The Agentic AI marketplace opportunity

Creating new CSP revenue streams

The communications industry is entering a new phase of AI adoption — defined by autonomous, goal-driven agents capable of reasoning, coordinating, and acting across complex environments. Agentic AI fundamentally changes how intelligence is delivered and how value is created.

For Communications Service Providers (CSPs), the real opportunity lies not in deploying individual agents, but in orchestrating multi-partner agent solutions to deliver measurable business outcomes at scale. This shift requires a new economic model one built on marketplaces, shared value creation, and outcome-aligned monetization.

Traditional AI monetization models, such as licenses, subscriptions, usage tiers, were designed for software executing predefined tasks. Agentic AI breaks this paradigm, so agents operate continuously, adapt dynamically, and invoke other agents, systems, and APIs to fulfill intent. As a result, pricing access to models or users no longer reflects value creation. Monetization must instead align to autonomous behavior and outcome delivery, particularly in interconnected and regulated telecom environments.

From Hyperscaler marketplaces to CSP-owned ecosystems

Today’s agent marketplaces are largely hyperscaler-driven. While they support third-party participation, they operate as cloud-centric, single-principal ecosystems in which hyperscalers define governance, commercial terms, and billing. CSPs can publish, sell and consume agent-based services, but often relinquish control over solution composition, value attribution, and solution monetization — positioning them primarily as consumers rather than economic owners.

A CSP- or consortium-owned Agentic AI Marketplace represents a different construct. In this context, an Agentic AI Marketplace refers to a governed platform that enables the onboarding of autonomous agents from CSPs and partners, supports intent-based requests from enterprises, users or operational systems, and dynamically discovers agents, composes, and executes multiple agents to fulfil those intents. The marketplace

provides the control plane for agent lifecycle management, policy enforcement, execution orchestration, and economic settlement across participating parties.

The marketplace supports multi-industry solutions even though it is deeply integrated with BSS, OSS, network connectivity, and edge environments, while supporting multi-vendor, multi-agent solution composition. Ownership plays a critical role in determining how value is captured. When CSPs control the marketplace, they shift from paying for intelligence to monetizing autonomous outcomes enabling governed collaboration, ecosystem innovation, and sustainable value creation.

The monetization stack: offering, revenue, and pricing

Monetizing an Agentic AI ecosystem requires separating three tightly linked layers: the offering model (what is sold), the revenue model (who pays and why), and the pricing model (how value is measured and settled). In agentic environments, traditional telco pricing constructs alone are insufficient. Autonomous agents execute iterative workflows and coordinated interactions across operational boundaries, requiring pricing models to evolve toward dynamic, usage- and outcome-based economics. Tokenomics provides the foundation for measuring, governing, and monetizing autonomous behavior.

Partner agents and marketplace monetization

A defining feature of the Agentic AI Marketplace is partner participation. Vendors, systems integrators and CSPs contribute agents for telemetry, orchestration, order management, billing, and discovery. These agents are monetized through a mix of acquisition fees, execution-based charges, and platform usage fees.

Agents become economic participants within a governed ecosystem. Each time a composed solution executes, contributing agents generate measurable value and earn revenue based on their role in fulfilling intent. This creates a scalable innovation engine: partners are incentivized to enhance capabilities, while CSPs gain transparency, governance, and automated settlement.

Tokenomics: pricing autonomy, not models

Tokenomics acts as the economic control plane for Agentic AI. Unlike LLM token counting, agentic tokens represent units of autonomous work: reasoning steps, planning actions, API calls, agent-to-agent negotiation, and SLA decisions. Tokens make autonomous behavior measurable, governable, and billable.

By tying cost to intent fulfillment rather than model access, tokenomics aligns pricing with outcomes and enables policy-driven cost controls. It also supports automated, agent-level

settlement across ecosystem participants — allowing partners to earn usage-based compensation, CSPs to capture outcome- and SLA-based value, and marketplace operators to earn orchestration fees without manual reconciliation.

Enabling B2B2X through revenue sharing

Agentic AI marketplaces enable a new class of B2B2X solutions, dynamically composed from agents contributed by CSPs and partners. By combining capabilities such as network slicing, edge compute, and telemetry, CSPs can deliver industry-specific solutions across manufacturing, sports, and immersive experiences. Revenue is shared based on each agent’s contribution, with rules defined during onboarding and enforced automatically by the marketplace.

Beyond incremental revenue, Agentic AI ecosystems drive cost optimization through autonomous operations, productivity gains through self-optimizing workflows, and growth through new digital services and premium SLAs. More fundamentally, CSPs evolve from connectivity providers to AI-native service orchestrators — curating ecosystems where intelligence, infrastructure, and innovation converge.

Agentic AI: Catalyzing industry transformation

Agentic AI represents not just a technology shift, but a new economic model for the communications industry. CSP-owned marketplaces, grounded in ecosystem participation and token-based monetization, offer a sustainable path to value. As the industry moves toward intent-driven autonomy, the CSPs that succeed will be those that treat Agentic AI not as a tool to deploy, but as an ecosystem business to lead.

The views reflected in this article are the views of the author and do not necessarily reflect the views of the global EY organization or its member firms.