Amazon.com Inc.’s Q4 earnings call offered a window into how the company believes purchasing will evolve. In that future, the company’s vision is not centered on marketplaces or websites.

Instead, executives repeatedly described a future in which AI “agents” research products. In addition, agents may compare options and place orders on behalf of users inside enterprise systems. For wholesale distributors and B2B sellers, the signal is less about any single product and more about direction. Commerce may increasingly happen through software, not screens.

“The primary way companies will get value from AI is with agents — some their own, some from others,” CEO Andy Jassy told analysts on Feb. 5.

The comments came as Amazon outlined new AWS capabilities designed to help enterprises build and deploy what it calls agentic workloads, including tools that connect agents to data, identity, governance policies and compute resources at scale.

Amazon ranks No. 1 in Digital Commerce 360’s Top 2000 Database. The database is how Digital Commerce 360 tracks the largest North American online retailers by their annual ecommerce sales.

Amazon is also No. 3 in Digital Commerce 360’s Global Online Marketplaces Database. That database ranks the 100 largest such marketplaces by third-party gross merchandise value (GMV).

How Amazon views agents and B2B commerce

On the commerce side, Amazon pointed to Rufus, its AI shopping assistant, as an early example of how AI can shorten the path from research to purchase. Amazon said Rufus was used by more than 300 million customers in 2025, and that customers who used it were about 60% more likely to complete a purchase.

Jassy framed Rufus as a preview of how shopping could work when software agents, rather than people, handle comparisons and checkouts.

“Customers will increasingly rely on agents that can navigate, compare and transact on their behalf,” Jassy said.

He also acknowledged Amazon expects to work with “horizontal” third-party agents — AI systems that sit outside individual retailers — while arguing that retailer-specific agents may have advantages because they can draw on richer pricing, inventory and purchase-history data.

For B2B commerce, that distinction is significant. Many business purchases are repeat, rules-based and data-driven.

Much of the discussion centered on AWS, which reported $35.6 billion in quarterly revenue, up 24% year over year, and now operates at a $142 billion annualized run rate.

Jassy described new services that allow companies to train custom AI models on their own data and deploy agents that can operate across enterprise systems.

“Enterprises will want models trained on their own data at an early stage,” Jassy said. “That becomes part of their learning foundation.”

Amazon’s custom chips

Amazon also emphasized its custom silicon strategy, saying Trainium and Graviton chips now represent a combined annual revenue run rate of more than $10 billion, offering 30% to 40% better price performance than comparable processors. Amazon said Trainium underpins much of the inference workload for Bedrock, its AI model platform.

For distributors, these tools lower the cost and complexity of deploying AI for quoting, forecasting, customer service and order automation.

Amazon did not directly discuss its B2B marketplace during the call. But its AI strategy has several implications for distribution.

AI agents depend on clean, structured, machine-readable product content, pricing and availability data. Distributors with inconsistent data may struggle to participate as purchasing shifts to automated workflows.

If agents execute orders inside ERP and procurement systems, the ability to expose catalogs, pricing, credit terms and order status through integrations will matter more than storefront design.

Recurring purchases of maintenance, repair and operations (MRO); electrical; heating, ventilation and air conditioning systems (HVAC); plumbing; and safety supplies could occur increasingly without a human browsing a site. Agents may simply validate contract terms and place the order.

In an agent-driven world, buyers may favor suppliers whose systems consistently return accurate pricing, inventory and delivery commitments.

Amazon said it plans to invest $200 billion in capital expenditures in 2026, primarily in AWS and AI capacity. That scale of investment underscores how centrally the company believes AI and cloud infrastructure will be to future customer experiences.

For distributors, the takeaway is not that Amazon launched a new B2B feature. It is that Amazon is preparing for a world where buying is increasingly delegated to software — and the suppliers that succeed will be those whose systems are ready to transact with machines as easily as with people.

Sign up

Sign up for a complimentary subscription to Digital Commerce 360 B2B News. It covers technology and business trends in the growing B2B ecommerce industry. Contact Mark Brohan, senior vice president of B2B and Market Research, at mark@digitalcommerce360.com. Follow him on Twitter @markbrohan. Follow us on LinkedIn, X (formerly Twitter), Facebook and YouTube.