Against the AI Premium: Why Toast's Q4 Won't Justify a Re-rating

Friday morning, second coffee, three sell-side previews of Toast’s Q4 print fanned across the desk. They all want the same thing. They want next Thursday’s earnings to be the catalyst that finally pulls Toast out of the fintech bucket and reprices it as a software platform. Toast IQ is the lever in every memo I’ve read this week. “AI-attach is accelerating,” one of them says. “Software ARR mix is the re-rating story.” Another one literally walks through a sum-of-the-parts where Toast IQ alone is worth eight to twelve dollars of share price.

I think they’re going to be disappointed, and I think they’re going to be wrong about why.

The thesis, plainly. The Feb 12 print will probably be clean. ARR will plausibly cross two billion. Location adds will be healthy. The market will get its AI-attach commentary. None of that justifies a re-rating, because the gravity holding Toast at roughly two and a half times sales isn’t a missing AI story — it’s the structure of the business. Toast’s revenue is dominated by fintech take rate that has a hard ceiling, its payment volume runs through a vendor concentration that nobody on the sell side wants to talk about, and the CEO himself has been the most honest voice in hospitality tech about what AI does to software pricing. If Aman Narang isn’t going to underwrite the software-multiple narrative, I’m not going to either.

This isn’t a short call. I don’t short platforms with positive operating cash flow and a stronger CEO than most of their comps. It’s a setup call: the people positioning into Thursday for the re-rating trade are paying for an outcome the business won’t deliver in this print or the next four.

What we actually know going in

Toast reported Q3 2025 on November 4, 2025. The headline numbers are what the buy side should be anchoring on this morning, not extrapolating from. Revenue came in at $1.63B, up 25% year over year. ARR crossed $2.0B for the first time, growing 30% YoY. Locations hit roughly 156,000, up 23%. Gross payment volume was $51.5B in the quarter, up 24%. Free cash flow was $153M for the quarter against $97M a year prior. GAAP income from operations was $84M versus $34M (Toast Q3 2025 press release, Nov 4 2025).

Read the composition. Subscription revenue grew 29%. Subscription gross profit grew 32%. Payments — fintech — ARR grew 31%, and fintech gross profit grew 35%. Fintech is still growing faster than software. That’s the entire valuation argument in two numbers. When the sell side tells you Toast deserves a software multiple, what they’re saying is that the slower-growing line item should price the asset.

The Q4 print Thursday will likely show ARR closing the year somewhere north of $2.0B. The locations number will probably show another ~8,000 net adds. Full-year revenue will land in the low six-billions. The buyback authorization may get refreshed. None of that is news. None of that is a re-rating.

What would be a re-rating? Software ARR growing meaningfully faster than fintech ARR for two consecutive quarters, with Toast IQ attach disclosed as a specific dollar contribution rather than narrative color. I haven’t seen the disclosure framework that gets us there, and I don’t think Thursday brings it.

Argument 1: The take-rate ceiling is the actual ceiling

The structural reason Toast trades at ~2.4x sales is that roughly two-thirds of revenue is payments take rate. I’m not going to relitigate the whole math; I did some of it in the indie buy-side note. The short version: payments revenue is a spread on gross payment volume, and that spread has a hard ceiling defined by interchange, scheme fees, and the merchant’s tolerance for vendor-side capture.

Q3 GPV of $51.5B against fintech gross profit growth of 35% YoY is excellent execution against a fixed ceiling. It is not a software business. A software business compounds margin as it scales because the cost of serving the marginal customer falls toward zero. A fintech business compounds dollars as GPV scales because the take rate is roughly flat. Those are different multiples. The market is not confused about which one Toast is. The sell side is.

Toast IQ does not change this in any version I can construct. Toast IQ is a feature layer sold to merchants who are already on the platform. To the extent it drives subscription ARR — which is the line item we should care about — it has to grow that line item meaningfully faster than the fintech line for the multiple to move. In Q3 it did not. Subscription ARR grew 28%; payments ARR grew 31%. The mix shift is going the wrong direction for the bull case.

Judgment, flagged: I have been told by three separate sell-side analysts since November that “the mix will turn in 2026.” I would like to see one quarter where it actually does before I underwrite the eight-to-twelve-dollar Toast IQ contribution.

Argument 2: The CEO’s own framing is the tell

The most underweighted public commentary in hospitality tech right now is Aman Narang’s own framing of AI. He has been remarkably consistent across earnings calls and interviews for about eighteen months: AI is an opportunity for Toast because Toast is a platform, not a software vendor.

That sentence is meant to sound bullish. It is not. Parse it.

What Narang is conceding — quietly, repeatedly, in language that the sell-side memos this morning are choosing to file under “vision” — is that the software layer alone is not where the durable economics live. He’s framed Toast as software, hardware, fintech, and partner ecosystem. He has used some version of “Toast is more than a software provider” on at least three calls I have transcripts of going back through 2025. That is a CEO telling the market not to value him on software multiples.

