There is no such thing as a fully self-built sportsbook. Every operator on the market—from regional mono-sports books to the major US sportsbooks processing billions in handle—runs on a stack assembled from external vendors. The question has never been whether to buy third-party components. The question is which ones, on what terms, and at what point the revenue share math turns against you.
This article maps the actual economics of the iGaming buy-vs-build decision across each layer of the sportsbook stack: what buying costs today, what lock-in actually looks like in practice, why AI is deepening third-party dependency rather than reducing it, and where the one layer worth building in-house sits.
Market StructureWhy ‘Buy’ Is the Default: A $13.48B Industry Built on Outsourcing
The B2B segment accounts for 55% of the global iGaming platform and sportsbook software market, which reached $13.48 billion in 2024 and is projected to hit $17.59 billion by 2026, according to Industry Research analysis. That figure represents the scale of a market entirely predicated on the assumption that operators will buy rather than build core technology.
The components involved are not peripheral. A functional sportsbook stack requires a Player Account Management system, an odds engine, a payments layer, a content aggregation platform, a CRM system, and a compliance/certification layer. No single operator builds all of these in-house. The question is never “should we use third-party components?” but rather “which components, from which vendors, under what contractual terms?”
This structure is not new, but its scale has accelerated sharply. White-label and managed services models expanded by approximately 35% in 2024 alone, enabling more than 400 new operator launches globally, according to industry market data. For new market entrants across regulated jurisdictions—whether entering the US state by state, expanding in LatAm, or launching in newly regulated European markets—white-label is the default entry path. Building a platform from scratch to compete with an established B2B vendor’s certified, already-integrated stack is simply not a viable option for most operators at launch.
The pattern holds at the component level too. According to industry analysis, approximately 55% of new operator contracts in 2023–2025 favor API-first integrations and managed services as the primary delivery model, cementing modular widget procurement as the dominant integration paradigm. Operators don’t install software anymore—they plug in managed services via API and iFrame, adding capabilities in days rather than months.
Build Costs €2.5M–€5M. Buy Costs €500K. The Math Is Clear—Until It Isn’t
The initial cost comparison is decisive. Building a full iGaming platform in-house to MVP costs between €2.5M and €5M and takes 12–24 months, according to GamingTec’s industry analysis. This figure excludes ongoing compliance maintenance, security certifications, and the engineering headcount required to keep the platform operational and competitive. A licensed B2B solution, by contrast, costs €500K–€1M in year one and launches in approximately six months. The capital efficiency of buying is not marginal—it is an order-of-magnitude difference at launch.
For new operators and expansion market entrants, this math is unarguable. Capital that would be absorbed by a two-year build cycle is instead deployed into player acquisition, bonus budgets, and marketing. The time-to-revenue gap alone—18 months faster to first depositor—makes the buy case dominant for any operator without existing proprietary technology assets.
But the cost structure changes dramatically at scale. White-label platforms typically charge operators between 10% and 50% of Net Gaming Revenue in ongoing revenue share. This is not a flat fee—it is a percentage of every euro of margin the operator generates. As player volume grows and GGR scales, the revenue share becomes an increasingly punishing line item.
The hidden costs compound further. White-label solutions add an estimated 15–25% to operational overhead through support dependencies, vendor-controlled release cycles, and integration constraints that require workarounds. An operator who needs a custom bonus mechanic, a non-standard market type, or a localized UX adjustment is entirely at the vendor’s discretion for prioritization and timeline.
Most critically, migration is prohibitively expensive—not primarily because of technical switching costs, but because player accounts, active bonuses, and historical transaction data are held by the vendor. Switching platforms means either negotiating a data export (which vendors have little incentive to facilitate cleanly) or asking players to re-register, destroying retention in the process. The initial buy decision locks in far more than the operator typically anticipates at signing.
Vendor Lock-InThe Dependency Trap: When Your Widget Vendor Owns Your Stack
Vendor lock-in in iGaming is structural, not accidental. Some B2B providers design integrations to be deliberately difficult to replace—using proprietary data formats, non-standard API patterns, or tightly coupled module dependencies that make removing a single component technically complex even when contractually permitted. Operators who assumed they were buying a modular widget discover they have acquired an architectural dependency.
The migration problem is compounded by data hostage dynamics. Player accounts are not just records in a database—they carry bonus balances, pending withdrawals, loyalty tier history, and regulatory identity verification data. Moving this data between platforms requires vendor cooperation, typically triggers KYC re-verification requirements, and creates compliance exposure during the transition window. In practice, most operators who want to switch platforms find the operational cost of migration exceeds the ongoing cost of staying, regardless of revenue share trajectory.
