Global Industrial AIoT Platforms

This market spent a decade being scored on how well it could connect things. In its 2026 edition, Gartner scored it on something else: whether the AI sitting on top of those connections is safe enough to be allowed to act.

That is not a scoring change. It is a gate, and it closed during the year the category was renamed.

What the 2026 Magic Quadrant for Global Industrial AIoT Platforms recorded

The evaluation published on 15 September 2026 under the name Global Industrial AIoT Platforms. That detail is the correction. The page that stood here described a market transitioning toward the AIoT name and treated the shift as unfinished business. The 2026 quadrant finished it: the new name is the report's own title, not a forward-looking label attached to the old one.

The analysts are Scot Kim, Sudip Pattayak, Emil Berthelsen, Sushovan Mukhopadhyay and Avinash Dev Nagumanthri. Two names that carried earlier editions of this research, Wam Voster and Akhil Singh, are not on this one.

Twenty vendors are named in the edition. Sixteen are plotted across four bands, and four more appear as honorable mentions without a position on the figure. That makes the 2026 field the widest this market has been scored against.

One piece of Gartner's own estate has not caught up. The Peer Insights listing for the market still reads Global Industrial IoT Platforms (Transitioning to Global Industrial AIoT Platforms). The quadrant completed the rename in September 2026. The peer listing still carries the transitional label.

Four Leaders, two Challengers, and twenty names in total

Microsoft, Siemens, Litmus and ABB hold the four Leader positions, with Microsoft plotted furthest toward the upper right of the figure as that coverage reads it.

AWS and Bosch are the two Challengers. Univers, Cumulocity, AVEVA, Velotic and IROOTECH are the Visionaries. Infinite Uptime, Braincube, SUSE, Davra and Exosite are the Niche Players.

Four vendors are named without being plotted. Eurotech, Honeywell Technologies, Inductive Automation and XCMG HANYUN appear as honorable mentions, a form of recognition that carries no position and no axis.

Read the bands against the previous edition and one name moves in the direction vendors do not announce. Cumulocity was confirmed as a Leader in the September 2025 quadrant, and the 2026 figure places it in the Visionaries band. That is a downgrade in everything but the word. The vendors that move down are precisely the ones with no press release to point at, which is why a roster like this one is legible from trade coverage of the figure and not from the vendor newsrooms.

Litmus went from Niche Player to Leader in two editions

The page that stood here said Litmus moved from Niche Player to Challenger. That was true of the 2025 edition and is now a year out of date.

Litmus is a Leader in the 2026 quadrant. The climb reads Niche Player in 2024, Challenger in 2025, Leader in 2026: three editions, three bands, no plateau. The company's own announcement of 21 September 2026 carries the placement and the report's exact title.

What makes the trajectory worth reading rather than merely recording is what it tracks. Litmus sells an industrial edge data platform, the layer that normalizes more than 250 industrial protocols and orchestrates workloads next to the machine rather than in the cloud. That is the layer the renamed category centers. The climb is the market paying for it, and the price was set one protocol driver at a time.

Agentic AI stopped being a score and became a gate

Here is the edition's biggest substantive change, and it is a change of kind rather than of degree.

Vendors had to demonstrate at least five credible agentic AI capabilities embedded in their workflows. Safe agentic AI, governable through prompt filtering, human-in-the-loop controls, audit logs and policy-bounded actions, is now a mandatory platform feature rather than a differentiator a vendor may or may not have.

The consequence is structural. When a capability is scored, a vendor that lacks it places lower and stays inside the field. When a capability is required, a vendor that lacks it is not ranked at all. AI moved from a dimension of the evaluation to a threshold for inclusion, and every placement in the 2026 figure was earned by a vendor that cleared the floor first.

That changes what a placement tells a buyer. A Leader in this edition is not the vendor with the best AI story. It is a vendor whose AI can be governed, selected from a field that was not previously asked to prove the point.

Sixty-five percent by 2030, and the trust hurdle in front of it

Gartner's headline number for the market is a forecast with a gap in the middle of it. By 2030, 65 percent of AI agents from industrial AIoT platform providers will be running in production across asset-intensive industries, up from 3 percent today.