He’s right, by the way. The reason Toast IQ is interesting at all is that it’s a wedge to drive attach across the platform, including the hardware refresh cycle and the fintech-adjacent products. Narang has talked about Toast IQ supporting “increasingly complex workflows that could eventually do some of the work themselves” — which is the right vision, but it is also the most honest possible articulation of the commoditization risk. If the AI agent does the work, the customer is paying for the outcome, not the SaaS seat. Outcome-based pricing in hospitality tech is going to be brutal on per-seat economics. Narang knows this. He’s not pretending otherwise. The buy side should listen.

The piece my colleague wrote on Sous Chef-to-Toast IQ (the pilot was the point) caught the operational version of this. Toast learned in the pilot that operators don’t want a chat box that tells them things; they want one that does things. Doing things is great for retention and attach. It is harder to price like SaaS.

Argument 3: The buyback is not the signal it looks like

I expect a refresh of the share repurchase authorization on Thursday or shortly after. The original 2024 authorization has been worked down, and a board that has watched the stock chop around for a year wants optionality. Fine. That is not a re-rating event.

Free cash flow is real and growing — $153M in Q3 alone, on a trajectory that puts the full year comfortably north of $500M when Thursday’s numbers land. A buyback authorization is the rational use of that cash if you believe — as Narang clearly does, and as I do — that the stock is range-bound at this multiple for structural reasons.

But a buyback is not a multiple-expansion event. It is a per-share-arithmetic event. When Toast prints Q4 and announces an incremental authorization, the stock will probably trade up two to four percent on the open and give half of it back by Friday. Anyone telling you the buyback “validates the AI thesis” is selling you something. Cash generation at 2.4x sales is a value story dressed up in a growth wrapper, and value stories don’t get re-rated on AI narrative.

Argument 4: What I’d want to see — and probably won’t

Three things would move my view on Thursday. None of them are in any of the previews on my desk.

One: a disclosed Toast IQ ARR contribution. Not “AI-attach is accelerating.” Not “increasing penetration across our base.” A dollar number, ideally with a year-ago comparable. If Toast IQ is contributing $150M or more to ARR at this point — fifteen months after the pilot started, four months after general availability — I’d update meaningfully toward the bull case. If they don’t disclose it, the most likely reason is that the number isn’t big enough yet to disclose flatteringly.

Two: a stated software-to-fintech mix target. A management team confident that AI is going to bend the mix toward software would tell you what mix they’re targeting in 2027. “We expect software to be 40% of revenue by 2027” would be a re-rating sentence. I don’t expect Narang to say it because I don’t think he believes it, and his discipline about not saying things he doesn’t believe is one of the reasons I respect him.

Three: any acknowledgment of price compression in the existing software stack. The honest version of the AI thesis is that it lets Toast charge more, not just sell more. If pricing power on the subscription tiers is visible in the numbers — average revenue per location growing faster than total location count, for instance — that would be evidence. If it isn’t, then Toast IQ is volume, not price, and volume doesn’t get the multiple.

Where I might be wrong

Two places. Standard caveats.

One: I’m underweighting the operational compounding. The boring sibling story from the Sous Chef pilot — the menu-upsell tool that drove a 6% lift in average order volume at one pilot site — is the version of AI that actually works in restaurants. If Toast IQ Grow or whatever ships next bundles enough of those small productivity wins to materially change the per-location economics, attach could compound faster than I’m modeling. I’m watching net revenue retention numbers in particular. If they break above the high end of the historical range on Thursday, I owe an update.

Two: the platform-layer aggregation moves faster than I’m modeling. If Toast credibly absorbs more of the third-party stack — reservations, marketing, payroll — and starts pricing those bundles like a true platform, the take rate on the software line could move in ways the comp set doesn’t anticipate. That’s the genuine bull case, and it’s not what the sell side is selling this morning. They’re selling AI. The real story, if there is one, is platform aggregation. Different argument. Different multiple. Same conclusion through the Feb 12 print: not enough new evidence to re-rate.

The Feb 12 print will tell us. I’ll be on the call, marker in hand, watching three lines: software-to-fintech ARR growth differential, ARPU trajectory, and any specificity at all on Toast IQ contribution. If the print breaks the pattern, I’ll write it up and say so. The Bottom Line is meant to be a place where the buy-side calls its own bluff in public.

If you’re long Toast into Thursday on the AI re-rating thesis, this is the note I’d want my partner to send me before I let it ride. The setup is wrong. The CEO is telling you it’s wrong. The Q3 numbers we already have are telling you it’s wrong. Don’t pay the AI premium on a fintech.

— Oliver writes the buy-side perspective for The Bottom Line. He sits on three operator boards. Tips, especially counter-arguments: [email protected].