A market analysis, product and technology audit, AI-readiness review, PM critique, and mock product strategy for SevenRooms — the hospitality CRM and reservations platform that DoorDash bought to own the restaurant's front door.
SevenRooms is a New York–founded hospitality software company. It sells reservations, table management, a guest CRM, marketing automation, payments, review management and, since 2025, voice AI to restaurants, hotel food-and-beverage teams, nightclubs, membership clubs, and entertainment venues. Its customer list skews premium and multi-venue: Marriott, MGM Resorts, Mandarin Oriental, Wynn, Nobu, Wolfgang Puck, and Dishoom appear in company and DoorDash materials.
Since 13 June 2025 it has been a DoorDash company, folded into DoorDash's Commerce Platform next to online ordering and storefront tools. For most of its history the pitch was the mirror image of a marketplace: the restaurant owns the guest relationship and the data, pays a software fee, and does not pay a per-cover commission. The acquisition changed who is standing behind that promise.
Runs SevenRooms inside DoorDash. Fronts the operator narrative ("more tables for diners, one book for restaurants") and the Channel Connect launch.
Owns the guest-intelligence roadmap: Client Insights, Global Guest Search, and the AI feature set. Quoted on the July 2026 launch.
Named as a co-founder in DoorDash's announcement; listed as CTO in company history. His current role is not confirmed in 2026 coverage, and some data aggregators mislabel him as CEO.
The DoorDash executive who fronted the deal and the "Going Out" consumer launch. She is the clearest signal that SevenRooms is now a component of DoorDash's in-store strategy.
Early investors included BoxGroup, Thomas Keller, and the Nathan's Famous family. Later backers: Comcast Ventures, Amazon's Alexa Fund, Providence Strategic Growth, and Enlightened Hospitality Investments.
Several widely repeated numbers about SevenRooms do not survive a second look. Here is how I weighted them.
| Claim | Source | Problem | Confidence |
|---|---|---|---|
| $43M ARR (2024), ~1,000 customers, $43K ACV | Latka (estimate) | Circular: 1,000 × $43K = $43M. Cannot be reconciled with 13,000+ venues. Also mislabels the CEO. | LOW |
| 13,000+ venues (May 2025) vs 15,000+ restaurants (site, 2026) vs ~2,100 US restaurants (mid-2024) | DoorDash IR / sevenrooms.com / BistroChat | Three different units (venues, restaurants, US-only) and three dates. The gap suggests a footprint weighted to hotels, nightlife, and international. | MEDIUM |
| Deal price and close date | DoorDash IR, 10-K | Some secondary write-ups misdate the deal to 2024. Primary filings confirm May 2025 announcement and Jun 13, 2025 close. | HIGH |
| Voice AI: "25% more reservations captured" | SevenRooms / ElevenLabs | Adopters' Q1–Q3 2026 versus the same months of 2025. No control group. Vendor-reported, and 600+ restaurants is small next to 13,000+ venues. | MEDIUM |
| Reservations via DoorDash marketplaces "+150% QoQ" | DoorDash Q2 2026 release | No base number. Going Out launched in two cities in Sept 2025, so a large percentage from a small base is expected. | MEDIUM |
| Trustpilot 1.7 / 5 (80 reviews) vs G2 4.7 / 5 (76 reviews) | Trustpilot, G2 | Trustpilot is diner-heavy and about deposits and booking failures. G2 skews toward vendor-solicited operator reviews. Neither is a clean NPS. | MEDIUM |
| "~100+ sales reps laid off post-acquisition" | One anonymous Glassdoor review | Single source, uncorroborated in press. Recorded as a watch item, not a finding. | LOW |
| 75% of consumers comfortable with AI reservations | DoorDash / SevenRooms trends report | The same survey is cited as 74% and 75% in different places. It is also vendor-commissioned (3,001 consumers, 509 operators). | MEDIUM |
Restaurant reservations used to be a software category. It is turning into a fight over who owns the diner relationship, with card issuers, delivery apps, POS vendors, and AI assistants all entering at once. Software fees matter less each quarter. Guest identity and payment-linked perks matter more.
Only about 19% of US sit-down restaurants used an online reservation or waitlist system by mid-2024, up from 13% in 2022 (BistroChat, citing a LinkedIn market analysis). The category is still early in the US even though the incumbents are famous.
