for Palantir prospects. DataWalk lives in this narrow band where customers need real graph analytics but have municipal budgets. Palantir can't come down here without destroying their pricing model. Their average deal is 10-50x larger than DataWalk's. The economics don't translate.

There's also the speed advantage, though it's eroding. DataWalk implementations take 19 weeks, vis-a-vis Palantir's 12-18 months. In government procurement, where decisions stretch forever, being 4x faster matters. But Quantexa claims similar deployment speeds now. This advantage has maybe two years left.

The Polish cost structure provides mathematical protection. Those Wrocław engineers cost way less than Silicon Valley equivalents. DataWalk can profitably serve customers that would be rounding errors for Palantir. But this assumes talent stays in Poland and stays good. Both assumptions grow shakier as Polish tech salaries rise.

What about moats they've lost?

The technical moat evaporated around 2019. Remember Piećko's hybrid graph-relational database? Revolutionary in 2014. Today, Amazon Neptune does the same thing. So does Microsoft's Cosmos DB. Neo4j added similar features. DataWalk's special sauce became table stakes.

They also lost the first-mover advantage in anti-Palantir positioning. When DataWalk started, being the affordable alternative was unique. Now everyone claims it. Databricks, Snowflake, even Microsoft position themselves as the "democratised" choice. According to their own CEO, only two real competitors exist in their specific niche: DataWalk and Palantir. But that's increasingly wishful thinking.

It's possible that DataWalk has convenience moats, not power moats. Their customers stay because switching is annoying, not impossible. They win deals because they're cheaper and faster, not fundamentally better.

The 71% revenue concentration with Rabobank reveals a crucial pattern: even their biggest wins start small, "less than 5% of users in the given department." They're not displacing systems. They're supplementing them.

Which brings us to the existential question: Can a company with weak moats and strong competition survive on execution alone? With negative equity and 2.7 quarters of runway before the Series S funding, the answer had better be yes.

Mr. Market

Mr. Market has taken DataWalk shareholders on a roller coaster ride. Like most tech companies do.

You buy at 50 PLN in early 2020. By 2021, you're up 400% at 260 PLN. November 2023? Down 42% from your original investment at 29 PLN. Today, at 115 PLN, you've doubled your money. But if you bought at the peak? You're still down 56%.

This is how the market completely changed its mind about DataWalk's worth, three times.

In 2020-2021, Mr. Market fell in love. COVID made every company a "digital transformation play" and DataWalk, with its graph analytics and spook contracts, was perfect narrative candy. At peak froth, investors paid 60 times annual revenue. Palantir, even in those bubble days, rarely touched 40 times sales.

The market wasn't buying a Polish software company. It was buying the dream of a European Palantir, complete with intelligence agencies and criminal networks.

Then came the hangover. Ukraine invaded in February 2022 sent European tech tumbling. DataWalk dropped from 210 to 130 PLN in days.

But the real pain? The numbers. Revenue growth slowed to 8% in 2022. The company posted a 117 million PLN loss, three and a half times their revenue. Mr. Market suddenly realised this wasn't hypergrowth SaaS but a cash-burning software vendor with 40-something customers.

By late 2023, the market had gone from euphoric to suicidal. At 29 PLN, DataWalk traded at less than 1x sales, bankruptcy pricing. And honestly? The market had a point.

Cash had dwindled from 74 million to 12 million PLN. Revenue was shrinking. Management issued going concern warnings. One day in November, over 300,000 shares traded vis-à-vis the usual 10-20k, suggesting forced liquidations. The market was pricing maybe a 10% chance of survival.

What changed? As we noted earlier, management cut costs by 40% and shifted to subscriptions. But the real catalyst was Rabobank in March 2025.

When a Dutch banking giant agreed to pay more annually than DataWalk's old perpetual price, Mr. Market recalculated. The stock doubled in April, hit 127 PLN by July. The 300% rally wasn't relief, it was repricing from "likely dead" to "possibly alive."

Today at 115 PLN, the market values DataWalk at roughly 19 times trailing sales (based on 34 million PLN run rate post-Rabobank). That's expensive for a cash-burner but cheap if they've cracked enterprise sales.

Reading between the lines, the market's saying: "There's maybe a 20% chance this becomes a 200 million PLN revenue company, and 80% chance it muddles along or fails."

The maths is brutal. At current prices, DataWalk's enterprise value sits around 635 million PLN. If they land those mythical 6 million USD enterprise deals, just ten annually would justify today's price.

But they have 36 customers total. One represents 71% of quarterly revenue. The market is betting on unproven strategy, not quite the blind faith of 2021, but definitely hope over history.

Bear Thesis

The bull case requires everything to go right. The bear case? Just one thing needs to go wrong.

Start with the Rabobank time bomb. We mentioned that 71% concentration earlier, but when that contract comes up for renewal, who has the leverage? DataWalk burning through 4.5 million PLN quarterly, or Rabobank knowing they're DataWalk's oxygen supply? The customer becomes the hostage-taker.

