AI is moving beyond software development and office workflows into highly specialized engineering disciplines. Endra, an AI platform for mechanical, electrical and plumbing (MEP) engineering, has launched Endra Power Studio, an agentic platform designed to automate electrical engineering workflows, while acquiring Swiss AI lab Planlabs to accelerate its expansion into mechanical engineering.
The construction industry faces a constraint that has little to do with concrete, steel or equipment: there are not enough engineers to design everything the world wants to build.
Data centers, electrification projects and large commercial developments are increasing demand for MEP engineering, while electrical and mechanical design remains spread across specialized tools and heavily dependent on experienced engineers.
Endra is attempting to address that bottleneck with agentic AI for engineering.
The company has unveiled Endra Power Studio, which it describes as an agentic platform purpose-built for electrical engineering. At the same time, Endra announced the acquisition of Planlabs, a Swiss AI lab focused on mechanical engineering.
The combined strategy is to move toward a unified AI platform capable of handling multiple MEP disciplines.
From Engineering Tools to Engineering Agents
Electrical engineering projects typically involve multiple stages, including load calculations, equipment placement, circuiting, cable routing, voltage-drop calculations, coordination and documentation.
Those activities can span applications such as Autodesk Revit, Navisworks and spreadsheets, leaving engineers responsible for moving information between different parts of the workflow.
Endra Power Studio is designed to bring those activities into a single environment.
Engineers load a building model, define project requirements and delegate specific tasks to AI agents. The engineer then reviews the results and signs off at each stage.
The platform includes Playbooks, which allow engineering firms to encode their own standards, calculation methods, unit-placement rules, circuiting practices and approaches to code interpretation.
That creates a potentially important distinction from generic AI assistants. Instead of asking a general-purpose model to produce an engineering answer, firms can establish a controlled set of organizational rules that agents use while carrying out defined workflows.
Endra says its agents can handle tasks such as developing routing strategies, placing electrical units across rooms and updating designs when architectural models change.
AI Meets Physics-Based Engineering
The platform also incorporates a simulation and physics engine intended to validate designs before construction.
According to Endra, the system can check electrical designs against physical constraints and building-code requirements, including cable sizing and voltage-drop calculations.
That combination of generative AI and deterministic engineering systems is significant.
Large language models are effective at interpreting instructions and generating information, but engineering applications require calculations, constraints and validation that cannot safely depend on language generation alone. Endra’s approach therefore places AI agents around specialized engineering engines rather than treating an LLM as the complete engineering system.
Another feature, Type Rooms, allows engineers to design a room once and instruct an agent to apply the design across matching rooms in a building model. The company is initially targeting use cases including healthcare, multifamily residential and hotel projects.
The platform’s document engine then connects engineering outputs with project documentation, including Revit models, single-line diagrams, shop drawings, panelboard schedules and bills of materials.
Building Toward Multidisciplinary MEP AI
The Planlabs acquisition expands the company’s ambitions beyond electrical engineering.
Founded in Switzerland in 2024, Planlabs developed geometry, data and physics engines for automatically designing mechanical systems while accounting for physical constraints and collision avoidance.
The five-person team, comprising PhDs in mathematics and computer science, will join Endra. Mechanical capabilities are expected to become part of the Endra platform after the technology is integrated.
Endra says a plumbing module is planned for 2027, creating a roadmap spanning electrical, mechanical and eventually plumbing engineering.
The strategy reflects a broader movement toward vertical AI systems that combine foundation-model capabilities with proprietary workflows, domain-specific simulation and structured enterprise data.
Enterprise Control and Data Governance
Engineering firms adopting AI also face a different set of requirements from organizations deploying general-purpose productivity assistants.
Design decisions can affect safety, compliance, construction costs and project schedules. Endra therefore emphasizes human review throughout its workflows and says its models are not trained on customer data.
The company says its platform is SOC 2 Type II and ISO 27001 certified, has undergone Cyber Essentials Plus assessment and is GDPR compliant, with regional data hosting.
Those controls are particularly relevant as AI moves deeper into professional engineering workflows where project information can include sensitive building designs and infrastructure specifications.
Endra says it already works with enterprise consulting and engineering firms including AtkinsRéalis, Ramboll, Buro Happold, Hoare Lea and AFRY.
The company previously raised a $50 million Series A led by Andreessen Horowitz, with earlier backing from Notion Capital and Norrsken VC.
Endra is now expanding beyond Stockholm into New York, San Francisco and London.
The larger opportunity is not simply making engineers faster. It is turning accumulated engineering expertise into software that can be reused across projects while leaving professional engineers responsible for requirements, judgment and final approval.
If Endra’s approach scales, agentic AI could become less about replacing engineering software and more about connecting the fragmented systems, calculations and institutional knowledge that engineers already use.
Market Landscape
The AEC industry is becoming an increasingly important target for vertical AI and engineering automation.
Traditional building-information-modeling platforms remain central to project design, while AI companies are adding automated design generation, model analysis, simulation and workflow orchestration around those systems.
The technical challenge is greater than deploying a general-purpose AI assistant. Engineering systems need deterministic calculations, physical constraints, regulatory requirements, traceability and human approval.
This is creating an emerging category of AI-native engineering software, where agents coordinate specialized computational engines rather than independently generating final designs.
For engineering firms, the potential value is concentrated in repetitive design work, documentation and coordination. The longer-term question is how much engineering expertise can be encoded into software without weakening professional oversight or accountability.
Top Insights
- Endra Power Studio applies agentic AI to electrical engineering workflows including design, routing, simulation and documentation.
- Its Playbooks allow firms to encode internal engineering standards and apply them consistently across AI-driven workflows.
- Physics-based validation complements AI agents by checking calculations and engineering constraints rather than relying solely on generative models.
- Planlabs brings geometry, data and physics capabilities for mechanical engineering into Endra’s broader MEP platform roadmap.
- The strategy illustrates how vertical AI is moving into engineering domains where specialized computation, compliance and human oversight are essential.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
