built here.
shipping everywhere.
The companies of ai:gothenburg. More than thirty-five teams building AI that works, across industries from fintech to robotics to developer tools. This is what AI companies in Gothenburg look like today.
Attacks start with information that is already out there. Forgotten accounts, leaked credentials, and exposed infrastructure sit in the open long before anyone tries to break in. 0&1 Clarity scans open sources, breach data, and underground markets, then ranks what it finds by how much it matters.
A locally run model separates danger from noise and translates findings into plain language with clear, prioritised actions. Data is anonymised before analysis. Built for organisations and individuals where exposure carries real consequences.
Co-founded in 2026 by Magnus Broman and Stefan Wikander.
It speaks in their voice, knows their world and turns attention into action: book, buy, sign up, get an answer. Share it as a link. Embed it on your site. Let people's own AI agents talk to it directly.
Pages are over. Branded Intelligence is here. Built on EU-sovereign infrastructure.
Co-founded in 2025 by Max Ekman and Dawid Dawod.
Your next hire isn't browsing job boards, they're scrolling. Adway puts your open roles in front of them automatically, hyper-targeted social ads, employer branding and full ATS integration end to end.
The AI is the engine: it decides who sees each role and where, reaching people who aren't actively searching. Without it, this is just manual posting and guesswork on reach. Recruitment teams and agencies across Northern Europe fill roles faster through one platform.
Founded in 2017 by Viktor Nord and Richard Sörensson. Viktor went looking for software to connect an ATS to social media, couldn't find it, and found Richard instead.
Most companies have plenty of marketing data but few real insights. Dashboards get built and reports get sent, yet decisions are still made on gut feeling. Agent Zero builds AI agents that plug straight into marketing tools through its own MCP server, so a marketer can ask what worked and why, and get a clear answer back instead of a chart. The agents are built on twenty years in analytics and performance marketing, and alongside them Agent Zero helps companies with measurement and KPIs.
Founded in 2025 by Carl Löjdqvist.
Paid social burns through creative fast, and a produced video costs thousands and takes weeks. Small businesses either run the same asset until performance collapses or skip video altogether. AiBrush replaces the expensive steps of a shoot with generation and delivers finished, on-brand ad videos. The studio remembers the product and the brand, so a new hook, voice or market is a new cut rather than a reshoot.
The work that remains is orchestration: routing each shot to the right model, keeping the brand consistent across generations and assembling it on a timeline. Inference runs on AiBrush's own GPU infrastructure. Built for small businesses and the agencies that serve them.
Founded in 2025 by Tarek Abdellatef.
Data shouldn't wait for you to ask. Most BI tools do: someone builds a dashboard or writes a query. Aida flips it: it monitors your metrics proactively, flags what needs attention, and explains why numbers are moving, before anyone asks.
Talk to it like you would an analyst. It connects to your data warehouse and semantic layer, and delivers insight in Slack, Teams or web. For data teams, product teams and finance.
Co-founded in 2025 by Niklas Woxlin and Emma Woxlin.
Organisations keep solving problems they have already solved because most knowledge never reaches the people and decisions where it can create value. Aipiens captures knowledge from documents, data and experts, structures it and delivers it in the right context. Unlike traditional AI, Aipiens also asks questions, actively uncovering critical knowledge that has never been documented.
It identifies business-critical knowledge gaps and guides teams in closing them where the impact is greatest. The result is an organisation whose expertise continuously grows over time. For engineering, R&D and knowledge-intensive teams.
Founded in 2026 by Amer Catic.
Growing online gets harder every year, channels multiply, algorithms shift, ad budgets leak. BBO turns that chaos into measurable growth across every channel. Fifteen-plus years of results for Swedish companies, specialists across Gothenburg, Stockholm and Barcelona, and the mindset of a team that has survived every algorithm change since 2009.
Founded in 2009 by John Axelsson, building agentic growth from Gothenburg long before it was called AI.
The battery is the most expensive and least understood part of any electric system. Yet most batteries are still managed by static models, generic software, and fixed rules written before they ever shipped.
Cognivity moves AI where the data is, directly onto the battery. The physics-informed AI software layer, with reasoning embedded, diagnoses current and future battery performance, adapts to how each pack ages, and optimises operation over time. Designed for automotive, stationary storage, space and defence applications.
Founded in 2023 by Christian Fleischer.