Then there's the SaaS transition math. They're asking customers to pay 560k USD annually instead of 300k perpetual, nearly double for the privilege of renting instead of owning. Meanwhile, they're burning cash faster than they're booking it. The J-curve of SaaS transitions is brutal for healthy companies. For one with 2.7 quarters of runway? It could be fatal.

The credibility gap might be even wider than the financial one. You're a Fortune 500 CISO. Your job is on the line. Do you bet on DataWalk, negative equity, Polish headquarters, 38 employees, or pay up for Palantir's peace of mind?

That "middle market of sophistication" sounds clever until you realise it might mean too expensive for small companies, too unproven for large ones.

Geography compounds the problem. Those Wrocław engineers at one-fifth Silicon Valley cost? Fantastic for the P&L, woeful for customer acquisition. Enterprise software gets sold at RSA Conference in San Francisco, not via email threads to Poland. Their U.S. subsidiary generating 12% of revenue while holding 92% of equity obligations reveals everything about this mismatch.

The talent equation is particularly vicious. Can't afford Silicon Valley engineers on Polish burn rates, but can't win Silicon Valley customers without Silicon Valley presence. And Krystian Piećko, whose "knowledge is unique on a global scale"? One LinkedIn message from a Mega-Cap and DataWalk's technical moat evaporates.

It's possible that several risks materialise simultaneously: Rabobank flexes at renewal, cutting revenue 30%. The SaaS transition stalls as other customers balk at higher prices. U.S. expansion burns cash without corresponding wins. Piećko gets poached. Suddenly that 15-quarter runway looks like 5, and the Series T raise happens at a painful valuation.

The market's pricing in the dream. But companies at this stage don't get points for dreams, they live or die on execution. And with this many single points of failure? The bear case isn't that DataWalk fails. It's that they run out of time before they can succeed.

Bull Thesis

The bull case for DataWalk starts with a simple observation: when the US Department of Justice fires Palantir and hires you instead, you're no longer a startup, you're a legitimate threat. Those 36 customers we mentioned? They include the DOJ's Money Laundering unit, Morgan Stanley's compliance team, and Barclays' intelligence unit. Throw Amazon in there as well. When three of the world's most sophisticated data analysis buyers independently choose your platform over the $20 billion gorilla, you've crossed from promise to proof.

It's worth noting that the CEO who "invested all his money" and works "7 days a week" has delivered these blue-chip validations. The three founders still own 18% and control 27% of votes, exactly the skin-in-the-game that owner-operator investors prize.

What Mr. Market misses: DataWalk is perfectly positioned for Europe's data sovereignty revolution. While American tech giants face increasing EU scrutiny, DataWalk's Polish roots transform from weakness to strategic advantage. Germany won't let Palantir near certain government data. France wants European alternatives. The UK, post-Brexit, seeks non-EU but non-US options. DataWalk offers all three what they want, sophisticated analytics without the geopolitical baggage. With EU defence spending at historic highs and data localisation requirements tightening, being the "European Palantir" isn't marketing fluff. It's a moat worth billions. We could argue this is textbook geographic arbitrage, world-class technology trading at Warsaw Stock Exchange multiples instead of NASDAQ premiums.

The subscription transition that terrifies short-term investors should excite business model transformation specialists. Yes, recognising $140,000 quarterly instead of $300,000 upfront hurts today's numbers. But Rabobank isn't paying $560,000 once, they're paying it every year. Do the maths: five years under the old model meant $300,000 total. Five years under the new model means $2.8 million. Every single customer who renews at the new pricing adds $2.5 million in lifetime value. With 96% gross retention, the compound maths becomes beautiful. Adobe's stock quintupled after its Creative Cloud transition, same playbook, different vertical.

We're witnessing the "good enough" revolution in enterprise software. Not every organisation needs Palantir's Death Star capabilities. Most need Toyota Camry reliability at Toyota Camry prices. DataWalk delivers 80% of the functionality at 20% of the cost, implementing in weeks not years. There are thousands of regional banks, police departments, and government agencies with sophisticated problems and municipal budgets. Palantir can't serve them without destroying their unit economics. DataWalk can serve them profitably. It's the Southwest Airlines of intelligence software. Peter Lynch would recognise this immediately: a simple business solving real problems for willing customers.

Each customer DataWalk lands makes the next sale easier, that's the network effect nobody discusses. When the DOJ chooses you, state agencies follow. When Rabobank signs on, other European banks take notice. The 96% retention rate isn't about sticky software; it's about results that create references. In enterprise software, nothing sells like success. DataWalk is building a reputation flywheel where each implementation becomes a case study, each case study becomes a reference, each reference becomes three prospects. At 36 customers, they're at the inflection point where word-of-mouth accelerates. They're becoming what Hermann Simon calls a "hidden champion", dominant in a narrow niche most investors ignore.

