Mirakl

1.1 Identity, Product Scope, and Current Scale

Mirakl is an enterprise commerce infrastructure company built around a simple thesis: large retailers, distributors, and brands increasingly need marketplace-style assortment expansion without owning all of the inventory themselves. The company, founded in 2012 by Philippe Corrot and Adrien Nussenbaum after their SplitGames experience, now presents itself as the “operating system for intelligent commerce.” That framing is broader than the original marketplace thesis: Mirakl still anchors on third-party marketplace and dropship operations, but it now layers catalog onboarding, seller payouts, retail media, multichannel seller tooling, and agentic-commerce infrastructure on top of the core platform. The company’s dual Boston/Paris headquarters and global leadership bench reinforce that Mirakl is run as a cross-Atlantic scale-up, not a France-only SaaS vendor. Scale claims are material and mostly company-sourced. Mirakl says it serves 450+ enterprise customers and a network of more than 100,000 brands and sellers; those numbers appear consistently across the 2025 results release, the about page, and adjacent product pages. Public customer proof also suggests the platform is deeply embedded in the operating models of major retailers: Macy’s used Mirakl to add 220,000+ marketplace SKUs in year one, while Best Buy, Lowe’s, and Ulta all launched or scaled marketplace programs on Mirakl in 2025. The strongest takeaway for later chapters is that Mirakl is no longer just a marketplace-enablement vendor; it is trying to become a broader commerce control plane for enterprise operators and sellers.[CO001, CO002, CO003, CO004, CO005, CO006]

1.2 Leadership, Governance, and Key-Person Dependence

The about page shows a fairly mature leadership bench: Marie Best (CFO), Laure Le Gall (CRO), Nagi Letaifa (CTO), Jean-Yves Simon (Chief Product Officer), Sophie Marchessou (Chief Customer Officer), Scott Eckert (CEO Americas), and Tzipi Avioz (CEO APAC & Japan) all appear as current leaders. That matters because Mirakl’s current strategy spans software, financial flows, media monetization, and AI infrastructure; it would be hard to execute that transition without a more specialized executive layer beneath the founders. The board-and-advisor surface visible on the site also includes investment-linked figures from Silver Lake, 83North, Permira, Felix, Bain, and Elaia, which is consistent with a late-stage private company whose governance has broadened with each financing round. Even so, founder dependence remains high. Corrot and Nussenbaum are still the public strategic narrators across funding, annual results, and new-product launches, and the company’s identity is closely linked to their long-held platform-economy thesis. There is no public disclosure of detailed board rights, voting control, or succession planning. That is normal for a private SaaS company, but it leaves an underwriting gap around how much practical control sits with the founders versus later-stage investors. For diligence purposes, the leadership bench looks credible; the governance disclosure still does not match what a public-market investor would expect.[CO001, CO007, CO008, CO009]

1.3 Capital Formation, Valuation, and Financial Milestones

Mirakl’s funding history maps cleanly onto the rise of the marketplace-software category. Tracxn and Clay both show a ladder from a small 2012 Series A through a $20 million Series B in 2015, a $70 million Series C in 2019, a $300 million Series D in 2020, a $555 million Series E in 2021, and a €100 million revolving credit facility in 2023. The Series E remains the canonical valuation mark: more than $3.5 billion, led by Silver Lake with support from existing investors including 83North, Elaia, Felix Capital, and Permira. On disclosed equity alone, Mirakl has raised about $948 million; including the 2023 debt, total disclosed capital is roughly $1.06–1.1 billion depending on currency translation and data vendor rounding. The financial operating story improved materially after the 2021 financing peak. Mirakl reported $177 million ARR and $11.2 billion GMV in 2024, alongside positive EBITDA for the historical core platform. It then reported $218 million ARR, $14.6 billion GMV, and full group-level profitability in 2025. Those are strong absolute numbers for a private verticalized enterprise SaaS platform. They also imply that Mirakl is growing faster than many broader commerce peers while carrying a much older valuation mark. That combination — improving fundamentals but stale price discovery — is why valuation discipline becomes a central issue later in the report.[CO010, CO011, CO012, CO013, CO014, CO019]

