The End of One Throat to Choke
For three decades, the pitch for a single, monolithic ERP suite was simple: one vendor, one data model, one contract, one throat to choke when something breaks. Forrester's Enterprise Resource Planning Solutions Landscape for the first quarter of 2026, covering 29 vendors, confirms that pitch is losing ground. Forrester principal analyst Faram Medhora describes a market that is mature and saturated in large enterprise adoption, increasingly defined not by new ERP purchases but by modernization of what is already installed, a shift that opens a very different architectural question than the one CIOs faced a decade ago.
That question is whether to replace the monolith with another monolith, or to break ERP into a federated set of best-of-breed, API-connected applications orchestrated from the top. Justin Overdorff, a partner at Lightspeed Ventures, puts the alternative bluntly: buyers increasingly want tools that snap together like Lego pieces rather than a single suite that tries to do everything adequately. Lightspeed values the addressable ERP market at roughly 150 billion dollars, large enough that startups now target individual ERP functions and the orchestration layer itself, rather than trying to unseat the incumbents outright.
The Deadline That Forces the Question
CIOs do not get to make this decision on their own timeline. SAP ends mainstream support for its Business Suite 7 systems on December 31, 2027, and Microsoft ends support for Dynamics GP on September 30, 2029, both hard deadlines that force a decision regardless of appetite for change. An estimated 60 to 70 percent of SAP's installed base has still not moved to the cloud, and cloud migrations of this scale typically take 18 to 36 months, meaning organizations that have not already started are running short on time to make a deliberate architectural choice rather than a rushed one.
Morgan Stanley equity strategist Simon Clark frames these deadlines as doing the deciding for companies that would otherwise keep deferring: aging on-premises systems approaching end-of-life create what he calls a forcing function for cloud migration. In a Morgan Stanley survey, half of CIOs said they plan to upgrade or modernize their ERP systems in the coming years, a figure that tracks closely with the share of the SAP base still running on legacy infrastructure. The deadline, not the AI roadmap, is what is actually moving budgets right now.
What "Unbundling" Actually Looks Like
In practice, unbundling means keeping a core ERP system of record for the ledger, source-to-pay, and order-to-cash processes that genuinely benefit from a single data model, while routing everything else, manufacturing execution, project delivery, supply chain orchestration, planning and analytics, through specialized applications connected by APIs. Forrester's landscape report names this orchestration layer as the fastest-growing part of the market, and notes that change management and organizational readiness, not technology integration itself, are now the primary barriers to pulling it off well.
Gartner analyst Johan Jartelius describes the broader direction as transforming ERP from a transactional system into an intelligent platform, where AI handles routine processing and humans focus on exceptions and judgment calls. That framing matters for CIOs deciding where to draw the unbundling line: core transactional modules with well-understood, codified logic are good candidates for AI-driven automation inside the core ERP, while judgment-heavy extended processes are better served by specialized best-of-breed tools that can evolve independently of the ERP's release cycle.
The Market Numbers Behind the Shift
The size of the opportunity explains why so many vendors and analysts are converging on the same prediction from different angles. Mordor Intelligence projects the cloud ERP market will grow from 56.53 billion dollars in 2026 to 138.56 billion dollars by 2031, a 20 percent annual growth rate, with private cloud ERP growing even faster at 22 percent a year, driven largely by banking and healthcare regulatory requirements that make pure public cloud deployment a harder sell. Grandview Research and Fortune Business Insights put overall ERP software growth at a more modest 9.5 percent and 13 percent annually through the early 2030s, while Morgan Stanley's Chris Quintero projects a more aggressive 17 percent.
The spread between those estimates is itself informative. Lower-end projections tend to model a continuation of the current replacement cycle, enterprises swapping one monolith for a newer one, while higher-end projections assume the unbundling trend accelerates and captures new spending on orchestration and specialized applications that would not previously have been counted as ERP at all. Gartner analyst Neha Ralhan adds a geopolitical dimension, pointing to data sovereignty rules that are pushing some organizations toward local and regional cloud providers rather than the dominant global hyperscalers.
AI Is the Argument for Both Sides
Gartner predicts that by 2030, more than 50 percent of foundational ERP tasks will be executed autonomously by AI, and Mordor Intelligence reports that organizations with AI-integrated ERP have already seen a 25 percent cut in delivery times and a 15 percent drop in operating costs. Those numbers are routinely cited by vendors defending the monolithic approach, on the theory that AI works best with a single, clean, unified data model to reason over rather than data scattered across a dozen best-of-breed applications with their own schemas and integration quirks.
Deloitte's So Chan, a US enterprise performance leader at the firm, makes the strongest version of that case: ERP is the system that makes AI's value real, because the clean, structured, governed data ERP enforces is exactly what gives AI outputs their accuracy and trustworthiness. That argument redirects the unbundling debate toward a narrower question: which functions need to supply the clean, governed data AI requires, while specialized applications handle orchestration around them.
The Decision in Front of You
For a CIO or CTO weighing an ERP decision ahead of a 2027 or 2029 support deadline, the practical takeaway is not to pick a side in the monolith-versus-unbundled debate in the abstract. It is to identify which processes in your business genuinely need a single system of record with unified data, and which processes are being forced into the ERP suite today only because that is where the contract happened to land them. Forrester's data suggests that line moves every year, and that the applications sitting on the wrong side of it are increasingly available as modular, API-connected alternatives.
The deadline pressure from SAP and Microsoft means this decision cannot be deferred indefinitely, but it also creates a rare opening to make the architecture choice deliberately rather than by default. Enterprises that use the forced migration to draw a clean line between core system-of-record functions and orchestration-layer functions will be positioned to adopt AI-driven automation inside the core while still capturing the flexibility of best-of-breed tools everywhere else. Enterprises that simply replatform the old monolith onto new infrastructure will be back at this same decision point well before the next decade is out.
