Toast Just Bought the Front Door: Inside the Amex Deal, Toast Go 3, and Sydney

I read the Toast Q2 print on the train back from a brewery operator in Long Island City who had spent forty-five minutes telling me, with the precision of a man who has been burned, exactly why he was not switching off Resy. Then I opened the Toast 8-K exhibit on EDGAR, then the BusinessWire release, then — the one that actually made me put down my phone — the American Express partnership announcement on Toast’s own newsroom. The brewery operator’s loyalty to Resy was the most relevant piece of background I had walked into the day with. By the time I got off at Court Square he was, structurally, a Toast customer who didn’t know it yet.

The contrarian read on Toast’s Q2 is this: the ARR number is a distraction. The beat is real — $1.9 billion in annualized recurring run-rate, up 31% year over year — but it is the trailing indicator. The leading indicator is that on the same morning Toast reported earnings, it announced a partnership that gives it a structural claim on the table-stock reservation flow of every Amex-card-using diner in the United States, the launch of a new generation of front-of-house hardware, and its first Australian customer. Those three things, read together, are not a quarter. They are a pivot. Toast is no longer trying to be the POS for restaurants. It is trying to be the demand-generation layer that sits in front of the POS — and that, increasingly, decides what the POS gets to ring up.

If you are an operator, you should plan your next eighteen months as if Resy and Tock booking data is, by the back half of 2026, flowing into Toast capacity decisions in your dining room. Not because Toast has said that. Because Toast has just bought the front door, and there is no commercial reason on earth to leave the front door unconnected to the rest of the house.

The ARR number is the trailing indicator

Let me get the financial gravity out of the way, because the numbers are not small and ignoring them would be dishonest.

ARR of $1.9 billion, up 31% year over year. Gross payment volume of $49.9 billion in the quarter, up 23%. Net income of $80 million. Adjusted EBITDA of $161 million. A record 8,500 net new locations added in the quarter, bringing total locations to roughly 148,000. The enterprise, international, and food-and-beverage retail segments — the verticals Toast spent 2024 explaining to a skeptical analyst community — collectively crossed ten thousand live locations, and Toast disclosed a new 1,300-plus-unit chain win whose name they did not put in the headline. The full set is laid out in the BusinessWire announcement.

Those are good numbers. Yahoo Finance’s coverage of the Australia launch and Amex deal leaned, predictably, on the location count and the buyback authorization that came with the print. Sell-side will quote the EBITDA. The bull case is intact and the bear case — that the SMB restaurant TAM is going to choke Toast’s growth in 2026 — got harder to make this week, not easier.

But I want you to notice the shape of the beat. Eight thousand five hundred net adds is a record. Restaurants do not, in aggregate, decide to switch POS systems on the same week. The deal flow that produced 8,500 net adds in Q2 was set in motion in Q4 of last year and the first half of this year, with implementations bunching toward the end of the quarter. The number tells you about Toast’s sales motion six to nine months ago. It tells you almost nothing about what Toast intends to sell six to nine months from now.

The Amex deal tells you about the second thing.

What Toast actually bought from Amex

Read the partnership announcement carefully. The Toast newsroom version and the Restaurant Business write-up say slightly different things, in the way that one of them is the press release and the other is the trade reporter trying to figure out what the press release actually means.

The mechanical commitment is this: Amex’s Resy and Tock platforms will be integrated with Toast’s POS. Reservations made by an Amex cardholder — and, by extension, by anyone using Resy or Tock at all, because Amex does not have a separate reservation rail for non-cardholders — will land in the Toast back-of-house. The partnership covers loyalty integration, on-premise payment acceptance for Amex cards across the Toast estate, and what the release calls “more seamless and personalized experiences for guests.” The language is deliberately soft. The plumbing is not.