1.4 Milestones, Customer Proof, and Important Caveats

Mirakl’s milestone cadence shows a company that kept broadening the product surface after finding fit in marketplace operations. The about page anchors the chronology: first dropship platform in 2013, first B2B marketplace in 2014, U.S. office in 2015, Mirakl Connect launch in 2019, Payout and the Octobat acquisition in 2022, Ads in 2023, Adspert in 2024, and Nexus in 2025. The 2024 and 2025 result releases deepen that picture by quantifying newer engines: Connect reached $11.7 million ARR in under a year; Ads reached $12.7 million ad spend in 2025 after >100% growth in 2024; and a May 2026 trust-and-safety launch connected the roadmap directly to regulatory compliance. The customer proof is strong because it spans multiple retail formats and measurable outcomes, not just logo slides. The main caveat is scope creep and platform complexity. Mirakl’s bullish story is that adjacent modules deepen moat and raise wallet share. The skeptical story, represented most clearly by competitor-oriented commentary, is that Mirakl remains a specialist marketplace layer that still requires operators to integrate storefronts, CMS, ERP, and order-management systems elsewhere. That criticism does not negate product-market fit, but it does matter for diligence: Mirakl’s differentiation depends on remaining the best neutral orchestrator for complex marketplaces rather than becoming just another costly integration layer in a composable stack.[CO015, CO016, CO017, CO018, CO026, CO027]

1.5 Exhibits

2.1 Market boundary and what Mirakl is really selling

Mirakl is often described against enormous e-commerce or B2B transaction markets, but those totals overstate the revenue pool that an infrastructure vendor can monetize directly. The company sells software and workflow orchestration to marketplace operators; it does not own the GMV that moves through those operators. The most relevant lens is therefore the enterprise marketplace-platform software layer: seller onboarding, assortment expansion, catalog normalization, payout orchestration, operations tooling, and increasingly retail media and AI distribution. Broader e-commerce platform studies are useful for context because enterprise buyers do compare Mirakl with VTEX, Adobe, Shopify, Salesforce, and composable stacks, but Mirakl’s addressable pool is a subset of that larger platform market. Sizing evidence confirms both the opportunity and the ambiguity. The broad e-commerce platform market is a single-digit-to-teens billions software category today, while B2B e-commerce transaction estimates are in the tens of trillions. Between those poles sits the more directly relevant B2B marketplace-platform software segment — large enough to matter, but far smaller than the GMV headlines. The right underwriting frame is therefore to ask whether Mirakl can capture a durable share of enterprise marketplace-software budgets as platform operators adopt third-party assortment, not whether it can somehow monetize a fixed percentage of all marketplace transaction value.[CM001, CM002, CM003, CM004, CM005, CM006]

2.2 Why the category is expanding

The demand case rests on channel economics and buyer behavior. Independent data points from Swell, Accio, and Mirakl-adjacent research all point in the same direction: marketplaces are taking a growing share of digital commerce, B2B buyers are behaving more like consumer buyers, and operators need broader assortments without carrying all of the inventory themselves. If only a small fraction of enterprises currently run marketplaces, the runway is still long. That is especially relevant for Mirakl because its best customers are incumbents trying to defend search visibility, price breadth, and long-tail assortment against Amazon, Alibaba, and category specialists. The strongest structural drivers are digital self-service, mobile usage, cross-border shopping, and AI-enabled merchandising. B2B buyers increasingly want instant product discovery, negotiated pricing online, and rep-free transactions. Operators, meanwhile, want capital-light growth through third-party sellers and new monetization levers such as retail media. Mirakl’s adjacent modules map directly onto those demand drivers: Connect helps sellers distribute across channels, Catalog Platform fixes supplier-data bottlenecks, Ads monetizes traffic, and Nexus/agentic tooling tries to keep merchants visible in AI-mediated discovery. The market is not merely growing in volume; it is broadening in functional scope.[CM008, CM009, CM010, CM011, CM012, CM013]