SevenRooms' own research: 40% of US operators run four or five separate systems, 83% think connecting them would improve profitability, and one in five cannot recognize the same guest across on- and off-premise. It is vendor research, but the pain is real.
American Express is merging Resy and Tock (about 25,000 venues, Tock brand retired, restaurant tech combined in 2027). Chase and Visa lean on OpenTable. DoorDash ties reservations to DashPass. Each has a payments or loyalty program behind it.
Single-source estimate (a mid-2024 LinkedIn analysis relayed by BistroChat), US only, before the Amex merger and before DoorDash's sales push. Platforms define "active restaurant" differently. Resy and Tock together would be about 19%. Use as a rough shape, not a ranking to quote.
| Company | Model | Moat | SevenRooms Overlap | Threat |
|---|---|---|---|---|
| OpenTable (Booking Holdings) | Marketplace + SaaS. Public plans of $149 / $299 / $499 a month plus $1.00–$1.50 per network cover and a 2% fee on deposits and no-show charges (vendor-blog figures) | Diner network, ~65,000 restaurants globally, Chase/Visa perks, ChatGPT integration | Core reservations and table management; direct head-to-head in premium dining | HIGH |
| Resy + Tock (American Express) | Flat membership fees; Amex card benefits; Toast POS partnership | Card-member demand, a $400 Platinum dining credit, six- and seven-figure signing incentives reported by restaurateurs | Fine dining and destination restaurants, the same customers SevenRooms courts | HIGH |
| Toast Tables / Square | POS-bundled reservations | Already in the restaurant's hands; "one bill, one vendor" | Independent and small-group restaurants below SevenRooms' price line | MEDIUM–HIGH |
| Yelp Reservations / Guest Manager | Discovery-linked waitlist and reservations | Diner traffic; Google and ChatGPT integrations | Casual and mid-market dining | MEDIUM |
| Google AI Mode / ChatGPT | Agentic booking front doors | Distribution monopoly; they choose which platforms are bookable | SevenRooms is a named Google partner in 8 international markets; not named in ChatGPT's Aug 2026 launch coverage | GATEKEEPER |
| TheFork / Eat App / regional | Regional marketplaces and SaaS | Local diner traffic (TheFork lists ~80,000 partner restaurants) | Partners in some markets and rivals in others | LOW–MEDIUM |
| Voice-AI point solutions | Answering-service AI built on the same voice infrastructure as Retell-class platforms | Fast to deploy; no need to replace the reservation system | SevenRooms Voice AI, which is built on ElevenLabs' agent platform | EMERGING |
OpenTable's April 2026 client terms ask restaurants to make it their primary table-management system and share inventory across the marketplace, along with a shift from month-to-month to annual contracts. Operators such as Altamarea Group publicly pushed back ("we want to have multiple partners"). SevenRooms answered with Channel Connect, a free aggregator that syncs across platforms. Counter-positioning is now a product line.
Google AI Mode reached 75 million daily users by January 2026 and now books tables across partners; ChatGPT added in-chat reservations on Aug 10, 2026. In both cases the platform loses the diner's direct relationship, because bookings flow through an assistant. Being a listed partner is now a distribution question, not a technical one.
The reservation slot is becoming a payment-linked benefit: Amex dining credits, Chase and Visa perks, DoorDash credits worth an average $9 per redemption. Restaurants report being paid to switch platforms. Guest data is the real prize, as one Boston operator put it: "It's less about the reservations … all about the data."
SevenRooms' research says 64% of diners still call to book and 40% of those calls go unanswered during service, yet only 28% of operators use AI for calls. Vendor numbers, but consistent with the pitch of every voice-AI startup pointed at restaurants.
In the DoorDash survey, 74% of diners say dining in led them to order delivery later, and 62% say delivery led them to dine in. That is the empirical case for one guest profile across channels, and the reason DoorDash wanted a CRM.
The moment a marketplace's consumer records appear in an operator's CRM, consent rules under GDPR (UK/EU via Deliveroo) and US state privacy laws matter. Coverage of Global Guest Search describes no consent framework, so this is an open item rather than a solved one.
Profiles that merge reservation history, POS spend, notes, and cross-channel identity. Recognition at the door is the visible payoff.
Being bookable wherever the diner starts: Google, agents, marketplaces, the restaurant's own site. Lock-in fights are being fought at the channel layer.