Regulatory complexity amplifies dependency further. Compliance certifications across regulated jurisdictions—MGA, UKGC, DGOJ, Spelinspektionen, and others—are maintained by the B2B provider, not the operator. Initial security and compliance infrastructure for a single regulated market costs between $50,000 and $200,000 to set up from scratch. Operators relying on white-label providers effectively rent these certifications—paying in perpetual revenue share for access to compliance infrastructure that would be prohibitively expensive to replicate independently.
The speed that makes buying attractive also creates stickiness. iFrame and API widget deployments can be live in one to three weeks for pre-assembled betslip components, sport line displays, and event banner integrations. That speed advantage at deployment becomes an architectural sunk cost over time: the faster an operator integrates, the less likely they are to have built the internal engineering capability to eventually replace what they bought. Speed-to-market and strategic flexibility pull in opposite directions, and the buy model optimizes firmly for the former.
AI Widgets Are Deepening the Buy Case—and the Dependency
The emergence of AI-powered B2B widgets has not disrupted the buy model. It has reinforced it. Approximately 30% of sportsbook vendors adopted AI-based odds engines and fraud detection capabilities between 2023 and 2025, according to market analysis. Building comparable AI in-house requires not just engineering resources but training data at scale—a dataset that most individual operators simply cannot accumulate. The AI buy case is not just about cost; it is about data access.
The performance numbers from AI widgets justify the procurement cost clearly. AI fraud detection integrations have reduced suspected fraud incidents by approximately 22% for operators who deployed them. AI-driven personalization widgets—betslip recommendations, event suggestion engines, CRM trigger logic—deliver up to 20% lift in player engagement versus rule-based alternatives.
The most direct evidence comes from a documented case study: Stats Perform’s OptaAI integration with Winner.ro in Romania generated a +31% increase in bet volume on markets enriched with contextual AI insights. Average stake size doubled. Average odds on enriched markets increased by 100%. This is not a marginal uplift—it represents a fundamental change in how players interact with markets when given richer contextual information at the moment of decision.
The implication for the buy-vs-build calculus is significant. An operator considering building their own AI odds context layer faces not just the engineering cost, but the cold start problem: their AI model will train on their own player data alone, while a B2B AI vendor trains across dozens of operator datasets simultaneously. The B2B vendor’s AI improves faster and outperforms single-operator models by a structural margin.
This dynamic means that AI is systematically shifting the build case for most sportsbook components toward buy—while simultaneously raising the strategic stakes of the components that remain worth building. If AI betslip and personalization widgets are better bought, the question of what to build becomes more focused, not broader.
Hybrid ModelThe Emerging Best Practice: Buy Commodity, Build Differentiation
The largest operators in the market have already resolved the build-vs-buy question in practice, even if the industry has not articulated a formal framework. Major US sportsbooks like DraftKings and FanDuel build one primary layer in-house: the trading and risk management system. Everything else—PAM, payments, content aggregation, CRM tooling—is sourced from third parties. The trading layer is where proprietary risk models create direct margin advantage. Everything else is commodity infrastructure.
Mid-tier and emerging operators have fewer resources to build even one layer. For them, the question has shifted entirely from platform-level to component-level. The conversation in 2025 is not “should we build our own sportsbook platform?”—that question was settled years ago. The conversation is “should we build our CRM widget, our bonus engine, our bet builder logic?” Modular procurement is the dominant paradigm, and the decision granularity has compressed accordingly.
A third route has emerged alongside buy and build: purchasing platform source code outright. Operators who take this path get the speed-to-market of a licensed solution without perpetual revenue share obligations. They launch fast on proven technology, prove the product-market fit, then evolve the platform at their own pace. This approach is growing in adoption among operators who have visibility into their revenue trajectory and want to plan an exit from revenue share dependency before it becomes a material cost center.
Demand for composable, API-first architectures reflects this shift. Demand for customizable platforms featuring API integration, real-time odds, and automated risk management grew 40% in 2024–2025. Operators want to buy components, not monoliths. They want contractual data portability and the ability to swap vendors at the component level without triggering a full platform migration. Approximately 55% of new operator contracts now favor API-first integrations and managed services as the primary delivery model—a structural shift away from tightly bundled white-label agreements toward composable procurement.