2.3 Who buys, who uses, and what they require

The enterprise buyer for marketplace software is rarely looking for a generic web-store builder. The classic Mirakl customer is a retailer, distributor, or manufacturer that already has a storefront and needs to add third-party assortment, dropship, or B2B marketplace functionality without rebuilding the whole stack. That explains why official competitor messaging from VTEX, Adobe, Shopify, BigCommerce, and commercetools all converges on similar evaluation criteria: omnichannel operations, broad integration capacity, localization, B2B capabilities, performance, and AI extensibility. The market is increasingly defined by whether a vendor can sit cleanly inside a larger commerce architecture. That purchase logic also shapes user personas. Marketplace operations, merchandising, supplier-onboarding, payments, finance, seller-success, and retail-media teams all touch the system, which makes implementation a cross-functional rather than purely IT decision. Budget owners usually sit in digital commerce or transformation functions, but the platform only sticks if catalog, checkout, seller compliance, and marketing workflows all improve. Mirakl’s strongest fit is therefore in complex organizations where assortment breadth, seller governance, and multi-team workflows matter more than building a basic storefront quickly.[CM019, CM022, CM023, CM024, CM025, CM026]

2.4 Constraints, substitutes, and what could slow adoption

The same complexity that creates demand also caps adoption speed. Market studies repeatedly flag cybersecurity, fraud, inventory accuracy, logistics coordination, legacy integration, and customer-acquisition cost as major constraints. For B2B deployments, contract pricing, approval flows, and negotiated terms add another layer of implementation friction. That matters for Mirakl because its deal sizes and deployment motions are enterprise-grade; long sales cycles and integration work are part of the product, not exceptions. Buyers can always decide to postpone a marketplace launch, extend first-party assortment instead, or assemble a composable stack without a dedicated marketplace specialist. A second constraint is value capture. Even if marketplace penetration keeps rising, the budget can fragment across core commerce suites, feed/discovery vendors, retail-media specialists, payments providers, and internal engineering teams. Competitor-oriented commentary makes this explicit by criticizing Mirakl for being a specialist layer that still requires adjacent systems. That critique is self-interested, but not wrong. The key diligence question is whether Mirakl’s category expertise is valuable enough to justify a separate control point in the stack as AI, retail media, and multichannel distribution each spawn their own software spend. The answer likely varies by customer complexity and by vertical.[CM020, CM021, CM027, CM028, CM033, CM034]

2.5 Exhibits

3.1 Competitive landscape and category structure

Mirakl does not compete in a single clean lane. The company’s original market was enterprise marketplace enablement, but the effective competitive set now spans at least four groups: integrated commerce suites (VTEX, Shopify, Adobe, Salesforce), composable and headless challengers (commercetools, Spryker), marketplace specialists (Marketplacer), and large commerce-network or point-solution vendors (Rithum, Feedonomics, Topsort). That breadth matters because buyers are often choosing architecture, not just features. A retailer deciding between Mirakl and Shopify or Adobe is partly deciding whether it wants a specialist marketplace control point or a broader operating system with marketplace capability layered in. The clearest way to frame Mirakl is as a marketplace operator specialist that has been broadening into adjacent budgets. It still wins by helping enterprises govern third-party sellers, enforce catalog and performance rules, and run marketplace workflows at scale. But its product expansion into Ads, Connect, Payout, Catalog Platform, and Nexus means the company now bumps into more vendors in more buying centers. The result is a market where direct competitors differ less by industry label than by how much of the commerce stack they try to own.[CP001, CP002, CP003, CP017, CP018, CP020]

3.2 Integrated suites and composable challengers

The suite competitors sell simplification. VTEX markets a unified stack across commerce, marketplace, OMS, and retail media for thousands of stores and customers globally. Shopify sells global rollout, omnichannel, wholesale, and TCO efficiency, while Adobe and Salesforce emphasize integration with broader commerce, data, and customer-relationship systems. In practice, these platforms compete by reducing the number of vendors a large operator must coordinate. That is especially attractive to teams that care more about fast unification than about best-of-breed marketplace depth. The composable challengers sell flexibility. commercetools and Spryker both frame the future as modular, headless, and API-first. Their pitch is that operators should not accept a rigid marketplace model if they can assemble the exact stack they want. Those vendors become more threatening when marketplace workflows look similar across customers and when AI, retail media, and data orchestration can be plugged in from outside. They become less threatening when seller-governance, compliance, and marketplace-specific operating detail become the real bottlenecks.[CP004, CP005, CP006, CP007, CP008, CP009]

