Funding · Stockholm · October 13, 2026
The network layer physical AI is still missing
FleetMQ closes its pre-seed round, backed by Trio Impact Invest and Wallenberg Investments. A note from Frank Jiang, CEO & Co-Founder.
Autonomous machines are being deployed at an unprecedented rate, across nearly every industry. Keeping these machines safely connected while they are on the move still requires gluing together products to get basic remote operation working: video streamers, VPNs, data visualization, with some products costing thousands of euros a month.
When we started FleetMQ, the phrase physical AI did not exist yet. The current race to deploy physical AI is introducing a new wave of compute into the world, and unlike every wave before it, this compute moves and interacts with the physical world. It leaves buildings, navigates cities, flies across countries, and runs on networks that are constantly changing.
Every previous wave of compute scaled once a network layer was built for it. The internet scaled with TCP/IP, which let applications stop caring about which network they ran on. The cloud scaled with software-defined networking, where managing networking across massive data centers became a few lines of configuration. Each wave came with its own scaling pains that could only be addressed by evolving the network layer. That still has not happened for physical AI.
At FleetMQ, we are hard at work on this problem. We have always believed the answer is intelligent middleware. Now that machines are beginning to move around and interact with the physical world, they need a middleware layer that intelligently adapts how critical data is sent and received over networks that change as the machine moves.
Today we are announcing the closing of our pre-seed round, backed by Trio Impact Invest and Wallenberg Investments. Here is what our first phase proved, and what we plan for this new phase.
Our previous phase set out to prove that our intelligent middleware unlocks operational value. Over the past year we have done the research and development with key design partners to show that our intelligent middleware renders dedicated remote operation tools, like video streaming and VPN products, unnecessary, some of which partners were paying thousands of euros a month for. We've shown that running video, telemetry, and control through one middleware layer that manages them together is key to overcoming scaling challenges faced by physical AI applications.
This round sets out to prove that intelligent middleware accelerates physical AI deployment. A physical AI system should be able to diagnose its own connectivity problems and take full advantage of whatever network it is on, whether it's Wi-Fi, 4G, 5G, or satellite. Our platform already assesses link conditions continuously and adapts routing and streaming to how the network is actually performing. The work of this round is to take the platform we've built with our design partners to teams who need to accelerate the deployment of their physical AI solutions.
Today, physical AI teams still burn precious time on their roadmaps trying to work around wireless networks. The outcome we are building toward is that teams become fearless around networks, and focus on building and deploying autonomous systems.
Frank Jiang, CEO & Co-Founder
The investors backing FleetMQ's next chapter
Trio Impact Invest and Wallenberg Investments' investment unit Ventures have both backed deep tech companies in Sweden's most important industrial sectors. They invest in companies where research meets an industrial problem nobody has solved. Together, their support speaks to what this team has already shown in the field, the scale of the problem, and the company's approach to building out the network layer that physical AI is still missing.
In their own words:
“The hard part in remote operation isn't autonomy or network capacity, it's the middleware in between, where bandwidth and latency swing unpredictably once a machine is moving and no single network covers the whole operation. FleetMQ has productized that layer, working across cellular, Wi-Fi and whatever else is on site, and sitting between the network providers, the OEMs and the autonomy vendors. As fleets scale, that is where the operational picture comes together and usage data becomes an asset.”
“We have followed the FleetMQ team since the early days of their research and have always been impressed by the foresight and insights of the team regarding the challenges around connectivity for moving machines. When we look into the future, we see that the network environment will get more complex, not less, and this is why we believe that FleetMQ's solutions will be key for successful operations of physical AI systems at scale.”
About Trio Impact Invest
Trio Impact Invest is a Swedish venture capital firm investing in research-intensive, early-stage deep tech companies with the potential to create significant commercial and societal impact. They back enabling technologies across fields including life sciences, advanced materials, energy and climate, industrial technologies, AI and digital infrastructure, and security and resilience. As an active long-term partner, they work closely with founders to transform breakthrough science and technologies into globally competitive companies.
About Wallenberg Investments
Wallenberg Investments administers the Wallenberg Foundations and manages and develops their assets. The largest assets of the Foundations are the partly-owned industrial holding company Investor, listed on the Stockholm Stock Exchange, and the wholly-owned holding company Wallenberg Investments. Wallenberg Investments brings the Foundations' grant-making activities, asset management and corporate governance together in a single organization with five investment units: Industrials, Real Assets, Alternative Funds, Special Investments, and Ventures.
About FleetMQ
FleetMQ (FMQ Technologies AB) builds the network layer for physical AI. Its intelligent middleware keeps moving machines connected in real time, replacing assembled remote operation stacks with a single layer that adapts to live network conditions. FleetMQ was founded out of research at KTH Royal Institute of Technology and is based in Stockholm.