Toast IQ vs Square AI Voice Ordering vs Lightspeed Capital: A POS-AI Head-to-Head

I spent Tuesday afternoon at a counter in Greenpoint with three POS demo accounts open in three browser tabs and a Q3 transcript queue on the fourth. Toast’s BusinessWire release had landed the night before. Block’s Q3 shareholder letter had hit EDGAR the same evening. The Lightspeed print was queued for tomorrow morning. The contrarian read I want to put on the table before any of the equity desks start sketching their 2026 POS-AI comp tables is this: the three public POS-AI strategies are not converging on the same product. They are diverging — Toast is building an in-store assistant pointed at the operator, Square is building a guest-facing voice and kiosk stack pointed at QSR, and Lightspeed is building an outbound-led hospitality push in Europe whose AI surface area we will only really see priced when Q2 FY26 hits the tape tomorrow. The desk review I’d hand any operator deciding between them right now is not a feature matrix. It is a surface matrix.

That is not how the sell side is reading this. Most of the Toast-vs-Square notes I’ve seen in the last forty-eight hours treat the two AI stories as competing pitches for the same operator dollar. They are not. Toast IQ is being demoed to the back-of-house manager who already has a Toast handheld in her apron. Square’s Multichannel Menus, Kiosk, and AI Voice Ordering are being demoed to the QSR owner who is trying to take a phone order off the line cook’s hands. The buyer is different. The job-to-be-done is different. The pricing surface is different. And the third option — Lightspeed’s hospitality push, particularly in Europe — is being sold by outbound reps into a buyer who has not yet been asked, in any serious volume, to underwrite an AI feature at all.

I want to walk through each of the three in the order the prints landed, mark the comparison, and then put a stake in the ground on which stack I think wins which segment in 2026. I’ll keep the math conservative and the speculation labeled.

Toast: the assistant is the equity story

The Toast Q3 numbers are the easy part. ARR crossed $2 billion (the company’s headline). GPV came in at $51.5 billion, up 24% YoY. The install base is at 156,000 locations after 7,500 net adds in the quarter. Adjusted EBITDA was $176 million on a 35% margin. The full-year EBITDA guide moved up. Those are clean prints, EPS asterisk aside.

But the thing the Toast prepared remarks spent disproportionate airtime on was Toast IQ, and the proof point management chose to hand analysts was the 25,000 restaurants / 235,000 uses disclosure for the assistant’s first month of general availability. My colleague Maya covered this in the Pass companion piece on the Q3 print and I won’t re-litigate the funnel-vs-retention reading there. What I want to mark from the Vibe Check seat is the pricing surface — because that is the part of Toast IQ that the call did not disclose and that the analyst notes are mostly not modeling yet.

Here is what we know. Toast IQ is being positioned as a bundled capability in the existing SaaS tier, not as a standalone line item, and Toast Advertising — the network the call cited Pizza by the Sea’s ~$400,000 attributable-revenue number against — is the surface area where the IQ recommendation layer is being pointed. Read together, those two facts tell you Toast is not going to charge per-query for the assistant. Toast is going to charge per outcome via the ads network, with IQ as the recommendation engine that decides which campaigns to surface to which operator. That is a fundamentally different monetization shape than the one the bulls are penciling into their 2026 model when they assume “AI seat ARPU lift” the way they did for Salesforce’s Einstein tier or for the Square AI products I’ll get to in a moment.

The implication for an operator: if you are already on Toast and you are not running on the ads network, the IQ capability is essentially free incremental software riding on the SaaS line you are already paying. If you are running on the ads network, IQ is going to start steering your spend in ways that may or may not be transparent. The desk review I’d write today says the IQ value capture mechanism is the ads-network rev share, full stop, and the equity case is whether the recommendation engine can move ads-network attach materially in 2026.

The 25,000-restaurants-in-month-one number is the early signal that the assistant is opening doors. The harder signal — and the one I’ll be watching for in Q4 — is whether the IQ-recommended ads spend on Toast Advertising starts to compound the way the locations-line did over the last decade. If it does, Toast IQ is the engine. If it doesn’t, IQ is a feature, the ads network is the engine, and the equity multiple stays where it is.

Square: the voice stack is pointed at QSR

Block’s Q3 hit at roughly the same time, and the Square line items inside it tell a different POS-AI story than the one Toast told. Square’s F&B GPV grew 17% YoY, which is the number I keep seeing the sell-side underweight in their comp tables. Total Block GPV was $67.2 billion in the quarter. Block gross profit was $1.02 billion across the consolidated company, up 9% YoY. None of those are the headline numbers — Cash App is the headline at Block — but they are the numbers that matter for the POS-AI comp.