3.3 Specialists, networks, and adjacent pressures

Mirakl also faces pressure from vendors that do not look identical on paper. Marketplacer competes as a marketplace overlay for operators that want range expansion without a full replatform. Rithum competes at network scale across listings, inventory, dropship, and retail media, especially for brands and retailers already selling across many channels. Feedonomics and Topsort compete at even narrower layers — distribution and monetization — but those layers matter because they are exactly where Mirakl is trying to expand wallet share. This is why Mirakl’s newer modules matter strategically. If Ads, Connect, and agentic-commerce tooling are merely helpful add-ons, adjacent specialists can cherry-pick those budgets while leaving Mirakl boxed into a slower-growth core. If those modules are deeply integrated into operator workflows, they increase switching costs and make Mirakl harder to displace. That tradeoff is one of the central questions for the rest of the report.[CP012, CP013, CP014, CP015, CP018, CP027]

3.4 Pricing, switching cost, and who Mirakl fits best

The strongest explicit skeptical case against Mirakl comes from competitor-oriented comparisons aimed at mid-market buyers. Those sources consistently argue that Mirakl is expensive, slow to implement, and over-specified for operators below true enterprise scale. Even if the exact quote ranges are not independently audited, the directional message is credible: Mirakl is built for organizations that can support long procurement cycles, deep integrations, and multi-team change management. That is a very different buyer than a founder-led marketplace or a mid-market distributor trying to launch quickly. The deeper issue is switching cost. Specialist marketplace platforms become embedded in catalog structures, seller-onboarding flows, and reporting hierarchies. That lock-in can be a feature when the operator wants stability and best-practice workflows; it becomes a liability when adjacent capabilities commoditize and the customer wishes the same budget had gone toward a broader suite. Mirakl’s best win zone remains complex operators that value marketplace-specific control more than one-vendor simplicity. Its hardest zone is customers who mainly want “marketplace enough” functionality inside a broader commerce platform.[CP021, CP022, CP023, CP024, CP025, CP026]

3.5 Exhibits

4.1 Revenue model and monetization surface

Mirakl no longer looks like a single-product marketplace SaaS vendor. The company still anchors around the core enterprise marketplace platform, but its disclosed product set now includes Ads, Connect, Payout, and Catalog Platform alongside newer AI-led initiatives. Financially, that matters because it should broaden monetization from a single software contract into a bundle of adjacent budgets: marketplace operations, supplier-data onboarding, cross-channel seller enablement, retail media, and marketplace payments. Public evidence is good enough to establish that breadth, but not good enough to show how much revenue each module contributes. The central public limitation is mix disclosure. Mirakl reports ARR and GMV, but not what share of ARR comes from core subscription software, transaction-linked monetization, services, or newer modules. That means a high-level investor can see that the revenue base is large and growing, but cannot yet tell whether the most recent growth is coming from durable, high-margin recurring revenue or from a mix with heavier services and implementation content than the headline implies.[CI001, CI002, CI003, CI004, CI005, CI006]

4.2 Public traction and quality of growth

Mirakl’s public traction is unusually strong for a late-stage private company. The company disclosed $177 million of ARR, $11.2 billion of GMV, and EBITDA profitability for 2024, then $218 million of ARR, $14.6 billion of GMV, and group-wide profitability for 2025. Digital Commerce 360 and Sacra both corroborated the core direction of those numbers. That does not eliminate reporting risk, but it does make Mirakl materially easier to underwrite than private peers that publish only vague customer counts or unattributed “triple-digit growth” statements. The deeper question is quality of growth. ARR and GMV moved up together while profitability improved, which is a positive sign. But the pace of growth also shows Mirakl has matured out of the 2021 fundraising environment. That is not a negative by itself; it simply means the investment case now depends more on durable attach, margin, and retention quality than on headline category creation alone.[CI009, CI010, CI011, CI012, CI013, CI014]

4.3 Unit economics, cost structure, and peer benchmarks

Mirakl’s unit economics are still mostly private. Public materials do not disclose gross margin, services margin, CAC, payback, NRR, GRR, or customer concentration. That means the investment team cannot answer basic underwriting questions such as how much gross profit incremental GMV or module attach creates, whether new logos pay back within a reasonable period, or how dependent the business is on large expansion deals. The best public answer is to use public-commerce peers as directional benchmarks, not as substitutes for company data. Those peers show that enterprise commerce can scale into large revenue bases, but that the model remains operationally demanding. Sales, partner ecosystems, product R&D, and global support all matter. For Mirakl, the implication is that reported profitability is meaningful, but still incomplete without mix and margin context. If Ads, Connect, and Catalog Platform are high-attach, software-like expansions, the earnings path is more attractive; if they carry heavy implementation or support costs, the quality of revenue is less compelling than ARR alone suggests.[CI018, CI019, CI021, CI022, CI023, CI029]