Trucks drive empty every day while companies pay full price for freight, because matching available capacity with transport demand is still largely handled by phone and email. Empty Miles is a digital marketplace where hauliers publish their return routes and available capacity, while shippers enter their origin, destination, date and load size. The system then finds a suitable truck that is already heading in that direction. The shipper gets a lower transport price, the haulier earns money on a journey that would otherwise be empty, and the entire deal can be agreed digitally on the spot.
Shippers book through the web platform, while drivers manage the transport through the app, with GPS tracking, payment guarantees and ratings built in. AI enables real-time matching and helps optimize pricing based on supply and demand. Empty Miles is built primarily for small and mid-sized hauliers, especially companies operating between one and twenty trucks.
Co-founded by Björn Johansson, Frank Ahlbom and Adrian Berntsson.
Finstig combines structured methodologies with AI to help founders think critically, make better decisions and turn strategy into action. A guided journey, living execution map and smart CRM help founders get things done while turning business and customer data into actionable intelligence. Finstig also supports startup programs with cohort management, measurable progress and pattern visibility across ventures and cohorts.
Beyond the platform, Finstig is building a community for entrepreneurs to connect, exchange knowledge and grow together.
Co-founded in 2025 by Eva Gómez and Renan Borba.
Most finance teams still run the company on spreadsheets and a monthly reporting marathon. Fintower retires it: plug into accounting and ERP, and the AI automates the close, forecasting and reporting, the work a finance team would otherwise grind through by hand.
Founders and CFOs get a live view of cash, runway and margins every day, not once a month. They shipped it live at the hub, no slides.
Founded in 2024 by Salman Eskandari and Ehsan Yazdani.
Teams building mission-critical products in automotive, aerospace, medical and industrial machinery can't adopt generative AI without compromising on product safety. Existing engineering infrastructure was never built to manage and trace agentic workflows.
Forge closes that gap. Built on end-to-end traceability and grounded in the team's engineering context, it is the verification layer that lets engineering teams deploy agentic workflows responsibly, at scale.
Founded in 2025 by Dennis Giegold.
Knowledge lives in people, not documents. NASA built the Saturn V and put people on the Moon. Today they could not build another. The blueprints survived, the know-how left with the engineers. Intended Future builds digital employees that fix this. Your experts teach them the job the way they would a new colleague. Then they do the work.
At its core is MIMIQ, a memory engine that turns expert judgement into explainable software that runs on its own. The first product, Agent Q, starts with design quality. The platform is expert-agnostic and growing into new domains such as materials circularity and defence.
Co-founded in 2021 by Kostas Stylidis and Julia Orlovska.
Most enterprise AI runs on someone else's models in someone else's cloud, which becomes a problem the moment your data is the sensitive kind. Kog takes the repetitive workflow work off people's plates and runs it on a fully generative agent stack, hosted and open-weight models, kept inside the EU.
The agents do the grind; the setup keeps the data and the models where the organisation can actually trust them. For organisations automating internal work.
Founded in 2025 by Anoop Mathew.
Construction projects hide costly risks when requirements and responsibilities do not align across drawings, descriptions and specifications. Konfliktkollen analyses the document package as a whole before tendering, contract signing or construction starts, turning it into a prioritised review list where each issue links back to the exact source.
Comparing requirements across a drawing, a table and a scanned PDF, often worded three different ways, is where AI can do what keyword search cannot. For construction companies, contractors and project teams.
Founded in 2026 by Niclas Lundgren.
Swedish SMEs still juggle separate systems for every part of their finances, each requiring accounting expertise the owner does not have. LedgerLoop replaces the stack with one platform built for Swedish regulations, BAS, SIE, AGI, K2, all of it.
Finn is the AI employee who does the actual work. Forward a supplier invoice, Finn reads it, suggests the right accounts and creates the entry. When a customer's invoice falls overdue, Finn follows up with a reminder. You review and approve. Not an analytics layer on top of existing software, a replacement for the manual work itself.
Founded in 2026 by Faik Riza.
Fresh food is still made almost entirely by hand, and that shapes where and when people can eat well.
Food has long resisted automation because it is organic and variable, and conventional robotics struggles when the real world doesn't match the model. Linneus builds on recent advances in machine learning for robotics to handle that variability.
Founded by Nils Ingelhag, out of robotics research at KTH.