Inside the Square segment, three product launches anchor the AI pitch: Multichannel Menus (sync menus across in-store, online, third-party delivery, and the new Kiosk in one place), the Square Kiosk hardware (self-order surface for QSR and counter-service), and AI Voice Ordering (the phone-order capture product that takes the inbound call off the line cook’s hands and turns it into a structured order ticket). The three are being sold together, and the buyer the demo is being pointed at is the QSR operator who is bleeding labor cost on the phone and on the order-taker headcount.

That buyer is not the Toast IQ buyer. The Toast IQ buyer is the back-of-house manager trying to get faster answers out of a system that already knows her menu, her ticket times, and her staff schedule. The Square Voice Ordering buyer is the front-of-house owner trying to remove a labor seat from the order-taking surface. The labor math is the pitch. If Voice Ordering can take 70% of inbound phone orders without escalation — which is the rough threshold I’ve heard cited in vendor decks across the category for the last two cycles — then the operator can either redeploy the labor seat or remove it. Either way, the ROI math is concrete and is calculable on a per-location basis.

That is a more legible pricing story than Toast IQ’s, and it is one of the reasons the Square F&B GPV +17% line is, I think, undermarked in the post-print notes. The voice stack and the Kiosk are the surface area where Block is competing for new F&B doors against Toast, against Clover, and against the regional QSR specialists. It is not competing in the existing-Toast-account upsell motion. It is competing at the install moment. And the install moment is exactly where AI voice ordering, if it works at the demo threshold, becomes a wedge.

The risk on the Square pitch is the one any voice product carries: accent and noise robustness in real kitchens, escalation handling when the model is wrong, and the menu-sync hygiene that determines whether the voice agent can even quote the right price on a special. The Multichannel Menus product is the unglamorous infrastructure under the AI demo — and it is the part of the stack I’d want to see audited before I underwrote the Voice Ordering ARR ramp. The AI demo will land or not land depending on whether the menu graph behind it is clean.

Lightspeed: the print lands tomorrow

Here is where I have to be disciplined about what I can and cannot say on November 5. Lightspeed’s Q2 FY26 print is queued for tomorrow morning, and the company has been telegraphing an outbound-led hospitality push in Europe since the F1 print earlier this fiscal year. The pitch, as I understand it from the demos I’ve sat through and from the prior-quarter commentary on the record, is that Lightspeed has retooled its sales motion away from the inbound, self-serve SMB funnel that defined the legacy Lightspeed Restaurant book and toward an outbound-rep motion targeting full-service hospitality operators in continental Europe.

What I want to mark before the print is the comparison surface. Lightspeed’s AI story is not yet as concretely shaped as Toast IQ’s or Square’s. There is no equivalent of the Toast IQ “25,000 restaurants / 235,000 uses” disclosure on the Lightspeed tape that I am aware of. There is no equivalent of the Square Voice Ordering launch. What there is is a sales motion — outbound, hospitality-specific, Europe-weighted — and a product roadmap that the company has been hinting will lean into AI-augmented capabilities for full-service.

The forthcoming Q2 FY26 print is the one that should, if the company chooses to disclose, tell us three things: whether the outbound motion is producing the location adds the hospitality push needs to be real, whether the gross profit per location is moving the way an upmarket motion would predict, and whether AI-feature attach has begun to show up in the disclosures the way it has on Toast’s tape. I’ll be reading the release alongside the Mews/PMS roll-up dynamics I’ll be covering in a forthcoming May piece, because the Lightspeed hospitality push runs straight into the question of whether the POS layer or the PMS layer is going to own the AI surface for full-service hotels and restaurants. That question does not have a settled answer in the public disclosures yet.

The cleanest move I can make today, before the print, is to mark Lightspeed as the third leg of the comp without pretending I have data I don’t. If the Q2 FY26 tape includes an AI-feature disclosure analogous to the Toast IQ number, the head-to-head becomes a three-way comp on real data. If it doesn’t, the Lightspeed story stays as a sales-motion story for one more quarter, and the head-to-head is, for the moment, a two-way comp between Toast IQ and Square Voice Ordering with Lightspeed as the bench.

The surface matrix

Here is the way I would line the three up on the desk today. The point of this matrix is not to crown a winner. It is to show why the three are not the same product.

Toast IQ is an in-store assistant pointed at the operator (back-of-house manager or owner-operator), bundled into existing SaaS tiers, monetized via the ads-network rev share, and pitched against the 156,000-location installed base. The buyer is already a Toast customer. The expansion motion is upsell.

Square Voice Ordering (plus Kiosk and Multichannel Menus) is a guest-facing capture stack pointed at QSR, sold as a labor-displacement ROI, monetized as a SaaS line on the Square POS subscription, and pitched at new-door installs where the F&B GPV +17% growth is being generated. The buyer is often not yet a Square customer. The expansion motion is acquisition.