4.4 Capital adequacy and financial verdict

Mirakl’s capital story has improved. The 2021 Series E injected $555 million, and the 2023 revolving credit facility added another €100 million of optional liquidity. By 2025 the company said it had reached group-wide profitability. Taken together, those facts imply Mirakl is less dependent on immediate primary fundraising than many private software companies of similar scale. That is valuable both strategically and in a high-rate environment where private capital is more selective. But the capital picture is still not fully underwritten. There is no public cash balance, no draw detail on the revolving line, no free-cash-flow disclosure, and no visibility into liquidation preferences from prior rounds. For a growth-stage investor, that means the current business looks financially credible, but not yet sufficiently transparent to justify paying any price. The right interpretation is “lower financing risk than average, still meaningful information risk.”[CI025, CI026, CI027, CI028, CI034, CI036]

4.5 Exhibits

5.1 Product scope and customer workflow

Mirakl’s public product narrative has clearly widened. The company still anchors on enterprise marketplace operations, but it now presents a broader operating system spanning multichannel seller distribution, retail media, catalog onboarding, payouts, trust and safety, and agentic commerce. That breadth matters because Mirakl is no longer just a point solution for third-party seller management; it is trying to own more of the surrounding workflow that determines whether enterprise operators can launch, govern, and monetize large marketplace ecosystems. The strongest workflow proof remains in the mature modules. Customer stories for Macy’s, Best Buy Canada, Graybar, Lowe’s, Ulta, and Best Buy all indicate that Mirakl is used in production to expand assortment, onboard suppliers, or stand up marketplace motions. The newer modules like Trust & Safety and Nexus look strategically important, but they are clearly earlier on the maturity curve than the core marketplace stack.[CE001, CE002, CE003, CE004, CE005, CE006]

5.2 Architecture and developer surface

Mirakl’s public technical story is unusually concrete for a private SaaS platform. The company says it is API-first, event-driven, and built from 100+ stateless microservices. It claims large daily operating volumes across SKUs, API calls, inventory updates, and Black Friday order peaks. The technology page also advertises interoperable components, native connectors, and multi-cloud redundancy across AWS, Google Cloud, and Azure. That collection of claims supports the idea that Mirakl is engineered for large enterprise operator workloads rather than only mid-market storefront add-ons. The developer surface is real but not especially open. Mirakl has a public developer portal, API references, webhooks, and at least one official PHP SDK. Yet the public GitHub footprint is relatively small for a company of this scale, and much of the deeper documentation appears oriented toward registered customers or partners. That does not mean the platform is technically weak; it means Mirakl behaves more like an enterprise software vendor with controlled integrations than like a broad ecosystem platform with a giant outside developer community.[CE007, CE008, CE009, CE010, CE011, CE012]

5.3 Security, compliance, and reliability controls

Mirakl’s security and compliance posture is one of the strongest publicly verifiable parts of the product case. The technology page now claims SOC 1/2 Type 2, ISO/IEC 27001, ISO/IEC 27018, and ISO 22301 certification coverage, alongside multi-cloud redundancy, MFA, SSO, role-based access, and active bug bounty programs. Older Mirakl security posts corroborate the sequence of some of those milestones. The Cloud Security Alliance STAR listing adds a further outside signal that Mirakl is willing to map controls into recognized cloud-security frameworks. The limitation is depth of proof, not absence of proof. Public pages show claims and milestone announcements, but not the underlying audit artifacts, current incident history, or detailed control mappings that a security diligence workstream would actually want. That means Mirakl clears an important first bar for enterprise readiness, while still leaving meaningful diligence for any buyer or investor that needs to evaluate operational resilience in depth.[CE013, CE014, CE015, CE016, CE017, CE018]