The product is used by a host stand at 7 p.m. with a line at the door. Sync errors, app freezes, and double bookings destroy trust faster than any feature gains it.
Zero cover fees versus $1.00–$1.50 per cover changes the math for restaurants that already have demand. It changes it the other way for restaurants that need the marketplace to fill seats.
Quote-based pricing (third-party estimate ~$499 a month per location, add-ons extra), no free trial, and repeated "expensive for smaller establishments" reviews. Toast and Yelp bundle their way in below this line.
Reviewers cite an Android/mobile gap ("impossible to use"), Oracle Micros POS friction, thin reporting filters, "bare bones" waitlist tools, and freezing during seating. The same reviewers praise the core CRM.
Trustpilot is dominated by diners reporting failed bookings and unexpected deposit charges (one reports $167). Restaurants set those deposit rules, but the SevenRooms name is on the checkout page.
Eater reports restaurateurs declining to comment on record over "fear of DoorDash retaliation." Whether or not that fear is justified, it is a perception a data-ownership brand cannot ignore.
SevenRooms organizes its platform into three hubs, Marketing, Guest Experience, and Operations, on top of a single guest profile. The product is easiest to understand as five layers: how bookings arrive, what the system knows about the guest, how the floor is run, how money and experiences are sold, and what AI now does across all of them.
Own-website booking widget, Google (Search, Maps, Assistant), Instagram and Facebook, Tripadvisor, TheFork, and regional marketplaces (Chope, OpenRice, Reserveout, The Chefz), plus DoorDash and Deliveroo. Channel Connect (Jun 2026) is a desktop app that syncs inventory across channels in real time and is free to any restaurant. Its stated proof point: Chelsea Living Room reports 90% less time on manual inventory management and 99% fewer double bookings.
Profiles merge reservation history, POS spend, staff notes, tags, and feedback. Email, SMS, and WhatsApp campaigns run off segments. New in July 2026: Global Guest Search, which lets staff search hundreds of millions of DoorDash and Deliveroo consumer records, and Client Insights, which summarizes a guest's history for pre-shift briefings.
Seating and pacing controls, cover forecasting, an auto-seating algorithm the company says evaluates 10,000+ table combinations per second, and newly announced Flexible Table Capacities and Slot Mode. Smart Boost flags busy and slow periods for targeted incentives.
Prepaid bookings through Stripe, Adyen, CyberSource, FreedomPay, and Network International. The redesigned first-party booking flow claims a 25% average lift in premium-experience sales. This layer is also where diner complaints concentrate (unexpected deposit charges on Trustpilot).
Diners book inside the DoorDash app and earn credits toward delivery orders; restaurants receive the booking without a per-cover fee and a 360-degree view that includes delivery and pickup history. In DoorDash's own data, 80% of customers using Going Out visited restaurants they had not tried before.
SevenRooms does not publish prices. Third-party reviews describe three quote-based tiers (Starter, Growth, Premium), roughly $499 a month for a single location, and separate add-ons for email marketing, Voice AI, and WhatsApp/text. There is no free trial. The commercial differentiator is no per-cover fee, which matters most for restaurants that already have demand. The counter-position is OpenTable's marketplace model, where the restaurant pays $1.00–$1.50 per network cover and gets discovery in return.
SevenRooms is secretive about its architecture, so this section separates confirmed facts from inference.
Job postings for backend roles ask specifically for Python and Django experience on a "hospitality-focused SaaS platform." The cloud provider is not named. A Django monolith with service extraction around the hot paths (availability search, booking widget) is the most probable shape, given a 2011 origin and the pace of feature addition. INFERRED
Credentials are scoped to a venue group. Endpoints cover reservations (create, modify, cancel), shift-level availability, guest profiles and feedback, and venue data. Developer documentation is restricted to approved operators and integration partners. The platform advertises 100+ integrations, including 30+ POS systems (Toast, Square, MICROS Simphony, Lightspeed, NCR Aloha, Revel, SpotOn), delivery and ordering (DoorDash Drive, Olo, Deliverect, ItsaCheckmate), and hotel systems (Revinate, Journey).
The guest profile is the core asset: entity resolution across email, phone, and name variants, POS check-level spend joined to reservations, and free-text staff notes. Global Guest Search implies a query path into DoorDash and Deliveroo's identity graphs. Nothing public explains how matching, consent, and data separation are enforced, which is the most important technical unknown in this report.
