Altana.

Somewhere right now, a shipping container is crossing an ocean and nobody on the receiving end can say, with confidence, what is really inside it - who mined the metal, who stitched the lining, whose factory three tiers down might be on a sanctions list. Global trade moves roughly $30 trillion a year on this kind of educated guessing. Altana's quiet ambition is to end the guessing. The Brooklyn company has spent since 2018 building what it calls a living map of the world's supply chains, and lately the people reading that map include Fortune 100 importers, the largest ocean carriers, and the customs officers who decide whether your container clears or sits.

It is an unglamorous problem dressed up as an impossible one. Supply chains are deliberately opaque - not out of malice, usually, but because no single company has ever been able to see past its own suppliers' suppliers. Altana's bet is that the map was always buildable. The data existed. It was just scattered across millions of parties who had every reason not to share it.

Everyone could see one link. Nobody could see the chain.

Ask a brand where its cotton comes from and it can name a supplier. Ask where that supplier's cotton comes from and the line goes quiet. Multiply that blind spot across a multi-tier network and you get the modern compliance headache: forced-labor rules, tariff regimes, export controls, and carbon disclosure obligations that all demand answers companies structurally cannot give. The penalty for a wrong guess keeps climbing. The tools for a right one barely existed.

The conventional fix was to ask everyone to pool their data into one giant database. That fix has never worked, for the obvious reason that no competitor, and certainly no government, wants to hand its records to a shared bucket. So the data stayed put, and the map stayed blank.

A team that had already done the smaller version.

Evan Smith, Peter Swartz, and Raphael Tehranian were not strangers to this corner of the world. Before Altana, they built Panjiva, a global trade-data company that S&P Global acquired in 2018. Panjiva proved you could turn messy trade records into something searchable. Altana is the same instinct aimed at a far larger target - not a dataset, but a network.

Their wager rested on a then-unfashionable technique: federated machine learning. Instead of collecting everyone's data, Altana trains models that travel to the data, learn from it where it sits, and bring back the intelligence without the raw records ever leaving home. Privacy and visibility, two things usually at war, were invited to the same table. It is the kind of architectural choice that sounds like a footnote and turns out to be the entire business.

A map you can act on, not just admire.

At the center sits Altana Atlas, an AI platform that assembles the map and lets customers query it: trace a product to its origins, screen suppliers against sanctions and forced-labor risk, model what a new tariff does to a portfolio, or estimate the carbon hiding in a bill of materials. The newer flourish is the Product Passport - pitched, with a straight face that turns out to be earned, as "Global Entry for Goods." The idea is to turn a customs entry from the start of an interrogation into a record-keeping event for traders who have already proven themselves trustworthy.

Customers who do not hand out trust easily.

The most persuasive thing about Altana is its customer list, because it includes the world's most paranoid buyers. More than half of the largest logistics providers run on the network. So does U.S. Customs and Border Protection, the UK Department of Business and Trade, the U.S. Space Force, and a roster of Fortune 100 importers. Getting a rival shipper and a federal border agency to lean on the same map is the sort of thing that only works if the federated promise is real.

The Maersk deal, announced in 2026, is the clearest tell. Rather than sell software to a carrier, Altana is being woven into the operational fabric of one - deploying Product Passports across the Gemini Cooperation's East-West routes and powering a service Maersk markets as Trusted Value Chains. When a shipping giant decides your map should sit inside its own plumbing, you have stopped being a vendor and started being infrastructure.

"Fix globalization" is a big phrase. They mean the boring parts.

Altana's stated goal - to make global commerce more secure, resilient, fair, and sustainable - sounds like the kind of line that gets printed on a tote bag. What grounds it is the specificity of the work: catching forced labor before goods ship, letting a government enforce a tariff without paralyzing legitimate trade, giving a company the paper trail to prove its carbon claims. None of it is glamorous. All of it is the connective tissue that decides whether globalization stays trusted or quietly stops working.

The map gets more valuable every time someone reads it.

Networks compound. Every supplier, carrier, and agency that joins makes the map sharper for everyone already on it, which is the same dynamic that made a few other unglamorous infrastructure companies very hard to dislodge. As tariffs multiply, sanctions tighten, and disclosure rules pile up, the demand for a single trusted source of "where did this actually come from" only grows. Altana did not invent that demand. It just built the answer first, and got the world's most skeptical customers to agree.

Return to that container crossing the ocean. The difference now is that somewhere, on a screen in Brooklyn or a customs office or a shipping line's operations room, its contents are no longer a guess. They are a record - traced, screened, and passported before the ship reaches port. The container has not changed. What people can see inside it has. That, in the end, is the whole company.