5.4 Roadmap, dependencies, and technical verdict

The roadmap story is attractive but unevenly mature. Trust & Safety and Nexus both address real marketplace problems: moderating ever-larger catalogs under tightening regulation, and preparing catalogs and checkout flows for AI-agent-driven commerce. The J.P. Morgan partnership is particularly useful because it shows Mirakl trying to solve not just discovery but the secure transactional layer required for agentic commerce. That is an ambitious extension of the product from operator tooling into future commerce infrastructure. At the same time, these newer products increase dependency complexity. Mirakl relies on cloud providers, enterprise connectors, outside payment infrastructure, and customer readiness to adopt new workflows. The core platform looks strong and battle-tested. The question for diligence is how much of the newer product narrative is already monetizable and repeatable, versus how much is roadmap-shaped positioning that still depends on partner execution and customer experimentation. materially.[CE020, CE022, CE023, CE024, CE030, CE033]

5.5 Exhibits

6.1 Customer segments and buyer / user map

Mirakl sells into enterprise operators, not to end-consumers. The economic buyer is usually a marketplace, ecommerce, merchandising, digital, or procurement leader inside a retailer, distributor, or other enterprise that wants to expand assortment or supplier access without holding more inventory. The user map then broadens: marketplace operations teams, seller-onboarding teams, catalog teams, finance teams, and increasingly supplier- and seller-side users through Connect or Catalog Platform. This makes Mirakl less like a single-seat SaaS tool and more like shared operating infrastructure across several internal teams. The public customer base is also broad enough to matter. Mirakl’s official materials point to 450+ marketplaces and 100,000+ sellers, while outside datasets such as Landbase and Apps Run The World point to hundreds of companies and many large enterprise names. These two views are not perfectly comparable, but together they support the conclusion that Mirakl has real enterprise adoption across retail and B2B distribution.[CU001, CU002, CU003, CU004, CU005, CU006]

6.2 Named production proof and adoption trajectory

The strongest part of the customer story is named production proof. Macy’s, Lowe’s, Best Buy, Best Buy Canada, Ulta, and Graybar all offer current or recent evidence that Mirakl is live inside material commercial workflows. The proof is not just logos: Macy’s disclosed seller, brand, and customer-engagement metrics; Best Buy Canada disclosed SKU and growth metrics; Graybar disclosed workflow-speed gains; Lowe’s described returns, loyalty integration, and verified-seller controls. This is much stronger than a typical late-stage startup logo slide. The trajectory evidence also suggests that Mirakl’s best customers use the platform as a growth engine, not a side experiment. Holiday 2025 GMV and API-call figures point to heavy production usage at peak times. That does not automatically prove retention quality, but it does show that Mirakl’s core operator customers are moving meaningful traffic and catalog volume through the platform.[CU008, CU009, CU010, CU011, CU012, CU013]

6.3 Review signals, retention, and expansion logic

Independent review surfaces are directionally positive but not decisive. Gartner Peer Insights highlights stability and catalog-management strengths, while G2 includes both praise for integration and warnings about usability or integration headaches. That mix is plausible for enterprise infrastructure software: strong operator depth often comes with process complexity. FeaturedCustomers and review aggregators further suggest Mirakl has built a reasonable public proof base, but they cannot substitute for renewal data. Public evidence is much better on expansion than on retention. Case studies show operators adding more brands, more sellers, more SKUs, new categories, or deeper operational integration over time. But no public NRR, GRR, logo churn, or contract-length disclosure was found. So the customer story looks sticky and expandable in theory, yet still lacks the metrics that would prove durability quantitatively. The review evidence is therefore useful as a texture layer around satisfaction and deployment complexity, but not sufficient to quantify renewal probability or net retention on its own.[CU020, CU021, CU022, CU023, CU024, CU025]

6.4 Expansion, concentration risk, and customer verdict

The same enterprise focus that makes Mirakl credible also creates blind spots. A company selling large, complex deployments to major operators may have excellent contract quality but still face hidden concentration risk if a few major accounts drive a large share of ARR or GMV. Public evidence cannot resolve that question for Mirakl. Nor can it show whether any very large customers are in renewal windows, aggressively negotiating, or delaying expansion modules. The right conclusion is that Mirakl’s customer base is real, enterprise-grade, and operationally meaningful. The company has better named proof than many private software firms. But from an underwriting perspective, the chapter still ends with missing account-level retention and concentration data that only internal reporting can answer. That means an investor can be confident that Mirakl has real customers and real production usage, while still being unable to determine whether future growth is broad-based across cohorts or overly reliant on a smaller flagship group.[CU028, CU029, CU032, CU034, CU035, CU036]

