Desktop as a Service

The 2026 Magic Quadrant for Desktop as a Service, published August 5, 2026, contains a projection that renames the product's customer: by 2030, 20 percent of DaaS workspaces will serve AI agents rather than humans, up from nearly zero in 2026. The virtual desktop, the technology built to deliver a human a screen, is being repurposed as a home for software that has no screen.

The workspace with no human at the desk

DaaS was the answer to a simple question: how does a worker get their desktop, with its applications and files, on any device, from anywhere, with IT keeping control of the middle.

The 2030 projection says the question is changing. A fifth of the workspaces will not belong to workers. They will belong to agents, the software that needs a controlled environment to run its tasks, with an identity, a permission set, and an audit trail, exactly the properties a managed virtual desktop already provides.

A workspace for an agent needs no screen, which changes everything, from how the environment is sized to how it is secured to how it is priced. The market is quietly inheriting a second customer, and the quadrant is the first place it shows up as a number.

The August 2026 Magic Quadrant for Desktop as a Service, in six Leaders

The edition published August 5, 2026, authored by Stuart Downes, Sunil Kumar, and Todd Larivee.

Six vendors hold the Leader rung. Microsoft, for the fourth consecutive year, highest on Ability to Execute, on Windows 365 Cloud PCs. Citrix and Amazon Web Services, the incumbent virtual desktop estates. Omnissa, the VMware lineage continuing under its new name. Alibaba Cloud, entering the Leader quadrant for the first time, the only Asia-Pacific vendor on the rung. And Parallels, a Leader for the first time, on its vendor-assembled DaaS and its self-assembled Remote Application Server.

The Visionary is XTIUM, the outcome-based DaaS provider whose merger with Evolve IP widened its scale. The Challengers are Access, Apporto, and Anunta. The Niche Players include ZTE, Dizzion, oneclick, and IronOrbit.

The three-step jump

Alibaba Cloud's entry is the edition's fastest trajectory: Niche, then Challenger, then Leader, three placements in three years.

The citation is built on AI-native positioning. The Wuying AI Cloud PC supports GPU workloads up to 48 gigabytes of VRAM, positions DaaS as a foundation for AI agents, and carries a published 99.975 percent SLA. The AI cloud PC is the market's own answer to the 20 percent projection: the desktop platform being pre-built for the agent era rather than adapting to it afterward.

Three quadrants in three years is a market vote, and the vote was cast by the GPU economics the agent era runs on.

What agent workspaces actually need

The 20 percent projection has specific engineering consequences, and the Leaders are already shipping toward them.

An agent workspace needs compute shaped like the agent's job: GPU where the agent does vision or inference, plain CPU where it does workflow. It needs identity and permission boundaries at the workspace level, so the agent can touch only what its role allows. It needs observability, because an agent's session log is the audit trail. And it needs to be cheap enough to spin up and down at machine speed, because agents do not respect business hours.

The virtual desktop already solves most of that. That is the market's quiet luck: the technology built for the remote worker is, feature for feature, close to what the remote agent needs.

The market math

The market's economics support the pivot. Gartner forecasts DaaS spending to grow from 4.9 billion dollars in 2026 to 6.5 billion by 2030, and the agent projection rides on top of that growth rather than replacing it.

The interesting reading is the overlap. The human workspace market is growing on hybrid work, and the agent workspace market is growing on agent adoption, and the same platform serves both from the same capacity pool. The vendors that win the human market get the agent market's infrastructure for free, and the quadrant's Leaders are the ones positioned to take both.

What the projection leaves open

The honest limits are the projection's own conditions. Twenty percent by 2030 is a forecast, not a law, and the agent population's appetite for full desktop workspaces is unproven; many agents will run in lighter environments, containers, functions, and the DaaS market will get only the fraction that needs a full managed desktop.

The other limit is the usual one: six Leader placements are confirmed, the field's remaining write-ups have not been widely republished, and the cautions, particularly Parallels' published constraints around organizational size, certifications, and geographic concentration, are the model for what a buyer should demand from every vendor on the rung.

Four questions for the workspace buyer

Who is the workspace actually for? The 2030 projection makes this a real procurement question. A platform bought for humans and later repurposed for agents needs different sizing, security, and licensing terms. Ask about the agent path now.

What is the SLA, in writing? Alibaba Cloud's 99.975 percent is public. Demand the same specificity from every vendor, with the credits attached.

Does the vendor publish its cautions? Parallels' constraints are public because the vendor engaged. Ask each vendor for its caution list, because the one that will not share it is the one carrying the largest one.

Where does the workspace data live? The DaaS estate holds the organization's files, applications, and now agent credentials. Ask the residency and exit questions before the pilot, not during the renewal.

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.

A closely related read is Digital Adoption Platforms. Gartner's guide says the category is shifting from show-me tutorials to do-it-for-me agents that run the software on the user's behalf, which is also the biggest threat to the category's own reason for existing.

The neighboring coverage here is Digital Employee Experience Management Solutions. Most technology problems inside a company never become a support ticket. This category exists to make the gap between what IT measures and what employees actually experience visible.