Voice AI runs on ElevenLabs' agent platform, natively connected to SevenRooms guest profiles, availability, and venue rules. AI Notes, AI Responses, Feedback Summaries, and Client Insights are LLM-based text features whose model provider is not disclosed. The seating engine's "10,000+ combinations per second" is a combinatorial search claim, which says nothing about whether a learned model is involved. Treat it as optimization software with an AI label until shown otherwise.
G2 and Capterra reviewers describe a host-stand app that can freeze during seating, a missing Android experience, and a mobile app called "impossible to use." Channel Connect being a desktop app suggests operators want a back-office control surface, not another tablet.
SevenRooms' AI strategy is unusually coherent for a company its age: put AI where staff already work (the reservation screen, the pre-shift brief, the phone), fed by the guest profile that only SevenRooms has. The company describes it as "agentic AI" that can "access your reservation platform, CRM and guest data" rather than operate in isolation. That is a credible design principle. The open questions are ownership, proof, and reach.
| AI Feature | What It Does | Type | Evidence | Read |
|---|---|---|---|---|
| Voice AI | Answers calls 24/7; books, modifies, and cancels; recognizes repeat guests; collects deposits; escalates to staff | Bought platform, native integration | 400K+ calls, 600+ restaurants, up to 25% more reservations captured, 1 in 5 calls after hours | REAL |
| AI Table Management | Seating optimization using party size, pacing, and demand; feeds revenue management | Optimization algorithm | "10,000+ combinations per second" (a speed claim, not an outcome claim) | REAL, LABEL STRETCHED |
| Client Insights | Summarizes preferences, spend, and staff notes for hosts and servers | LLM summarization over the CRM | Maître d' testimonial; no metric | REAL |
| AI Notes | Cleans and standardizes hurried guest notes into CRM fields | LLM extraction with human review | None published; a data-quality play | REAL, QUIET |
| AI Responses | Drafts replies to emails, SMS, and review-site comments | LLM drafting | 898% more reviews responded to; 85% faster response time (customer-selected) | REAL, WEAK PROOF |
| Feedback Summaries | Weekly digest of reviews and survey responses | LLM summarization | None published | COMMODITY |
| Global Guest Search | Searches DoorDash and Deliveroo consumer records to recognize new guests | Cross-platform identity lookup | Announced Jul 21, 2026; no consent framework described | HIGH VALUE, HIGH RISK |
Scores are my judgment from public evidence, not a benchmark. Composite: roughly 5.5 / 10. Strong on the ingredients that are hard to copy (data, workflow), weak on the ones a rival can copy in a quarter (open interfaces, published methodology, consent design).
| Segment | Proof Points | Problem Solved | Fit Today |
|---|---|---|---|
| Multi-venue restaurant groups & fine dining | Nobu, Wolfgang Puck, Union Square Hospitality, Dishoom, Ethan Stowell (14 locations), Brotzeit (12) | Recognize VIPs across venues, own the email and SMS list, avoid per-cover commissions | STRONG |
| Hotels & resorts (F&B) | Marriott, MGM Resorts, Mandarin Oriental, Wynn, Jumeirah; 14 Harrods outlets (2021) | Merge F&B guest data with hotel guest profiles; concierge and in-room booking | STRONG |
| Nightlife, clubs & entertainment | Live Nation, Topgolf, NoHo Hospitality, Madera Group | Table minimums, guest lists, events, prepaid experiences, VIP tracking | STRONG |
| International hospitality | 75+ countries cited; channel ties to TheFork, Chope, OpenRice, Reserveout, The Chefz | Local marketplace access plus one guest database across markets | MEDIUM–STRONG |
| Breweries, wineries & ticketed experiences | Listed as a segment on sevenrooms.com; Tock has 1,200 wineries | Ticketed tastings and add-ons | MEDIUM |
| Independent single-unit restaurants | Reviewers describe it as "expensive and could be more flexible for smaller establishments" | Little pain that Toast, Yelp, or OpenTable Basic do not already cover at a lower price | WEAK |
| Customer | Claimed Result | How to Read It |
|---|---|---|
| Nobu London | $4M commission savings over 7.5 years; $203K revenue in two years; 349K emails added to CRM | Savings assume guests would have booked through a paid marketplace anyway |
| IGC Hospitality | A$1.6M savings over 2.5 years; A$500K first-year revenue | Same counterfactual problem; revenue attribution method not shown |
| The Apollo | 10% cover increase; $6K a month in commission savings | Single-venue result; time window unstated |
| Brotzeit (12 locations) | $2M revenue and 350K+ guest relationships created | "Relationships" are contact records, not repeat visitors |
| Chelsea Living Room | 90% less manual inventory work; 99% fewer double bookings (via Channel Connect) | Strong operational claim, single customer, launch-day quote |
All figures are from sevenrooms.com and company press releases. Ethan Stowell figures on the site mix currencies and are omitted. "Commission savings" compares against a marketplace fee the venue may never have paid, which flatters the number.