There is a quote in the Toast release from Alex Drummond, who runs Membership Portfolio Services at Amex, that is doing more work than it looks. “Our Card Members spent over $87 billion on dining in the U.S. alone in 2024,” Drummond said. “This partnership is a natural evolution of our investment in this growing category.” Eighty-seven billion is a number designed to make you stop reading and think about denominator. Toast did $49.9 billion in GPV across all categories in one quarter; Amex did $87 billion in dining alone across one year. The two numbers are not directly comparable, but they tell you why Amex is interesting to Toast. Amex’s dining spend is a structural slice of the high-margin, full-service, reservation-led restaurant economy that Toast historically under-indexes on.

Restaurant Business read the deal as a reservations play and they are right as far as that goes, but the framing undersells it. The thing Toast just bought is not a reservations product. It is a demand signal — visibility into who is about to walk in the door, what they have spent before, what they ordered, what their party size is, what they tip, and which other restaurants in the city they have been to in the last ninety days. That is the dataset Amex has spent twenty-five years assembling on Resy, and now Toast is wiring it into the back-of-house screen where the kitchen, the GM, and the prep cook plan their evening.

I covered the strategic logic of the Amex stake in Resy in an upcoming spring piece — the short version is that Amex bought Resy in 2019 not for the booking fee but for the panel data, and has been waiting six years for a POS partner large enough to give the panel commercial weight. Toast at 148,000 locations is that partner. There was not really a second option.

”POS for restaurants” was always going to lose to “demand layer for restaurants”

This is the part of the analysis where I have to ask you to hold two things in your head at once.

The first thing is that POS is, structurally, a commoditizing business. The terminal is glass and a chip. The software is increasingly table-stakes. Margin in pure POS has been declining for ten years and there is no force in the market that reverses that. Toast knows this. Square knows this. SpotOn, Lightspeed, and the smaller players know this. Everyone is racing for the adjacencies — payroll, capital, banking, loyalty, advertising — because the ring-the-bell business is becoming a feature.

The second thing is that the most defensible adjacency, the one with the highest gross margin and the deepest moat, is demand. Whoever owns the booking flow owns the relationship with the guest before the guest enters the door. That position lets you charge the restaurant for the seat — OpenTable’s model — or it lets you charge the brand for the data — Amex’s model, via Resy. Either way, the dollar that buys the cover is more valuable than the dollar that processes the cover.

Toast has been edging toward the demand layer for at least two years. The ad product. The marketing automation. The loyalty stack. The work on Toast IQ, which I covered in a forthcoming May piece, is the data-science backbone for this whole motion. But until this week Toast did not own a reservation rail and could not credibly claim to be the front door. The Amex deal solves that without Toast having to buy OpenTable or build a Resy competitor from scratch — both of which would have been financially or strategically prohibitive.

So when I tell you Toast’s Q2 wasn’t about ARR, this is what I mean. The ARR is the engine room. The Amex deal is Toast quietly buying the bridge.

Toast Go 3 is a positioning shot, not a hardware refresh

The Toast Go 3 launch landed in the same news cycle and got the least analyst attention, which I think is a mistake. The trade press treated it as a hardware refresh — faster chip, better battery, contactless improvements — and moved on. The Yahoo Finance roll-up filed Toast Go 3 under “product cadence” and didn’t dig further.

But hardware in this category is a positioning artifact. You don’t ship a new handheld unless you have decided where the operator’s labor model needs to go in the next forty-eight months, and Toast Go 3 is shaped by two specific bets.

The first bet is that the line between front-of-house and back-of-house, in full-service restaurants, is going to keep collapsing into a single role armed with a handheld. The kitchen display gets its data from the same device that took the order, and that device has now been integrated, at the protocol level, with the reservation system upstream. The server walking up to your table in 2026 will know, before they say good evening, that you are table 14, that you booked through Resy three weeks ago, that you are an Amex Platinum cardholder, that the last time you ate at the sister restaurant in Brooklyn you tipped 22%, that your party is 7:45 with a hold on the second cover, and that the kitchen is currently 12 minutes behind on entrees. That entire stack — reservation, identity, loyalty, kitchen state — is delivered through the Go 3.