The obstacle is not capability. It is permission. As Gartner puts it, the market views industrial AIoT platforms as a foundational enabler of physical AI for efficiencies in autonomous operations, while vendors observe that customers are hesitant to let AI independently control physical assets and processes.

The vendor response is visible in what the field ships. Read-only and advisory modes, approval gates and immutable audit ledgers are the mechanisms appearing now to make an agent's action reviewable after the fact, which is what a buyer needs before granting authority in the first place.

ABB's roadmap is the clearest illustration of where this ends up. It moves from advisory recommendations toward closed-loop autonomous execution, and the same assessment notes that such deployments can require custom grounding, prompt governance and robust role-based access. Closed-loop control of a production line is not a switch a buyer flips. It is an integration project with a governance program attached, and the vendors selling it describe it that way.

Velotic arrived, Google is absent, and SUSE's Leader is a different report

Three names in and around this edition need separating, because two of them are easy to conflate and one of them is a silence.

Velotic is new to the field and placed as a Visionary on its first appearance. The company formed in March 2026 out of GE Vernova's Proficy business combined with PTC's Kepware and ThingWorx assets under TPG. A vendor assembled from three well-known industrial software properties, appearing in the quadrant in the same year it was created, is a consolidation data point as much as a placement.

Google does not appear in the quadrant, and it does not appear among the honorable mentions either. The report documents no reason for the absence. A hyperscaler of that size simply not being in this field is a fact about the market's shape, and the honest reading is that the published research does not explain it.

SUSE's Leader recognition in September 2026 belongs to the Container Management quadrant, a different report covering a different market. It is not a placement in this one, and the two are easy to mistake for each other.

What the rename still has not reached

The remaining gap is administrative rather than conceptual, and it will resolve on its own. The quadrant moved in September 2026. The Peer Insights listing has not.

There is a larger limit on what this edition can tell a buyer, and it is worth stating plainly. No market size, TAM or revenue forecast for this market appears in the available research. Anyone quoting a figure for it is quoting something else.

The question the rename implies is also the one the format answers least directly. Which AI is actually running in production at customer sites, on which assets, under whose authority? A quadrant scores capability and strategy across a field. It cannot tell you which vendor's agents are switched on in a plant that matters, and in a market whose name now promises exactly that, the gap is worth holding in view.

Three questions before you let an agent act

Which of the five capabilities is in production? The inclusion threshold asked for at least five credible agentic capabilities embedded in workflows, and credible is the operative word. Ask for the ones running at customer sites today, on live assets, and separate them from the ones on the roadmap.

What stops the agent? Prompt filtering, human-in-the-loop controls, audit logs and policy-bounded actions are the four mechanisms Gartner names as the definition of safe agentic AI. Ask which of the four your vendor ships, and ask to see an audit log rather than a description of one. A boundary enforced by a prompt is a suggestion.

Who carries the risk when it acts? ABB's move toward closed-loop execution arrives with custom grounding, prompt governance and role-based access attached, which tells you the work is real. Put the same question to your own vendor directly: when an agent acts on your line and the action is wrong, what is the record, who can read it, and what does the contract say. A vendor that has not answered this for itself cannot answer it for you.

Gartner does not endorse any vendor, product or service depicted in its research publications, and does not advise technology users to select only those vendors with the highest ratings or other designation. Gartner research publications consist of the opinions of Gartner's research organization and should not be construed as statements of fact. Gartner disclaims all warranties, expressed or implied, with respect to this research, including any warranties of merchantability or fitness for a particular purpose.

See also Higher Education SaaS Student Information Systems. Ellucian and Workday are the only two Leaders on Gartner's 2026 quadrant for college student information systems, and the market only got there after Ellucian bought its biggest rival's SIS and ERP business outright.

The neighboring coverage here is Connected Product Engineering Services. Connected products are everywhere, and Forrester's own research says far too many of them are bad. Building one well requires excellence in five disciplines that almost no single organization has all of.