Inspire Brands picks its IPO bankers. Every franchised-QSR multiple gets re-anchored if this prices.

It’s a Friday afternoon, the kind where the desk has gone quiet because the bond guys left at three for the long weekend, and the only thing keeping me at the screen is a Bloomberg headline that landed at 2:47 PM Eastern with the kind of byline pairing that tells you the leak was real and the leak was sourced from inside the syndicate. Roark Capital, per Bloomberg’s April 17 story, has selected JPMorgan, Bank of America, Barclays, Goldman Sachs, and Morgan Stanley to lead Inspire Brands’ US IPO. Target raise of around $2 billion. Target valuation in the neighborhood of $20 billion. The brands underneath — Dunkin’, Arby’s, Buffalo Wild Wings, Baskin-Robbins, Sonic, Jimmy John’s — are the entire franchised-QSR comp set rolled into one filing.

I read it twice. Then I write a line in my notebook: if this prices at $20B, every franchised-QSR multiple in the public tape gets re-anchored. That is the contrarian read. It is also the entire reason this filing matters more than any single brand inside it.

The press will write the headline as “Dunkin’ goes public.” That is wrong. The headline is that Roark has assembled the largest single-asset multi-concept franchised-QSR public comparable in fifteen years, and the bankers have priced it where they have priced it because they had to clear five other names’ multiples to get there. Once it lists, those other names re-rate. Up or down. Not by a little.

Let me walk it.

The deal in plain English

Roark has picked the syndicate. That is the news. The five banks named — JPM, BAML, Barclays, Goldman, Morgan Stanley — is a full bulge-bracket lineup, not a left-lead-plus-pacifiers structure. That tells you two things. First, Roark expects the book to be deep enough that the syndicate’s job is order-collection, not order-creation; you don’t bring five top-tier names onto a deal that needs handholding. Second, the fees are getting spread thin enough that every one of those banks has to want a relationship outcome — research coverage, follow-on access, the next Roark name — more than they want the single-deal economics. That is a sponsor with leverage running a sponsor’s process.

The raise is $2B. The valuation Roark is steering toward, per the same reporting, is $20B. Both numbers are anchors, not commitments. The S-1 has not been filed. The range has not been set. But the way the Bloomberg desk has been allowed to print those numbers — sourced, on the record as “people familiar,” not denied — is itself a piece of information. Roark is telegraphing the target before the confidential S-1 lands. Anyone who reads the leaks for a living reads that as a confidential filing inside thirty days, with the public marketing window in late summer.

The brands inside the wrapper — Dunkin’, Arby’s, BWW, Baskin-Robbins, Sonic, Jimmy John’s — together cover roughly 32,000 units across the US, with international optionality on Dunkin’ and Baskin in particular. The exact unit count and the exact systemwide sales number are going to be the line items the buy side fights over in the roadshow. But the shape is clear: a franchised-heavy, multi-concept, multi-daypart QSR holdco at the scale of the biggest public peers, with a tech estate that has been quietly built out in private for the better part of a decade.

That is the asset. The question I keep coming back to is what its listing does to the rest of the tape.

The QSR comparables reset, brand by brand

This is the part where the math runs against the press narrative. The press narrative is that Inspire is a Dunkin’ IPO. The math says it is six IPOs in one filing, and each one re-anchors a separate slice of the public comp set.

Dunkin’. The crown asset. ~9,500 US units, dominant on the morning daypart, a coffee-and-bakery business with international franchise optionality Roark has not fully harvested. The public comp is Starbucks — a different model, heavily company-owned, premium positioning — but the category-leader read-across is unavoidable. Interpretation: if the roadshow prints a Dunkin’ segment EBITDA that supports a ~$10–11B sleeve of the $20B target, the implied multiple lands inside the Starbucks zone on a royalty-stream basis. That re-anchors SBUX downward at the margin. Small effect, not zero, and the SBUX research desks know it.