An interface floor staff can learn quickly, guest data "front and center," commission-free reservations, reporting some call "second to none," and customer service Capterra reviewers rate 4.3 / 5. G2: 4.7 / 5 across 76 reviews. Capterra: 4.4 / 5 across 24.
Price for smaller venues, a mobile and Android gap, Oracle Micros and other POS integration snags, "bare bones" waitlist and guest-list tools, missing features versus OpenTable, and a steep learning curve. One Capterra reviewer cites "poor security governance."
1.7 / 5 across 80 reviews, 75% one-star. Themes: booking flows that time out or fail at the last step, unexpected deposit charges, and, from a few restaurant owners, contract terms they felt trapped by. Many of these are venue policy problems that diners attribute to the platform.
3.9 / 5, 66% would recommend. Praise for accessible leaders and remote flexibility. One detailed negative review alleges post-acquisition instability. Uncorroborated; logged as a watch item.
SevenRooms has the best-fitted product for premium, multi-venue hospitality and a parent that can give it a consumer front door. The critique is not about whether the product is good. It is about three tensions the acquisition created and the product has not yet resolved: ownership (whose data is it?), reach (who can book through it?), and proof (who has verified the numbers?).
Vision: Every restaurant, hotel, and venue on SevenRooms knows its guests before they arrive, on any channel they arrive from, with data the venue can prove it owns. SevenRooms becomes the booking and guest platform that every marketplace, assistant, and agent has to work with, not the one they try to replace.
North Star Metric: Recognized Cover Rate, the share of seated covers where a guest profile with at least one prior interaction on any channel was in front of the host before seating. It rewards the outcome that only SevenRooms can deliver (recognition), rewards data breadth (delivery, dine-in, phone, agent), and cannot be gamed by signing more venues. Baseline: to be measured in Q4 2026. Target: 60% for venues on the full stack by end of 2027.
Guardrail Metric: Consent-Logged Enrichment Rate. Every profile enrichment from outside the venue's own data (including DoorDash and Deliveroo) has a recorded consent basis. Target 100%. If the north star rises while the guardrail falls, the strategy has failed.
The Strategic Pivot: From "the best CRM for premium restaurants, owned by DoorDash" to "the open, verifiably operator-first guest platform that DoorDash's scale makes stronger." The first half of that sentence is what SevenRooms is. The second half is what has to be earned.
Turn the biggest risk into the most defensible feature. Ship a region-aware consent and purpose model for Global Guest Search, a data-separation guarantee for operator-owned records, full data export, and a public Trust Center with third-party audit. Recognition improves because guests opt in, not because they were matched silently.
A self-serve REST API, an official MCP server, and venue-controlled agent policies, so any assistant can find and book a SevenRooms venue on the venue's terms.
Public documentation, sandbox, and an MCP server for availability search and short holds. Source tagging so every agent-originated booking is attributable. Free for venues.
Booking, modification, and cancellation with venue rules: party-size caps, deposit requirements, blackout windows, per-agent rate limits, and no-show protection for agent bookings.
Named listings in the major assistants in every market SevenRooms serves, starting with the US gap. Channel Connect becomes the venue's single control panel for all agent and marketplace inventory.
An agent that finds a venue on SevenRooms sees a live guest-aware experience (preferences, recognized returning guest, upsell rules) that a marketplace listing cannot offer. Openness feeds the guest graph rather than diluting it.
A lighter tier delivered through DoorDash's merchant surfaces: publicly priced, zero cover fees, guided setup, pre-built connectors for the largest small-venue POS systems, and Voice AI on usage-based pricing. Priced against OpenTable Basic ($149 a month) as the visible anchor. Feature fences protect the premium tier: multi-venue CRM, revenue management, hotel integrations, and advanced segmentation stay above the line.
