Techaisle Analyst Insights

Depending on geography, between 7% and 12% of SMB and midmarket buyers use a cloud marketplace to discover software. The rest arrive at AWS Marketplace, or at any of its competitors, already decided. A partner or an ISV brings them, and they transact there for contract consolidation, committed-spend drawdown, and procurement governance rather than for anything resembling search.

Call it the Discovery Deficit. Cloud marketplaces have functioned as procurement rails, not demand engines. They close deals that were originated somewhere else, by someone else, usually a partner.

That gap is why the AWS Marketplace agentic procurement announcements matter, and it is also why most coverage is aimed at the wrong question. Whether AI improves marketplace search is not interesting. Whether a marketplace that has never originated demand in the smaller segments can begin to do so, once the buyer stops being a person typing keywords, is a different question entirely, with different consequences for everyone downstream.

Three changes, and what each one is actually buying

AWS Marketplace has made three structural changes that are easy to read as feature releases. Read against the Discovery Deficit, each is doing something more specific.

The first is the replacement of lexical search with conversational discovery. Agent Mode, launched at re:Invent 2025, lets a buyer describe a requirement in natural language, upload an RFP or a requirements document, and receive ranked recommendations with side-by-side comparisons. Conversational search converts better than keyword search, which is unsurprising. The more important change is in what the interface is for. A keyword catalog fulfills a decision the buyer already made, and works only for someone who knows what to type. A conversational one helps make the decision, and deciding is the step SMB and midmarket buyers have always outsourced, because they have no procurement function to run comparative analysis internally. That is also why so few of them discover software in a marketplace: a catalog that cannot help you decide is little use to someone who cannot decide alone.

The second is building for machines to read rather than people. Most web pages assemble themselves in the browser, so a crawler or an agent that arrives sees almost nothing. AWS builds Marketplace pages to arrive complete, which means an agent reading one gets the whole listing. It has also opened the catalog to direct queries through an MCP server, so a buyer's own AI assistant can ask it questions without visiting a page at all. Most platforms building AI discovery are building a destination and trying to keep the buyer inside it. AWS is doing close to the opposite, and that choice says more about the strategy than anything else in the set. Making the catalog legible to agents AWS does not own is a distribution choice rather than an experience choice, and it concedes that the buyer's first conversation about software will happen somewhere else. The competitive unit shifts accordingly, from whose marketplace interface is best to whose catalog is most readable by someone else's agent.

The third is the automation of the transaction, which arrives from two directions at once. Express Private Offers let a seller define rate cards, discount tiers, volume breaks, and qualification criteria in advance, so an offer can be generated and accepted without a human negotiating it, which lowers the cost of serving a small software deal. In June 2026, AWS cut its professional services listing fee from 2.5% to 0.5%, and takes it to zero where those services are bundled into a multi-product solution or attached to a Migration Acceleration Program engagement, which lowers the cost of attaching services to that same deal. Multi-product solutions then let software, third-party components, and professional services sit in one listing under one procurement flow. A fee that vanishes at the exact moment services are attached to software is a routing incentive rather than price relief, and it routes toward the bundled listing as the default unit of sale. All of this aims at deals neither AWS nor its partners could previously work economically, which are the same deals where the Discovery Deficit lives.

Individually these read as product announcements; together they describe a platform trying to convert itself from a procurement rail into a demand engine, which is a considerably harder thing to be.

The Composed Shelf

Worldwide IT spending by firms with 1 to 4,999 employees will reach US$1.667 trillion in 2026, excluding communication services, and the majority of it will go to IT services rather than to technology products. A market of that size, spread across every economy and every industry, sets the direction for commercial IT rather than following it. These firms are now spending more on the implementation, integration, management, and security of technology than on the technology itself, and the margin between the two is wide and widening.

That composition is the product of two forces working against each other. AI is pulling money up and forward, into software, infrastructure, and services that were not in the budget a year ago. Cost is pulling the other way, as component inflation, tighter budgets, and a higher cost of capital are pushing firms to defer what they can and to rent what they cannot. That second force is the quieter one, and it explains the tilt toward services better than any capability argument does. Buying an outcome instead of an asset moves cost from the balance sheet to the income statement, and it moves operational risk from the firm to the provider. In a year of expensive capital and unforgiving threats, that trade is worth paying for, which is why the money is moving toward services even where the technology itself is cheap.