Arby’s. ~3,400 units, the brand that anchored the original Roark thesis. Sandwich QSR, beef-led, sparse public comp, with the natural read-across to Wendy’s. Interpretation: a $2–3B sleeve clears a multiple Wendy’s stockholders will not enjoy if it lands below where Wendy’s trades. A defensive research piece follows inside two weeks of the S-1 hitting.

Buffalo Wild Wings. ~1,200 units, the casual-dining trouble child, with sports-betting tie-ins driving a partial comp recovery but flat unit growth. Interpretation: the roadshow’s hardest sell with the long-only buy side. Casual dining at scale, in a public wrapper, is a category nobody has wanted to own since 2019. If Inspire prices BWW at anything approaching a clean multiple, Texas Roadhouse, Bloomin’, and Darden’s mid-tier brands all get a research note asking whether the print implies a segment re-rate. Most likely answer: no, BWW is sui generis. But the question gets asked.

Baskin-Robbins. ~2,500 US units, ~7,500 globally, a near-pure-royalty business with the cleanest cash conversion in the portfolio. Interpretation: the sleeve that gets the highest multiple inside the SOP, and the one the buy side will most want to spin out conceptually.

Sonic. ~3,500 units, drive-in concept, heavily franchised, a Southern and Plains-states footprint the coastal buy side will under-appreciate. Interpretation: the analog is Wingstop on unit economics and Texas Roadhouse on geography. Neither is clean. The roadshow will sell Sonic on its drive-thru tech refresh and the throughput math the AI scheduling layer has produced post-2024.

Jimmy John’s. ~2,700 units, the sandwich-delivery concept with the cleanest throughput model in the file. Interpretation: the brand the buy side has the least public information on, and the one where roadshow disclosure will be most informative. A strong segment EBITDA print here revises the multiple on the entire sandwich-QSR sub-sector.

Six brands. Six separate re-anchoring events. All of them inside one filing.

The Roark math: 2018 take-private cost vs. 2026 valuation

The other piece of the trade that nobody is going to write up cleanly in the first wave of coverage is the Roark IRR. Inspire was assembled by Roark across a series of deals — the Arby’s platform was the original Roark vehicle, BWW closed in 2018 at roughly $2.9B enterprise value, Sonic followed in late 2018 at roughly $2.3B, Jimmy John’s came in 2019, and the Dunkin’ take-private in December 2020 was the biggest at ~$11.3B EV including debt. Stack those, layer in the equity contributions and the dividend recaps in between, and the Roark cost basis on Inspire’s equity is — and this is interpretation, marked as such — somewhere in the $10–12B zone.

A $20B exit valuation on a $10–12B cost basis, struck over a 6-to-8-year hold depending on how you weight the contributions, is roughly a 1.7–2.0x equity multiple before considering the cash distributions Roark has already taken via recaps. That is a fine outcome for a sponsor. It is not the home-run outcome the headline number ($20B!) implies in isolation.

Mark this: the Roark IRR on Inspire is going to look good in the press because the gross number is big. The actual IRR, net of fund fees and the time-cost of the capital across the hold, is probably in the high teens. That is a good fund outcome. It is not the outcome you would model if you only read the Bloomberg headline.

The reason this matters for the trade is that Roark is not pricing the IPO to maximize their own exit — they are pricing it to clear the deal in a window that the rest of the QSR tape can absorb without choking. A $20B target leaves room for the after-market to work. A $25B target would not. The five-bank syndicate is there to protect the book against a too-aggressive print that breaks the after-market and damages the QSR comp set on the way out. This is sponsor discipline expressed as bank selection.

The AI angle across 32,000+ units — interpretation flag

Here is where the contrarian read gets sharper, and where I want to mark the work as interpretation in bright red.

Inspire has spent the last four years quietly building out a unified tech stack across the portfolio. Not every brand is on every layer, and the franchisee adoption rates vary by concept and by region, but the direction is clear: one POS platform across most of the corporate-influenced footprint, one digital-ordering and loyalty spine pushing toward consolidation, one demand-forecasting and labor-scheduling layer being rolled out concept by concept. The result, by the time the S-1 lands, is a holdco that can credibly tell the buy side it operates at AI-enabled scale across 32,000+ units.