Not a bet, a prerequisite: diner-side checkout completion SLO, itemized deposit receipts, a native tablet and Android experience, certified connectors for MICROS Simphony and other hotel-grade POS systems, and reporting filters operators have asked for. Reviewers say the CRM is excellent. They also say the app freezes.
| Risk | Severity | Likelihood | Mitigation |
|---|---|---|---|
| Data-trust incident or regulator action on cross-platform guest data | HIGH | MEDIUM | Consent-first design, region gating (UK/EU first), data-protection impact assessments, third-party audit, and a public incident process. |
| Major assistants keep SevenRooms out of their listings | HIGH | MEDIUM | Open interface with no-cost venue participation; route through Channel Connect's installed base; use DoorDash's platform relationships without making listing contingent on them. |
| Amex and OpenTable incentives pull premium venues away | HIGH | HIGH | Compete on data ownership and cross-channel value, not on signing bonuses. Sell the guest lifetime value report to the top 200 groups directly. |
| Essentials cannibalizes premium plans | MEDIUM | MEDIUM | Hard feature fences (multi-venue CRM, revenue management, hotel integrations), separate packaging, and premium-churn monitoring as an OKR. |
| Voice AI vendor dependency: cost, quality, or terms change | MEDIUM | MEDIUM | Abstraction layer over the voice provider, cost-per-handled-call tracking, and a second-vendor evaluation each year. |
| Post-acquisition talent or sales-organization disruption (unverified) | MEDIUM | LOW–MEDIUM | Retention tied to Charter and Agent Interface milestones; protect customer-facing continuity for top accounts. |
The claim is the brand. Until the charter, audit, and export are live, marketing it invites the exact scrutiny a parent-company narrative attracts. Prove it first, then say it louder.
Some restaurants already offer DoorDash-exclusive reservations by choice. Making it a condition of better ranking or pricing would confirm the retaliation story and end the neutrality pitch.
Two more summaries and a redesigned workflow are worth less than one fewer freeze during a Friday-night seating. Put reliability on the roadmap as a first-class line, not a background task.
DoorDash owns the diner front door and the assistants own the rest. SevenRooms' advantage is behind the host stand. Consumer surfaces compete with the parent and the partners that Channel Connect needs.
SevenRooms is a good product attached to a large idea. The product is the best-fit guest CRM and reservation system for premium, multi-venue hospitality, with reviews to match, a growing set of well-placed AI features, and one shrewd piece of counter-positioning in Channel Connect. The idea is that a restaurant that recognizes its guests, on any channel, wins, and that DoorDash's consumer graph can make recognition automatic.
Three facts qualify the optimism. Its estimated US share is around 3% and the price paid implied roughly 15–28x revenue, so growth has to come from somewhere new. The AI front door is being built by Google, OpenAI, Amex, and OpenTable, and SevenRooms' interfaces are still permissioned. And the sentence that built the brand, "you own your data," is now printed on the product of a marketplace.
None of these is fatal, and all three have product answers: a consent-first data charter, an open agent interface, a down-market tier, and a published method for the numbers. The window is roughly 18 months. It closes as the Amex merger completes in 2027, as assistants finish choosing their partners, and as OpenTable's annual contracts lock in the venues that accept its new terms.
Sources: sevenrooms.com (product, AI, integrations, press releases of Jun 16 and Jul 21, 2026); DoorDash IR (deal announcement, Q2 2026 results, Going Out launch, 2026 Restaurant Industry Trends report); DoorDash 2025 annual report amendment (acquisition fair value); ElevenLabs (Voice AI case study, Jul 2026); Restaurant Dive; Restaurant Business; Boston Globe (Aug 2026); Fast Company (Amex, Resy, Tock); Eater NY; Expedite; Food On Demand; BistroChat (mid-2024 market share, single source); Eat App and other vendor blogs (OpenTable pricing); Winbuzzer and explainx.ai (Google AI Mode and ChatGPT reservations); G2, Capterra, Trustpilot, Glassdoor; Wikipedia; Latka and Tracxn (estimates; low confidence). Analysis as of September 2026. Revenue, share, and outcome figures are estimates or company-stated unless noted.