The AI boom doesn't compete with DataWalk, it amplifies their value. Every company rushing to implement AI faces the same problem: language models hallucinate when they lack grounded facts. DataWalk's knowledge graphs solve this by providing the factual backbone AI needs to function reliably. They're not selling AI; they're selling AI insurance. As organisations discover that ChatGPT can't actually investigate financial crimes without real data connections, DataWalk becomes the adult supervision. Classic pick-and-shovel play, while everyone speculates on which AI will win, DataWalk sells the tools they all need.

At 19x sales, the market prices DataWalk like a struggling software vendor. But if they capture 5% of Palantir's addressable market at current pricing, we're looking at $400 million in recurring revenue. That's a 12x increase from today. The downside? They muddle along at 50-100 customers generating decent cash flow. The upside? They become the default alternative to Palantir for the 95% of organisations that need sophisticated analytics but can't afford nine-figure implementations. At $178 million market cap, you're paying for the downside and getting the upside free.

It's possible that the negative accounting equity scaring lazy investors is actually 50 million PLN of non-cash RSU charges, exactly the kind of complexity that creates opportunity. Sometimes the best investments are hiding in plain sight, 38 people in Poland steadily signing the world's most sophisticated buyers, one transformative contract at a time. Whether you're hunting geographic inefficiencies, owner-operator excellence, or asymmetric optionality, DataWalk offers all three wrapped in one underfollowed package.

So what do we make of all this?

After spending time with DataWalk, you start to see it's not really a technology story. It's a timing story. Graph databases aren't new. The CIA has been connecting dots on whiteboards since the Cold War. What changed is that suddenly everyone needs to be their own CIA. Every bank is drowning in transactions that might be money laundering. Every insurance company is bleeding money to fraud rings they can't see. The world got complex, and Excel stopped working. DataWalk showed up at exactly the moment when the pain became unbearable and the technology became possible.

The bull case is beautifully simple: complexity is growing faster than our ability to understand it. That's the mega-trend, and it's not reversing. Twenty years ago, a criminal moved money through three banks and called it a day. Now they're layering transactions through crypto, shell companies, and dozen jurisdictions before breakfast. The bad guys got sophisticated, but most of the good guys are still using tools from 2005. DataWalk is betting that this gap, between criminal innovation and investigative capability, is about to create a massive market. They're not selling software; they're selling the ability to see clearly in a world designed to obscure.

The bear case is equally simple: DataWalk is a features company in a platform world. Palantir isn't really selling graph analytics, they're selling transformation. They embed themselves so deeply into an organisation that leaving would be like performing a kidney transplant. DataWalk shows up with a better, cheaper kidney, but surgery is surgery. And here's the thing about enterprise software: nobody ever got fired for buying IBM. Or in this case, Palantir. The switching costs aren't technical, they're psychological. DataWalk might have built a better mousetrap, but the mice have Stockholm syndrome.

What happens next probably depends on which movie we're in. If we're in the "disruption" movie, DataWalk follows the classic playbook: start with the customers Palantir ignores, get better every year, and eventually eat the market from below. Think Salesforce versus Siebel. In this version, five years from now every regional bank has DataWalk, and Palantir is explaining to analysts why they're focusing on "enterprise value" not market share.

But if we're in the "consolidation" movie, DataWalk becomes a feature of something bigger. Microsoft or Oracle wakes up, realises they need graph analytics, and buys the easiest solution. The founders get rich, the technology gets buried in some enterprise suite, and we all pretend this was the plan all along.

But I keep thinking about something else. DataWalk accidentally built a company that mirrors how the world actually works now. They've got Polish engineers building the engine and Silicon Valley executives selling it. They're using regulatory arbitrage, American prices with Polish costs, the same way their customers are trying to catch people using financial arbitrage.

It's possible that the companies that win the next decade won't be the ones with the best technology or the most money. They'll be the ones that understand the new geography of talent and capital. DataWalk gets this, perhaps without even trying. They're not a Polish company or an American company, they're what companies look like when borders matter less than ability. And in a world where crime is borderless, that's arguably exactly the right structure for the solution.

---

LEGAL DISCLAIMER AND DISCLOSURE

© 2025 Silba. All rights reserved.

This analysis is for informational purposes only. The content represents the views and opinions of the author, based on publicly available information. While we strive for accuracy, we make no representations or warranties regarding the completeness or accuracy of the information presented. Errors will occur.

This analysis is not investment advice. It is not a recommendation to buy, sell, or hold any security. The author may hold positions in securities mentioned. Future performance of any investment discussed cannot be guaranteed.

All financial figures, unless otherwise noted, are approximations based on publicly available information. Readers should verify all claims, statistics, and analytical processes independently. The information contained herein is not intended to be a complete analysis of any security, company, or industry mentioned.

Market conditions change. Past performance is not indicative of future results. Investments mentioned may not be suitable for all investors. Readers should conduct their own due diligence and consult with professional financial, legal, and tax advisors before making any investment decisions.

By accessing this analysis, you agree that neither Silba nor its affiliates bear any responsibility for investment decisions you make based on this information.