Manufacturing is undergoing the most significant architectural shift since the introduction of PLCs in the 1970s. Products are becoming more complex, production cycles more compressed, and system variability more volatile. Every factory leader we meet converges on the same truth that the intelligence their operations now demand is impossible to achieve with the current factory tech stack they have.

Aris Machina is the most promising company we’ve seen that can close that gap: not with another tool or dashboard that led to the infamous ‘Industry 4.0 fatigue’, but by rebuilding the data and intelligence layer of the modern factory from first principles.

This is why we’re backing Sid & Peter of Aris Machina and led their $ 10.7m pre-seed round earlier this year, with participation from Village Global, AENU, Planet A and Snö, and some exclusive angels. Founded in Stockholm in 2025, Aris Machina blends agentic AI and data processing with interdisciplinary expertise to develop tools that enable industries to handle complexity and increase productivity across design and manufacturing.

Factories Are Drowning in Data They Can't Leverage

The dominant constraint in modern manufacturing is no longer hardware, capital, or talent, but the inability to connect, structure, and reason over the influx of factory data. It’s the core reason AI has not yet transformed industrial operations the way it has started to show ROI in other ‘rusty & dusty’ industries. 

1. Data volume is exploding, yet data utility lags behind: Factories generate millions of signals per minute, yet only 44% of data is used effectively. Factories today are drowning in telemetry data, etc. but starving for context. Data is abundant, but intelligence is structurally scarce.

2. Fragmentation breaks context and blocks AI: The factory software stack (edge;  PLC; SCADA; MES; ERP) and several abstraction layers make this data somewhat accessible. However, it was engineered for control, not intelligence. Each layer transforms and compresses data differently, severing relationships between machine behaviour, process parameters, and quality outcomes. AI cannot properly reason over missing links in the ontology, but the current stack destroys important links, given the vertical and horizontal data silos.

3. Integration tax on every innovation: Integration complexity of data from different sources, especially extracting data from legacy systems, is the top barrier towards more data-driven decision-making in factories, not even talking about AI yet. According to our references, large factories routinely spend tens of millions per site try stitching systems together and still fall short… They miss a single unified player, providing end-to-end capabilities.

Peter Carlsson, Co-Founder & CGO of Aris Machina, summarizes the reality bluntly: “The demand for precision, speed, and adaptability in manufacturing systems has never been higher — yet the tools to achieve this are lacking, and fail to fully leverage the wealth of data available.”

Aris machina fixes the foundation, not the symptoms

The team at Aris Machina is building an Agentic Operating System (OS):

• connecting machines, systems, and data streams end-to-end,

• structuring them into a unified, dynamic ontology, and

• activating them through agentic workflows that reason, recommend, and coordinate action.

This is not another point solution. This is not another digital twin. It is the missing foundational intelligence layer for factories. Think of it like this: If Palantir redefined the enterprise ontology, Aris Machina is the verticalized, deeply integrated, and more grounded version purpose-built for the physical world of manufacturing.

The two initial workspaces make this tangible:

1. Gemba – factory intelligence in real time: The wedge and the heartbeat. Gemba connects directly with machines and middleware, mapping their signals into the shared ontology and deploying agents for troubleshooting, process control, and quality optimisation.

2. Protos — R&D, design, and physics-informed modelling: Workspace for designing and validating complex products (starting with batteries), closing the structural gap between design intent and production reality.

Siddharth Khullar, Co-Founder & CEO of Aris Machina, captures the long-term arc of this vision: “Think of what an App Store for manufacturing could be if there was an easy-to-integrate operating system on all existing manufacturing infrastructure.”

The ambition is not an incremental improvement. It is to rearchitect how factories think, learn, and adapt.

The team: a case study for founder-market-fit

In IndustrialTech, resumes matter, but not because of the logos. What matters is whether founders have built and scaled mission-critical industrial systems, understood their failure modes, and earned the conviction to rebuild them.

Sid led software and AI efforts inside one of the most advanced and smartest gigafactories globally, aimed at sustainable battery production, where precision, traceability, and uptime were existential. Peter founded Northvolt itself after scaling Tesla’s supply chain, giving him a uniquely deep intuition for where the factory stack breaks and what world-class manufacturing requires.

Here are some of the strong signals we picked up from the founders:

1. Extraordinary clarity of thought: From day one, their framing was unmistakably first-principles: “You cannot bolt AI onto a fragmented stack. You must rebuild the ontology first.”

2. Bold, credible vision for autonomous factories: Not buzzwords. Not hype. A concrete, technically grounded path from today’s brittle systems to agent-driven adaptive factories, informed by what they’ve already built before.

3. Grit and talent magnetism: Industrial software requires a special kind of stamina. Sid and Peter operate with the resilience of founders who have built under pressure at global scale. And the top-tier engineering talent already gathering around them says everything.

Long story short: Sid and Peter are exactly the founders you want rebuilding the most fundamental layer of manufacturing.

Looking into the future

Factories will not become intelligent through dashboards or simple co-pilots. They will become intelligent when the data foundation is rebuilt, i.e., when every machine, process, and operator shares a common ontology and AI agents can reason across it. 

Aris Machina is unlocking a new paradigm of industrial superintelligence. Are you ready?

We at Earlybird are ready and proud to back Sid and Peter from day one, as they build the intelligence layer for the factories of the future from Sweden to the world.

If you’re curious, please learn more here: Website, LinkedIn, Sifted article.