CRM and Personalization: The One Layer Worth Building
If the hybrid model is “buy commodity, build differentiation,” the question becomes specific: what is differentiation in a sportsbook? The answer has become clearer over the past seven years of US sports betting market development.
US sportsbook handle grew 2,387% from 2018 to 2025. Revenue grew 3,673% over the same period—faster than handle, because margin expansion was driven by high-margin bet types: same-game parlays, player props, and bet builders enabled by third-party widget capabilities. But handle growth and revenue growth do not explain which operators won disproportionate share. The operators who converted early adoption into durable retention did so through CRM that turned casual players into engaged ones. The widgets created the product surface; the CRM created the relationship.
This is the strategic case for building the CRM layer. Third-party AI betslip and event recommendation widgets produce measurable ROI—the 20% engagement lift figure is real and documented. But operators who cede CRM logic entirely to a vendor lose something more valuable than a single-campaign metric: they lose the ability to differentiate retention strategy. When every operator on the same white-label platform runs the same CRM vendor’s logic, player retention becomes a commodity. The operators who win on retention are the ones who control segmentation rules, trigger logic, and player lifecycle decisions at the architectural level.
The practical hybrid approach is clear: use third-party AI widgets for speed-to-market and for capabilities that require training data scale (betslip recommendations, odds context, fraud detection). Retain ownership of how those widgets are deployed, to whom, when, and in response to which player behaviors. Segmentation logic, reactivation trigger timing, bonus eligibility rules, and cross-product CRM sequences are where proprietary control compounds into competitive advantage over 12–36 month horizons.
An operator running BidCanvas’s CRM AI Wizard alongside third-party odds and content widgets is not building a platform. They are building a retention edge. The widget generates the content; the operator controls who receives it, why, and what happens next. That distinction—between AI content generation and CRM logic ownership—is where buy-vs-build maps to competitive differentiation in practice.
The research literature on dormant player reactivation and sharp money segmentation both confirm the same underlying pattern: operators with proprietary segmentation logic outperform those running vendor-default CRM rules. The delta is not marginal—it compounds across every campaign cycle.
Decision FrameworkA Practical Buy-vs-Build Framework for the 2025–2026 Operator
With a market projected to reach $44.6 billion by 2033 at a 14.22% CAGR, according to Global Growth Insights, the architecture decisions operators make today will shape their cost structures and competitive positioning for a decade. The framework below is not a theoretical model—it reflects how the most successful operators in the market are actually making these decisions.
Buy when:
- The component is commodity infrastructure—PAM, payment processing, compliance certification, content aggregation. No operator differentiates on whose KYC vendor they use.
- Speed-to-market is critical and the component has a 12–24 month build timeline. Every month of delayed launch is an acquisition and revenue opportunity cost.
- AI capabilities require training data at scale you don’t have. B2B AI vendors train across multi-operator datasets; single-operator AI models are structurally disadvantaged.
- Regulatory certification is required and the vendor already holds it. $50,000–$200,000 per market for self-certification is a prohibitive cost at early scale.
Build when:
- The component is a direct source of brand differentiation—UX, personalization logic, CRM segmentation rules, retention playbooks. These are the components where proprietary control creates compounding advantage.
- The component involves proprietary data you need to retain. CRM event data, player behavioral signals, and segmentation insights are valuable assets; keeping them inside your own systems preserves strategic optionality.
- Revenue share at your projected three-year scale will exceed the cost of building. Model the crossover point explicitly. For high-GGR operators, the 10–50% NGR revenue share on white-label platforms crosses the build-cost threshold faster than most operators anticipate at launch.
Hybrid default:
- Procure API-first with contractual data portability and exit clauses. The architecture of operators who want to buy today without being locked in tomorrow requires these provisions at signing, not at renewal.
- Buy commodity; build CRM logic. Use third-party AI widgets for speed and scale; retain ownership of segmentation, trigger logic, and player lifecycle decisions.
- Treat source code acquisition as a migration path, not just an upfront option. Operators who build toward platform ownership from day one structure their vendor agreements differently from those who treat white-label as permanent.
The stakes of getting this wrong are not abstract. A $44.6B market by 2033 means every percentage point of NGR paid in unnecessary revenue share, every player lost to a competitor with better CRM logic, and every architectural migration forced by vendor lock-in represents a meaningful capital allocation failure. The operators who compound competitive advantage over the next decade are the ones who are deliberate about this calculus today—not those who default to buy across the entire stack and discover the compounding cost at scale.
More Research