Every engineering team building embedded systems knows the feeling: eighteen months quietly becomes three years. Not because the engineers aren't good, because the process compounds against them. Requirements shift, validation cycles stack, each quality gate finds what the last one missed.
Melvix builds a different structure. AI validates continuously as code is written. Not in a gate at the end, but at every change, every commit. What used to surface in a review cycle surfaces in seconds.
For engineering teams in aerospace, automotive, industrial and consumer electronics who cannot afford to keep absorbing the cost of time.
Founded in 2026 by Mauricio Barahona.
300 million two-wheelers are going electric in India alone. As AI changes how vehicles are financed, insured, serviced and resold, the key question is who owns the connected infrastructure beneath those markets. Nexular scales that infrastructure with large vehicle manufacturers across India, Southeast Asia and Africa.
Nexular turns each electric two-wheeler into a trusted connected vehicle and enables AI agents to understand vehicle health, support financing and insurance decisions, enable service workflows and manage actions around the vehicle. A Swedish-Indian team proving that Gothenburg vehicle DNA scales to the world's fastest-growing EV markets.
Founded by Joel Rozada, Emma Rozada and Mathias Hultén.
When AI agents work together they burn tokens repeating context and still talk past each other. Noetive sits in the middle as the broker: a semantic layer, built on LLMs and its own neural net, that lets agents coordinate on shared meaning instead of raw text, and cuts token usage along the way. For teams building multi-agent AI systems.
Founded in 2026 by Gustav Svalander.
Sweden is rebuilding its total defence, and every organisation is now expected to understand how well it can keep critical operations running. Few have a structured way to find out. ORDNAD makes that assessment something an organisation can actually complete, mapping its capabilities, dependencies and vulnerabilities, and surfacing the weak points that would otherwise go unnoticed.
Where expert consultants take weeks, ORDNAD delivers a full capability assessment in a fraction of the time, ending in a prioritised plan for where to act first. The method is established; the speed and intelligence are new. Built for every preparedness sector, not just IT security.
Founded in 2026 by Anton Nyberg.
It helps close the industrial skills gap and significantly reduces the cost of workforce training in manufacturing. Language models are good at conversation but bad at being reliably right, which rules them out wherever a wrong instruction can injure a worker or damage a machine. Ordström AI, a spin-off from the University of Gothenburg, replaces the LLM at runtime with its own dialogue engine, trained on real human conversational patterns, so the AI stays inside what it actually knows.
The first application is an instructor for onboarding and training in manufacturing, where experienced staff today leave their stations to train new colleagues. The engine comes out of more than twenty years of dialogue-systems research by Staffan Larsson.
Co-founded in 2026 by Staffan Larsson, PhD Computational Linguistics, and serial entrepreneur Dennis Juul Poulsen.
Most of Europe's charging operators run at a loss. They manage large networks of chargers across many locations, but operate without the insight into the market or the tools to act on it.
Pickel is an AI platform that gives charging operators real-time data on price and utilization for every charger on the market. It automates pricing and power management in response to local market changes such as demand, volatile energy prices, and the competitive landscape. Built for charging operators across the Nordics and Northern Europe.
Founded in 2025 by Hugo Cliffordson, Christoffer Segerdahl, Gabriel Rosengren and Melker Veltman.
Swedish SMEs face mounting regulatory pressure from GDPR, NIS2, ESG and tightening customer requirements, but rarely have internal compliance capacity. Polaris replaces fragmented one-off efforts with a continuous system: it assesses the current state, builds the right documentation and structure, and keeps the work alive with ongoing monitoring and updates.
The platform adapts to company size and industry, suggesting what to prioritise and when to act. Not a consultant engagement that ends with a report, a system that keeps compliance current as regulations change.
Co-founded by Johan Gerebro and Olle Nordberg.
Half the numbers and datapoints in a CRM are dead. RebelOS finds the decision-maker in official records, with registry IDs, industry codes, revenue and margin rather than last year's scraped profiles, and reveals a mobile or email that is verified on request and rechecked every 120 days.
Built for the people who sell: search, list, reveal, push to the CRM. Since Rebel is available over MCP and API, a rep can describe the companies and people they want in Claude, ChatGPT or Cursor and get a saved, enriched list back in one turn. Used by teams at Consid, Reliable, Nespresso, Garba and Videotech.
Founded in 2026 by Timmy Lundin and Robin White.