Interpretation: that story is worth roughly 1.5-to-2.5 turns of EBITDA on the blended multiple, against a comp set where most of the public franchised-QSR names cannot tell the same story without significant caveats. Call it $2–4B of the $20B valuation that is uniquely underwritten by the AI/tech narrative, not by the brand-level cash flows.

The reason this matters is the same reason the Olo App valuation thesis matters: a multi-brand holdco with a unified consumer surface and a unified data spine has network-effect economics that no single-concept QSR has. Inspire is the only public-bound name in 2026 that can tell that story credibly at scale. McDonald’s is bigger but is single-concept on the brand side. Yum is multi-concept but the brands are siloed on tech. Restaurant Brands is in between but has not yet rolled the integrations. Inspire is the one filing where the synergy story is both real and large.

If the buy side credits the synergy story, Inspire prices toward the top of the $20B band, possibly through. If the buy side does not, the deal clears closer to $16–17B, and the AI premium across the entire QSR tape compresses by a turn or more. That is the trade the next sixty days will resolve.

The bet: Inspire prices, FAT Brands re-prices, NRP Florida sells better

Here is how I’d lay out the cross-asset trade, with the appropriate caveats for a deal whose S-1 has not yet hit the EDGAR system.

Inspire prices inside the $19–21B band. The five-bank syndicate, the leak-controlled marketing run-up, and the Roark discipline on the target multiple all point to a deal that clears. The risk is not pricing — it is timing. If the broader IPO window closes between now and the listing date, the deal slips into Q4 and the comp-set reset gets pushed into 2027. Base case: the confidential S-1 lands within a month of the Bloomberg leak, the public filing follows in mid-summer, and the deal lists in the September-to-October window.

FAT Brands re-prices on the back of the Inspire print. I wrote on the FAT Brands Chapter 11 auction two weeks ago that the bid-depressing factor was tech debt, not comp trend, and that a strategic buyer with a credible standardization plan would clear the auction in the $700–800M zone. The Inspire IPO changes the upper bound of that range. A public-market comp at $20B for a tech-enabled multi-concept franchised-QSR platform gives the strategic bidder a defensible read-across. FAT Brands clears 10–15% higher than it would have without the Inspire calendar. The forthcoming Q2 M&A roundup tracks where that lands.

NRP Florida sells better. The Applebee’s franchisee bankruptcy and the Dine Brands stalking-horse bid were, structurally, a question about what a public franchisor pays for territory defense at distress. A higher public comp on the franchised-QSR tape raises the implied territory value across the entire franchise universe. Dine pays the stalking-horse price comfortably. A competing bid, if one surfaces, comes in higher than it would have in March.

The forthcoming confidential S-1 confirms the path. I expect the confidential S-1 inside thirty days of the bank selection, putting the filing in the first half of May. The counter-take on the AI premium and the forthcoming DoorDash/SevenRooms $1.2B transaction both sit in different layers of the same comp-set reset — a marketplace pricing a unified data spine, an IPO pricing a unified holdco. Both prices land within ninety days of each other. Both reset different sides of the same tape.

The bet, distilled: Inspire prices. The QSR comp set re-anchors. FAT Brands clears higher. NRP Florida sells better. The AI-premium narrative across franchised-QSR multiples gets reinforced, not weakened, by the largest single piece of public disclosure the category has produced in fifteen years.

It is a long weekend. I am going to print the Bloomberg piece, mark up the brand-by-brand math by hand, and have the segment-level model ready for the moment the confidential S-1 surfaces. The wider Inspire confidential filing coverage — when it lands — will be the next data point on this trade. Until then, the bank selection is the tell, and the comp-set reset is already in motion.

— Marcus edits The Bottom Line for TableTransfers. Tips: [email protected].