The second bet is that quick-serve and fast-casual operators want the same hardware as full-service, because the labor cost of training on two different terminal generations is no longer worth the savings on the cheaper one. Toast Go 3 is the unified device. Pricing was not disclosed in the launch material I’ve seen, but the strategic implication is that Toast is willing to subsidize the hardware to standardize the fleet, because the fleet is what makes the demand-layer monetization work.

The thing to understand about Toast Go 3 is that it is the physical surface of the strategy the Amex deal is the data surface of. Both go live this fall. Both reinforce the same thesis.

Australia is not a flag-plant. It is a forcing function.

The Australia announcement — Graze Craze, a quick-service brand, as Toast’s first Australian customer — got framed by Yahoo as the international expansion story. It is. But the way to read Australia is not “Toast has decided the U.K. and Ireland are saturated so it’s adding TAM.” The way to read it is “Toast has decided to test its full stack in a market where it has no incumbent advantage.”

Australia is a hard market. The reservations rail in Australia is dominated by TheFork and a fragmented set of local players. Amex penetration is meaningful but not dominant. Card present economics are different. Tipping culture is different — which means the loyalty primitives Toast built for the U.S. need to be reworked. The labor laws are different, which means the payroll product Toast spent two years building for the U.S. ESL workforce does not port cleanly. And restaurants in Sydney and Melbourne run, on average, smaller and more chef-led than the average U.S. independent.

Picking Australia first, ahead of Germany or France or Spain, tells you Toast is willing to absorb friction now to prove the full stack works in a non-U.K., non-U.S. market. The single-customer launch — Graze Craze, a U.S.-origin chain expanding into the region — is the friendliest possible beachhead. It is also a fine choice if the actual strategic goal is to harden the international product against an Australian operator base before Toast tries Germany.

Mark interpretation: Australia is the forcing function. It is where the international product gets stress-tested in 2026 before Toast walks it into Continental Europe in 2027. The Graze Craze customer matters less than the engineering discipline that the Sydney launch imposes on the team for the next twelve months.

The 1,300-unit chain win, and the enterprise read-through

Buried in the Q2 commentary is a line that should have been a headline: Toast signed a new enterprise customer of more than 1,300 units. Toast did not name them. The market did not seem to care. The market is wrong.

Enterprise, international, and food-and-beverage retail crossed ten thousand live locations in the quarter, which means that segment is now roughly seven percent of Toast’s total footprint and growing faster than the SMB base. The 1,300-unit win, on its own, is more locations than most regional POS players have in their entire pipeline. And enterprise wins at this scale do not happen because of price. They happen because of the integration surface — the ability to plug Toast into the chain’s existing loyalty program, the chain’s existing reservation system, the chain’s existing kitchen display, the chain’s existing payroll provider — without ripping out infrastructure the operations team has spent five years tuning.

The Amex partnership materially expands that integration surface. So does Toast Go 3. So does the Australia launch, which forces the engineering team to make the platform genuinely portable rather than U.S.-only with a U.K. veneer.

I went deep on Toast’s enterprise positioning in a forthcoming desk review and the short version is that Toast won the SMB segment by being the only opinionated full-stack product in the market, and is winning the enterprise segment by being the only product whose opinion is now flexible enough to fit the chain’s own opinions. That is a difficult posture to hold. The Q2 numbers suggest Toast is holding it.

What this means for the operator

If you run a single restaurant or a small group, here is what to do with this week’s news.

First, take the reservations-data implication seriously. The Amex/Toast deal does not switch on overnight, and the privacy and contracting work between Amex and Resy’s existing restaurant clients will be its own multi-quarter exercise. But by mid-2026 it is reasonable to expect that any Toast operator with a Resy or Tock integration will, by default, have booking-level guest signal flowing into the Toast back office. You should be planning your guest-experience workflow as if this is true. That means thinking now about which staff member is responsible for acting on the signal — the server, the GM, the host — and how you want to train them. The technology will be ready before your team is, which is the failure mode for every front-of-house data product in the last fifteen years.

