Could agentic commerce reach one trillion dollars by 2030?

Katharina Schumacher and Roger Roberts, with Katharina Giebel, writing for McKinsey & Company (QuantumBlack), argue that "agentic commerce—shopping powered by AI agents acting on our behalf—represents a seismic shift in the marketplace," moving toward "an integrated intent-driven flow" that replaces the boundaries between platforms, services, and experiences. McKinsey puts a number on the shift: "By 2030, the US B2C retail market alone could see up to $1 trillion in orchestrated revenue from agentic commerce, with global projections reaching as high as $3 trillion to $5 trillion, according to McKinsey research." These are the firm's own projections, not measured outcomes.

The mechanism McKinsey points to is a stack of emerging interoperability protocols rather than a single technology. It names Anthropic's Model Context Protocol (MCP), which lets agents "retain memory, reasoning, and objectives across environments"; the Agent-to-Agent Protocol (A2A), which lets agents "coordinate, negotiate, and complete tasks directly with one another" over standards like JSON-RPC and HTTP; Google's Agent Payments Protocol (AP2), which uses "cryptographically signed mandates" to link intent, cart, and payment; and the Agentic Commerce Protocol (ACP, named but not detailed in this excerpt). McKinsey also cites METR's finding that "the duration of tasks that large language models (LLMs) can reliably complete with at least a 50 percent success rate has been doubling every seven months since 2019," and says Claude 3.7 Sonnet reached a 59-minute time horizon in 2025 while "Claude 4.5 has extended this to more than 30 human hours" — offered as evidence that agents are becoming capable of the multistep chains agentic commerce requires.

This reframes the protocol layer that other notes treat mainly as a safety or accountability problem as also a market-opportunity argument. Where What infrastructure must exist beyond training agents? argues infrastructure is needed to attribute and remedy agent actions, McKinsey's framing is that the same kind of infrastructure (standardized protocols for identity, intent, and payment) is what unlocks the revenue. It also extends Will agents compete for attention just like users do?: McKinsey names the ecosystem of "adapters and enablers" — e-commerce platforms, payment providers, fraud prevention tools — that must become "agent compatible" for that shift in attention to pay off commercially, and attaches a dollar figure to the stakes of that competition.

The excerpt does not establish how McKinsey derived the $1 trillion or $3–5 trillion figures, what adoption curve or consumer-trust assumptions underlie them, or how much of the projected revenue depends on protocols like AP2 and ACP actually being adopted at scale rather than merely announced. It is a consultancy's market-sizing exercise paired with a capability trend (METR's), not an empirical measurement of agentic commerce happening today — the implication that the protocol stack will unlock this revenue follows only if adoption, trust, and regulatory acceptance proceed roughly as McKinsey assumes.