6.5 Exhibits

7.1 Regulatory and legal risk

Mirakl sits directly in the blast radius of platform regulation. Its customers operate marketplaces, seller ecosystems, catalogs, payouts, and increasingly AI-moderated or agentic buying flows — exactly the areas where Europe and other jurisdictions are tightening rules. The Digital Services Act raises diligence expectations around illegal products and platform integrity, DAC7 imposes seller-reporting duties, and payment rules such as PSD2 matter wherever marketplace payouts or agentic transactions cross into regulated payment behavior. Mirakl’s own product roadmap now explicitly references this pressure through Trust & Safety and agentic-commerce governance language. The good news is that Mirakl appears more legally mature than an early-stage startup. The privacy policy, legal center, and DPA collectively show a more developed contractual and data-protection posture than many private software companies. The bad news is that legal maturity does not remove the operating burden: product, compliance, moderation, and customer-success teams still have to turn those legal obligations into repeatable workflows in live customer environments.[CR001, CR002, CR003, CR004, CR005, CR006]

7.2 Operational, security, and AI risk

Mirakl’s core platform looks operationally serious, but that raises rather than lowers the standard it must meet. A company claiming multi-cloud redundancy, 99.997% uptime, and enterprise-grade controls is implicitly promising customers that its software can sit inside critical commerce operations without interruption. Certifications and the CSA STAR listing are helpful trust signals, and the Cyber Week traffic proof is encouraging. But cyber and reliability failures rarely arrive as clean refutations of a marketing claim; they arrive through misconfigurations, identity-policy gaps, SaaS role mistakes, vendor sprawl, and data leakage at the edges of a complex stack. AI creates another layer of risk. Adversa’s incident catalog makes clear that prompt injection, unsafe agent behavior, and action-oriented AI failures are already causing real losses. Mirakl’s push into Nexus and Trust & Safety is strategically sensible, but it means the company must manage not just catalog correctness, but also the governance of autonomous or semi-autonomous commerce actions. In other words, Mirakl is moving from marketplace infrastructure into decision-bearing infrastructure.[CR009, CR010, CR011, CR012, CR013, CR014]

7.3 Partner, model, and execution risk

Mirakl’s risk is distributed across dependencies. The company depends on cloud providers, customer-system integrations, seller and supplier data quality, payment partners, and the ability of enterprise customers to operate good governance processes on top of its tools. The J.P. Morgan relationship illustrates both the upside and the dependency problem: Mirakl can accelerate agentic-commerce readiness by partnering with a global payments player, but it also ties part of the promise to another company’s fraud controls, transaction infrastructure, and roadmap timing. Execution risk rises further because Mirakl is expanding on many fronts at once. Core marketplace operations are mature, but Ads, Connect, Payout, Trust & Safety, and Nexus each add new GTM, product, and support demands. Customer reviews suggest the platform can be stable and feature-rich while still being complex to integrate and operate. That combination makes execution risk less about code quality alone and more about whether Mirakl can keep the whole operating model coherent as the surface area expands.[CR017, CR018, CR019, CR020, CR021, CR022]

7.4 Mitigations, kill criteria, and risk verdict

Mirakl is not exposed to these risks without defenses. Trust & Safety, privacy and DPA scaffolding, security certifications, published controls, and peak-season resilience all provide meaningful mitigation. The platform looks more institutionally prepared than many private peers. Even so, the most dangerous risks are the ones public evidence cannot close. Retention, concentration, and module-economics opacity remain the clearest hidden model risks; compliance and AI-agent safety remain the clearest operating risks; and dependency complexity remains the clearest transmission channel from incident to revenue or valuation damage. That leaves the investment committee with a medium-high residual-risk picture. The company is not screaming “fragile,” but it is entering more regulated and safety-sensitive workflows while still asking outsiders to accept a meaningful amount of private information asymmetry. In practice, that means diligence should focus less on generic software risk and more on the specific events that would break the thesis: a major compliance incident, a failed AI-adjacency rollout, or evidence that the largest customers are not renewing or not buying the newer modules.[CR024, CR025, CR030, CR031, CR032, CR034]