Second, if you are on Resy or Tock and not on Toast, the calculus for switching POS just changed. I am not telling you to switch. I am telling you that the integration tax of running a non-Toast POS alongside an Amex-owned booking rail will start showing up in your operations in 2026, and you should price that into your renewal cycle.

Third, if you are on Toast and using a different reservation platform — OpenTable, SevenRooms, Yelp Reservations — you have a different problem. The Amex partnership is structured as an exclusive integration on the Toast side for Resy and Tock; nothing in the public release suggests Toast is killing its other reservation integrations, and there is no commercial reason for Toast to do so. But the premium features — the loyalty linkage, the cross-restaurant guest history, the Amex spend signal — are by definition going to be richest on the Resy/Tock side. The operator who runs SevenRooms on top of Toast in 2026 will get a degraded version of the product compared to the operator who runs Resy on top of Toast. That is a fact you need to model into your tech-stack roadmap.

Fourth, hardware. If you are due for a handheld refresh — and most of the operators I talk to who were early on Toast Go 2 are now four-plus years into that hardware — get on the Toast Go 3 waitlist now rather than later. The supply ramp on a new handheld generation in this category is always tighter than the vendor admits, and you do not want to be the operator trying to deploy in the middle of a Q1 2026 holiday-hangover rush against a back-ordered device.

Fifth — and this is the operator-level question that gets least asked — start auditing your guest data hygiene now. The Amex/Toast deal is going to put richer signal on your screens, but it is also going to put your guest data into a partnership-shaped data flow that I do not believe most operators have thought about. Know what data you have, where it sits, what consent you have for which uses, and who in your org is empowered to make decisions about it. By the time the integration goes live, the operators who win on this product surface will be the ones who walked in with their guest CRM already organized.

The market read

Sell-side is going to spend the next two weeks revising 2026 numbers on the back of the ARR beat and the buyback authorization. That work is necessary and not very interesting. The more interesting model adjustment is on the take-rate side: the Amex partnership is the kind of deal that, over a multi-year horizon, lets Toast monetize the demand layer in ways that are not yet in any analyst’s spreadsheet. Loyalty-program revenue share. Co-marketed cardholder offers. Premium reservation features. Demand-shaping ad units. None of those are big numbers in 2026. All of them are big numbers in 2028 if the integration works.

Mark interpretation: the right way to value Toast off this print is not to ask how much ARR comes from 8,500 net adds compounding. It is to ask what the gross margin and growth profile of the demand layer looks like in 2028, and how much of that demand-layer economics Toast captures versus Amex captures. The split is currently undisclosed and probably under-defined even inside the partnership. My working assumption is that Toast captures more of the long-tail demand monetization than the market currently assumes, because Toast owns the in-restaurant surface where the loyalty redemption actually happens, and that surface is where pricing power sits.

The thing the brewery operator should have asked me

The brewery operator in Long Island City — the one I started this piece with, sitting across from me at a steel table with a flight of pilsners and the patient certainty of someone who has been pitched by every POS company that ever existed — asked me whether Toast was going to raise prices on him next year. That is the wrong question. The right question, the one operators should be asking themselves the morning after a quarter like this, is: what is the product I am buying from my POS vendor going to be in eighteen months, and am I going to like it?

Toast’s answer, as of August 5, 2025, is that the product you are buying from them is going to be the front door of your restaurant. Not the cash register at the back. Not the kitchen display at the pass. The front door. The booking, the recognition, the loyalty pull, the upsell at the host stand, the demand shape of your Tuesday at 7:15, the reason the four-top with the Platinum card chose your room over the one three blocks south.

If that is the product Toast intends to sell, then everything about how you think of Toast — and how you negotiate with Toast, and how you stack other vendors around Toast, and which other vendors you keep — needs to be recalibrated. The Q2 print is the financial frame. The Amex deal, Toast Go 3, and the Sydney launch are the strategic content inside that frame. Read them together. The ARR will take care of itself.

— Priya covers operators for TableTransfers. Tips: [email protected].