Lightspeed’s hospitality push is an outbound-led motion pointed at European full-service hospitality, sold on a sales-rep basis, monetized as a Lightspeed Restaurant subscription with payments attach, and pitched against the regional hospitality specialists and the legacy on-prem POS base. The AI surface area is not yet fully priced on the public tape — the Q2 FY26 print tomorrow may or may not begin to price it. The buyer is rarely a Lightspeed customer at first touch. The expansion motion is also acquisition, but the target buyer profile is upmarket from Square’s QSR sweet spot.

Three buyers. Three motions. Three pricing surfaces. The reason I keep coming back to this is that the equity desks are mostly modeling these as one comp table, and the operator-side decision I keep seeing in the field — at the counter on Tuesday afternoon — is much more segmented. The QSR owner asking about Voice Ordering is not also asking about Toast IQ. The full-service hospitality operator looking at Lightspeed is not also looking at Square Kiosk. The independent restaurant on Toast is, in most cases, not looking outside the Toast stack at all.

What I think wins which segment in 2026

Here is the stake. In QSR and counter-service, where the labor seat being displaced is on the phone or at the counter, I think Square’s voice-plus-kiosk stack wins the new-door installs through 2026, conditional on the voice product clearing the accent and noise robustness threshold in real-kitchen conditions. The F&B GPV +17% line is the early signal that Block is taking F&B share at the install moment, and the AI surface area is the wedge.

In full-service casual and the independent fine-dining segment in the U.S., I think Toast keeps the installed-base lead and Toast IQ turns into an ads-network engine more than a standalone AI product line. The 25,000-restaurants-in-month-one number from the Q3 call is the early signal that operators are opening the assistant. The question that decides the equity case is whether IQ-recommended ads spend on Toast Advertising compounds in 2026 the way the locations line did over the last decade.

In European hospitality, I think Lightspeed has the open lane — but the size of the lane is the part the Q2 FY26 print tomorrow needs to start to price. If the outbound motion is producing the location adds and the gross-profit-per-location lift the upmarket pitch needs, then Lightspeed has a credible third-leg story. If it isn’t, the Lightspeed AI surface is the bench, and the comp table is two-handed.

What I’d ask the three CEOs

If I had a question in the queue for each of the three Q3 calls — and the Lightspeed Q2 FY26 call tomorrow — here are the three I’d ask.

For Toast: of the 25,000 restaurants that used IQ in October, how many used it again in week four versus week one? The funnel-to-retention conversion is the number that decides whether IQ is the engine or the launch-week curiosity bump.

For Square (Block): on AI Voice Ordering, what is the in-production escalation rate to a human order-taker across the launched cohort, and what is the menu-sync error rate the voice agent is encountering on Multichannel Menus operators? The labor-displacement ROI is a function of both numbers, and neither has been disclosed.

For Lightspeed (tomorrow): in the European hospitality outbound book, what is the AI-feature attach rate at install, and is it being priced as a standalone line item or bundled into the Restaurant subscription? The answer determines whether the AI surface is monetized or marketed.

None of those three questions, as of this morning, has a number on the public tape. The Toast number is the one closest to disclosure — Q4 is the obvious surface for management to either offer the retention cut or decline to. The Square numbers are the ones the analyst Q&A will start to push on every quarter from here. The Lightspeed numbers are the ones the print tomorrow either begins to surface or doesn’t.

Mark interpretation

The right way to read the three POS-AI strategies on November 5 is not as a single comp table. Toast IQ is an installed-base upsell story whose monetization runs through the Toast Advertising network. Square’s voice-and-kiosk stack is a new-door acquisition story whose monetization runs through labor-displacement ROI at QSR. Lightspeed’s hospitality push is an outbound sales-motion story whose AI surface area we will see priced — or not priced — when Q2 FY26 hits the tape tomorrow morning. Reading them as three different products with three different buyers is the only way the equity case makes sense; reading them as one product is how the sell-side gets the 2026 model wrong.

If I had to pick one of the three to overweight on the AI thesis alone, on the data we have on the desk this morning, I would overweight Square. The F&B GPV +17% line is the most legible early signal that the voice-plus-kiosk stack is taking share at the install moment, and the labor-displacement ROI is the most concrete pricing surface of the three. Toast IQ is the more interesting product story; the 25,000/235,000 disclosure is the most provocative AI adoption number on the public tape this quarter. But product interest and equity-case underwriting are different things, and the question I am putting under each of the three names is whether the AI surface area is priced. On Square, it is priced as labor saved. On Toast, it is priced as ads-network rev share. On Lightspeed, it is — as of this morning — not yet priced on the public tape.

The Lightspeed print lands tomorrow. The Toast Q4 print lands in February. The Square Q4 print follows. The desk review I’ll be writing into those three tapes is the same one I’m marking today: which of the three AI surface areas is monetized, which is marketed, and which is bench. The three are not the same product. They should not be modeled as the same comp.

— Sofia leads Vibe Check vendor reviews for TableTransfers. Tips: [email protected].