Your Counterparty Just Got Smaller

Quartz reported on October 2 that Burger King U.S. President Tom Curtis said the chain expects roughly 200 company-operated restaurants to change hands by year-end, down from an initial target of about 300, on the way to keeping roughly 300 of its more than 6,000 U.S. restaurants company-operated. CNBC reported the same plan. Per Quartz, Curtis said the company wants "franchisees who live and work in the communities that they serve" — smaller local operators, not the private-equity-backed platforms the brand once preferred. Per Restaurant Dive, the inventory came from Restaurant Brands International's 2024 purchase of Carrols Restaurant Group: about $1 billion for 1,022 Burger King units in 23 states, to be refranchised over five to seven years.

Our position: this is a counterparty story, not a brand story. On a Burger King site in our markets, the party deciding a lease, a relocation or a pad purchase is moving from a corporate balance sheet to an operator with a few dozen stores. That changes who we call and what the dirt is worth at exit. Underwrite the residual at the franchisee number from the first spreadsheet.

A hundred basis points sits in the guaranty

Per The Boulder Group's single-tenant net lease report for the second quarter of 2026, corporate QSR cap rates averaged 5.85 percent and franchisee QSR averaged 6.85 percent. Same category, one hundred basis points apart.

Whether a given refranchising changes the guaranty behind a given lease depends on documents that are not public, and we would not assume it does. The pricing gap is public. Ask who the obligor is before anyone quotes a cap rate on a Burger King.

The spend is going into boxes the brand already has

Per Restaurant Brands International's August 6 second-quarter release, Burger King U.S. comparable sales rose 8.5 percent in the quarter while the Burger King segment's net restaurant growth was negative 0.8 percent. The same release describes Reclaim the Flame as "investing up to $700 million through year-end 2028," including "high-quality remodels and relocations," with $194 million of up to $550 million in Royal Reset investments funded as of June 30. Its 10-Q ties the refranchising to the planned sunset of the Restaurant Holdings segment carrying the Carrols stores.

Comps up eight percent against a declining segment unit count means the capital is going into existing restaurants, not new ones. The word in the company's sentence that belongs to a developer is "relocations." A relocation is a site the current prototype cannot fit — queue depth, access, parcel shape — on a corridor the brand has already proven with its own sales. The trade area is validated, the replacement pad is usually close by, and none of it shows up on a list of new-market whitespace.

The buyer is smaller than the seller was

Carrols ran 1,022 restaurants. The operators buying now do not. Per Quartz, a franchisee group running 20 Florida locations posted 21 percent year-over-year sales growth and a 16 percent traffic increase after taking over.

Twenty stores is a different counterparty than a thousand. No in-house site-selection department, no national broker relationship, no real estate committee — and an owner who lives in the trade area and answers the phone. We would plan for a faster yes and a slower close.

What we watch for

We identify the operating entity before pricing a Burger King pad, and treat corporate and franchisee product as two different assets at exit.

We read a refranchising as a contact event. An operator who just bought a dozen boxes has a growth obligation and no pipeline, and we would rather be the first call than the fourth.

We hunt relocation candidates on corridors where remodel money is already landing. An aging freestanding box on a corner that cannot hold current stacking is the pad request, not the problem.

We do not read negative net restaurant growth as retreat. Closing weak units, reinvesting in strong ones and needing new dirt where the old parcel fails are three different things, and the net number hides all three.