7.5 Exhibits

8.1 Current pricing context and why multiple discipline matters

Mirakl enters the valuation chapter with two seemingly contradictory truths. First, the company is clearly real and scaled: it reported $177 million of ARR and positive EBITDA in 2024, then $218 million of ARR and group-wide profitability in 2025, while supporting 450+ marketplaces and more than 100,000 sellers. Second, the only widely visible equity valuation anchor is still the September 2021 Series E that priced the company at more than $3.5 billion. That means the underwriting question is not whether Mirakl is a quality business; it is whether a four-year-old private headline still offers upside after the denominator has grown and public-market commerce multiples have reset. The multiple math already answers part of that question. The old mark has compressed from an extreme 2021-era private multiple to about 19.8x 2024 ARR and about 16.1x 2025 ARR, so Mirakl has grown into the price. But growing into a valuation is not the same as creating a new discount. Public evidence still does not disclose the exact quality of that ARR, the gross margin behind it, or the capital-structure terms that determine what a new investor would actually own. TV001 and TV002 therefore set a track recommendation rather than a buy call, while FV001 shows how scale, profitability, and disclosure gaps interact.[CV001, CV002, CV003, CV004, CV005, CV006]

8.2 Public comp range and what it implies for Mirakl

The public comp set is useful precisely because it is uncomfortable. Shopify still commands a premium, around 13x market-cap-to-revenue in July 2026, because investors view it as a large, high-quality, profitable commerce platform. VTEX and BigCommerce sit much lower, around 3.1x and 1.15x respectively, because their public disclosures make it easier to see slower growth, narrower platform breadth, or lower confidence in long-run margin power. Mirakl’s last disclosed private mark at about 16.1x ARR therefore does not merely sit above mid-tier public comps; it lands in the same neighborhood as the strongest public commerce platform while offering far less disclosure and no liquidity. That does not mean the Mirakl mark is impossible. ARR can deserve a premium to GAAP revenue, and Mirakl’s enterprise marketplace leadership, profitability, and adjacencies give it a better qualitative story than the lowest multiple names. But the comp spread makes one point unavoidable: the burden of proof is on premium valuation support. TV004 and FV002 use the current comp data to show that even modest multiple changes move enterprise value materially, which is why entry discipline matters more here than generic admiration for the company.[CV012, CV013, CV014, CV015, CV016, CV017]

8.3 Thesis, anti-thesis, and scenario view

The positive thesis is straightforward. Mirakl is a category-defining enterprise marketplace platform with unusually strong public proof for a private company: it is profitable, it has large and recognizable customers, more than 35 of those customers now exceed $100 million of annual marketplace GMV, and newer businesses such as Connect and Ads are real enough to demonstrate that the company can monetize beyond the historical core. If those adjacencies scale, and if the agentic-commerce layer eventually becomes paid infrastructure rather than only narrative, Mirakl can reasonably defend a premium outcome. The anti-thesis is equally concrete. Public evidence still cannot answer the return-determining questions that matter most at a premium private price: NRR, churn, customer concentration, gross margin, free cash flow, debt draw, and the actual preference stack. In addition, the newest story layer—Nexus and agentic commerce—is strategically interesting but not yet shown as a meaningful ARR driver. That combination leads to scenario underwriting rather than precision. TV003 and FV003 place the center of gravity below the last headline mark, with bull-case support only if profitability and attach continue while disclosure quality improves toward public-market standards.[CV021, CV022, CV023, CV024, CV025, CV026]

8.4 Exit readiness, kill triggers, and final diligence asks

Mirakl is closer to exit readiness than many late-stage private software companies because it already publishes real ARR, GMV, and profitability signals. Even so, it is not ready for a confident buy recommendation from public evidence alone. Public-company and sponsor comps both show the same pattern: investors pay up for quality only when disclosure is good enough to defend margins, growth durability, and cash conversion. Mirakl has not yet provided that level of evidence. The absence of a fresh price-discovery event since 2021 increases the importance of the next financing, tender, or exit process because that will reveal whether outside buyers still support the headline mark. For that reason the final recommendation is price-sensitive and evidence-sensitive. A buyer should not underwrite upside mainly on brand, category creation, or AI messaging. The work list in TV005 and TV006 is specific: verify the audited ARR bridge, retention, margin, concentration, debt utilization, and full liquidation-preference stack before treating the last public mark as investable. If those checks are strong, the recommendation can improve; if they disappoint, the downside reset arrives quickly. That asymmetry is the core reason the chapter ends at track, not buy.[CV026, CV027, CV035, CV036, CV037, CV038]

8.5 Exhibits