The Broadline Two-Step: Sysco Watches as US Foods Circles PFG

I spent Friday afternoon on the phone with a buyer at a 40-unit casual-dining chain who reads the broadline trades the way most operators read the weather. He’d seen Bloomberg’s report that US Foods had approached Performance Food Group about a takeover before I had. His first reaction wasn’t about price or antitrust. It was: “So who owns my reorder data in eighteen months?”

That’s the question the headline arithmetic buries. The combined entity would do roughly $98 billion in annual sales — comfortably above Sysco’s $64.6 billion — and PFG shares jumped 6-10% on the Bloomberg story Friday, with both sides reportedly agreeing to share data as a first step. The M&A math is the easy part. The interesting part is what a US Foods/PFG combination does to the AI-inventory race that’s been quietly running underneath the broadline business for two years.

The number is a head fake

Every broadline story in 2025 starts with the same scoreboard: Sysco at $64.6B, US Foods second, PFG third. A US Foods/PFG deal flips that scoreboard for the first time since Sysco’s modern form took shape in the 1980s. Sysco loses the “biggest” line of its sales deck.

That matters less than people think. Sysco’s moat was never just scale; it was the density of routes and the depth of the SKU file inside any given metro. Combining US Foods and PFG creates national scale on paper, but the operational lift of merging two route networks — different DCs, fleet ages, teamsters’ contracts, ERPs — is the kind of work that takes three years and leaks customers the whole way. The Sysco playbook over the next twelve months is going to look like every incumbent watching a competitor try to digest an acquisition: open the wallet, poach the accounts that get left waiting on a Tuesday delivery.

What Sysco should be worried about isn’t the top-line number. It’s what US Foods does with PFG’s data on the way to the integration. PFG owns the Vistar convenience-and-vending business, Reinhart foodservice, and the Performance Foodservice broadline arm — three different demand signals across three verticals. Bolt those onto the demand-forecasting tooling US Foods has been building on its own ordering data and you have a training set no other distributor can touch.

That’s the AI-inventory race the trade press keeps treating as a footnote.

Mark interpretation: data agreement first, deal second

Read the Bloomberg framing carefully. The two sides reportedly agreed to share data before agreeing to terms. In M&A terms that’s a standstill-plus-NDA, normal enough. In broadline terms it’s something more interesting: a tacit acknowledgement that the asset being valued isn’t the warehouses or the trucks. It’s the demand history.

I’d mark this as the first deal in the sector where the AI-inventory thesis is being priced into the data-room work before the financial work. A US Foods CFO who tries to price PFG on the truck fleet and the customer book alone is leaving money on the table — and from the structure of the approach, they know it.

That’s why a forthcoming May piece on Sysco’s own software stack matters more than it might look on the surface: the question isn’t whether Sysco has the tools to compete on AI-inventory; it’s whether Sysco has the data depth to compete with what US Foods would have after digesting PFG. The answer in 2025 is probably yes — Sysco has more accounts and more SKUs than anyone — but the gap closes fast if the merger goes through.

What this does to the operator

For the buyer I talked to Friday, the practical question is leverage. He buys from two of the three players. If two of them become one, his negotiating posture gets worse before it gets better, and the only thing that meaningfully offsets that is his own data leverage — order history, forecast accuracy, the ability to bid the same SKU file across two vendors at once.

An upcoming May piece on DoorDash’s SevenRooms acquisition argues the front-of-house data layer is consolidating around two or three platforms. The back-of-house layer is, if this approach moves, doing the same thing. Operators who haven’t been auditing what their POS and their distributor know about them — versus what they know about themselves — are about to find out the answer is “not much.”

The defensive move is unglamorous: pull your reorder history out of the broadline portal monthly, keep a parallel file your distributor can’t see, and use it the next time the rep walks in with a “new program.” If the combined company forms, that file is your only leverage. If the deal collapses — and most deals at this scale do — you still have the file, and the exercise of building it is the part of the AI-inventory race the operator can actually win.

What to watch

Three things over the next ninety days. First, whether the data-sharing agreement leaks into a formal LOI or quietly dies. Second, whether Sysco does anything offensive — a tuck-in of a regional broadliner, a public AI-inventory product launch — to muddy the narrative. Third, whether either party puts a number on the data asset in any public framing. The day a distributor explicitly values its demand history on a per-account basis is the day the AI-inventory race stops being quiet.

The trade press will cover the antitrust angle and the truck count. They’ll undercount the training set. Don’t.

— Hana edits the newsroom for TableTransfers. Tips: [email protected].