The Watson Weekly is sponsored by Avalara — the agentic AI platform automating global tax and compliance for leading eCommerce brands with simple plans, transparent pricing, and complete coverage across Shopify, Stripe, WooCommerce, BigCommerce, and more.

Raj Subramaniam gave Fortune a number that deserves more attention than the headline it generated. FedEx moves about 18M packages a day, and by his account that traffic throws off two petabytes of data every 24 hours. He calls the company the referendum on global supply chains. It is a good line. The harder question is whether a referendum can be monetized.

His argument for 2025 being the strangest year in his 35 years is specific rather than atmospheric. US imports down. US exports up. Intra-regional volume climbing, with Latin America inbound and Southeast Asia and India growing fast. He has a name for it, reglobalization, which he says he coined himself. Naming a thing is not the same as being positioned for it, but FedEx has a decent claim here. If a shipper was moving goods from A to B and now moves them from P to Q, FedEx already sits at all four letters. That is what 53 years of network buildout buys you.

The customs point is the one I keep circling. Subramaniam says the number of packages requiring formal clearance rose six or sevenfold, and FedEx hired roughly two thousand people to handle it. Read that plainly. Tariff complexity became a hiring event. He frames it as an advantage because expertise is scarce, and he is probably right. But an advantage that costs two thousand headcount is a margin story before it is a moat story, and the interview does not tell us which way that trade nets out.

Then comes the pivot every logistics operator is attempting right now. FedEx wants to sell route optimization, predictive maintenance, and customs clearance as products. There is a Dun & Bradstreet partnership and work with ServiceNow. Subramaniam points to $1.8T of inefficiency sitting in global supply chains and suggests FedEx can take a piece of it. That figure is his, not an audited market. Healthcare is at nearly $10B in revenue, which is real and growing, but healthcare is transportation revenue won partly on the strength of visibility tools. It is not software revenue. Nobody has broken out what the data layer earns on its own, and until somebody does, the intelligence business is a story attached to a trucking business.

Meanwhile the operational work is unglamorous and unfinished. Network 2.0 is 45% complete against a fall 2027 target, with 65% expected by year end. Aurora has logged 700k autonomous miles for FedEx with a safety driver still in the seat, and highway point-to-point is where it starts. Dexterity handles truck loading, Berkshire Grey the unloading, which Subramaniam describes as an extremely hard robotics problem now solvable. Those are the only parts of the big surface hubs a human still touches

So FedEx has the data, the network, and a founder-built culture it is careful not to disturb. What it does not yet have is proof that anyone outside FedEx will pay for what FedEx knows. Freight is now trading on its own with a first earnings call ahead. That print will tell us more about the next two years than the petabytes will.

$14B buys Uber something it has been circling for years. Not food delivery. Frequency.

The number worth staring at is the one Uber put in its own rationale: cross-platform users generate roughly three times the gross bookings of single-product users. That is the whole thesis. Uber is paying a premium to convert 65 countries of standalone food orders into Uber One members who also take rides. Whether that conversion happens at anything close to the rate Uber assumes is the open question, and nobody will know until 2028 at the earliest.

The SSW carve-out tells you where Uber expects regulatory trouble. $11B in gross bookings, sold to a financial buyer who will resell it, in exactly the markets where Uber Eats and Delivery Hero compete head to head. That is a pre-emptive concession, structured before Brussels asks.

Niklas Östberg's framing was blunt. Food delivery is scale dependent, and it is hard to build that scale from a European base. Fifteen years of Delivery Hero ends with an American buyer and a Berlin headquarters guaranteed only through 2029.

Closing is eighteen months out. A lot of regulators get a vote between now and then.