Within services, the mix has shifted. Maintenance, support, and break-fix, the labor of keeping systems alive, once defined the SMB services market. The money is now concentrating in consulting, integration, and putting AI into production. Transformation work has overtaken recurring management, and it is not close.

Cisco Owns the Control Plane of the Agentic Era. Nobody knows it yet.

The market is currently operating under the assumption that the architectural gravity of AI belongs entirely to the orchestration layer of the hyperscalers or the workflow engines of SaaS giants. But those software surfaces only control logic within their own proprietary walls or virtual boundaries. When an autonomous agent goes rogue, encounters a looping cost explosion, or faces a machine-speed exploit, that liability manifests in the physical world as a network routing challenge, a telemetry event, and a data-fabric security crisis.

By building the infrastructure that unifies visibility and enforcement from the silicon to agent-action trust, Cisco has quietly captured the layer that governs how autonomous workloads actually execute.

Cisco did not join the AI conversation. It redefined it.

For 2 years, enterprises have funded the AI buildout as a capacity race, measured in GPUs, power, and capex, on the assumption that compute is the scarce input. It is not. Compute that cannot be connected, secured, and operated at scale is stranded capital, and most AI infrastructure budgets have underfunded the layer that decides whether the GPU spend ever produces a business outcome. Cisco used Cisco Live 2026 to name that gap and claim it. Capacity commoditizes. Control compounds. The contest that decides the next decade of enterprise infrastructure is the contest for the control plane of agentic AI, from programmable silicon to agent-action trust, and Cisco is the only company holding the full stack.

That reorders the buying decision. If control, rather than capacity, is where durable value accrues, the criteria most businesses use to select AI infrastructure are wrong-footed, because the vendor best positioned is not the one selling the most compute but the one that governs how compute is connected and trusted. Cisco just claimed that position, and every announcement at the event is a move to occupy it.

The swarm breaks the assumptions networks were built on

Two years into the generative AI gold rush, the spreadsheet is starting to call the question. IBM's own CEO study, released around Think 2026, found that only 25% of enterprise AI initiatives are delivering expected ROI, and just 16% have scaled enterprise-wide. Techaisle's own GenAI adoption research confirms the same gap from the buyer side: midmarket organizations plan a 27% average increase in GenAI spending for 2026, yet 45% of mid-sized firms remain stuck in pilot purgatory, unable to move workloads into production.

The capital has moved. The returns have not.

This is the gap Arvind Krishna walked onto the Boston stage to occupy. His framing was simple and, in its way, audacious. “The enterprises pulling ahead are not deploying more AI. They are redesigning how their business operates.” That sentence reframes the entire industry conversation. Not better models. Not bigger clusters. Not cheaper tokens. A different operating model.

It also reframes IBM.

Last year, after the IBM Analyst Forum, in September 2025, Techaisle defined IBM as the Vertical Integrator of Transformation, a company that owns the foundation (Red Hat OpenShift), the components (watsonx), and the factory (IBM Consulting), and ties them together with a single point of accountability. That frame held. Twelve months later, IBM has done something harder than extending it. The company has made the integration itself the product.

I am calling this evolution the Operationalization Premium: the durable economic advantage that accrues to vendors who solve the boring, expensive, regulated middle of enterprise AI, the part hyperscalers and frontier labs largely cede. Think 2026 was not a model launch. It was the most coherent operating-system play any incumbent has made for the agentic enterprise. The question for the next year is whether IBM can charge for it.

The Thesis: AI as an Operating Model, Not a Capability

IBM's central claim at Think 2026 is that enterprise AI failures are not model problems. They are architecture problems. Models are commodified. Inference will continue to fall. What organizations cannot buy off the shelf is the operating layer that lets agents act on connected data inside a governed infrastructure, with auditable outcomes.

IBM is now organizing its entire portfolio around four interlocking systems: agents, data, automation, and hybrid. The framing is not new; every firm has some version of it. What is new is that IBM has a product in the market across all four, with credible proof points, and a thesis that explicitly links them.

The boldness sits in the second-order claim. IBM is betting that the differentiated economic value of enterprise AI will not be captured at the model layer at all. That bet looks more credible the longer the ROI gap persists.