Regweaver is a European RegTech company building Automated Compliance Operations for companies managing multiple EU regulations across complex value chains. RegCheck identifies which regulations apply. RegWatch monitors regulatory change. Regweaver OS turns obligations into structured workflows, assessments, requests, responses and evidence across customers, suppliers and partners.
The same company can be a reporting entity, answer a customer's regulatory requests and request information from its own suppliers at the same time. Regweaver connects those relationships in one traceable compliance model spanning CSRD, VSME, CSDDD, NIS2, CRA, the AI Act, SFDR and EU Taxonomy. Agentic AI interprets regulatory change and supports execution, on a deterministic compliance model with control, traceability and audit trails built in.
Co-founded in 2025 by Daniel Rundgren.
Designing a hardware product is slow, error-prone and expensive: a typical cycle from concept to verified layout takes a professional a month or two. Root-E turns that into a conversation. Describe what the product should do, answer a few follow-up questions, and the AI generates the schematic, 3D design and electronics board, ready to send to a manufacturer.
Off-the-shelf models do not know product design with electronics, so the team has built a harness around them trained on circuit knowledge, professional routing, spacing and EMC rules, and is extending it toward multilayer boards. Built for electronics product companies and their engineers first, with the ambition that anyone with an idea can get a working board.
Co-founded by Chaitanya Gore, Dhananjay Yadav, Ripudaman and Vivek Singh.
Companies buy enterprise systems top-down and wonder why nothing works. Sefira starts from the other end: one process at a time, automated with AI where it fits and conventional automation where it doesn't. Each process is proven, deployed and running before the next one is built. What took a department months of spreadsheet work runs in hours.
The platform connects users, processes, requirements and systems in a unified architecture. The goal: companies where every operation runs autonomously, with humans governing decisions, not grinding through tasks.
Co-founded in 2026 by Mikael Tjärnhage and Alexander Mlakar.
Every large EPC project generates thousands of supplier documents and drawings, each reviewed by hand against project specifications, technical requisitions and standards such as NORSOK, ISO, IEC and API. Documentation typically goes through five to ten review cycles before approval. Simlogic's platform finds compliance gaps, missing information, inconsistencies and deviations, with every finding traceable back to the requirement behind it. Reading a requirement in a specification and finding where a datasheet or drawing contradicts it is where AI does what keyword search cannot. The first release covers technical and supplier documents, with drawings such as P&IDs and single-line diagrams to follow, starting with North Sea offshore projects and built to extend to nuclear, offshore wind, hydrogen and carbon capture.
Simlogic began as Norwegian startup Skillup, moved to Gothenburg in 2024 and still runs SkillupVR, a VR simulation platform for medical team training sold to several institutions across Europe and Asia.
Founded in 2024 by Ole J Olsen.
Most mobile robots are effectively blind: they follow fixed routes, need constant supervision, and stall when the world changes. Stær is the spatial intelligence layer that fixes it, semantic 3D mapping, multi-vendor coordination and continuous learning, so a fleet understands its space and runs without a babysitter on the floor. One robot learns the room and the whole fleet's map updates at once.
The first application is the warehouse: every pallet located in a live 3D map, blocked aisles and misplaced stock flagged, picks per hour lifted with data instead of guesswork. Built by computer-vision experts from Apple, Mapillary and Meta. For operators running mobile robot fleets.
Founded in 2025 by Jan Erik Solem.
Most SMEs run on spreadsheets and a monthly reporting marathon, and the analysis that matters waits for a CFO they cannot afford. Stift gives them that intelligence without the headcount: it connects banks, accounting and payments into one platform that reads everything in real time, spots cash leaks and duplicate invoices, and answers any financial question in plain language.
Traditional software shows what happened; you still need an analyst to interpret it. Stift's AI replaces that judgment layer. Software automates workflows, this automates judgment. For SMEs and scale-ups.
Founded in 2026 by Emil Ödman.
Fraud moves at AI speed. The decisions meant to stop it still run on static, disconnected records. LENS is the layer that closes the gap: it resolves entities across any source, maps the relationships between them, and surfaces the networks risk hides in, shell structures, hidden owners, the tie between a clean-looking applicant and one that isn't.
An operating layer for decisions, with regulated data access built in. The AI detects; you decide.
Founded in 2024 by Jens Nylander, with Marcus Thorelius as co-builder and investor.
Most platforms rank: the popular get all the attention, the rest get none, and the list never asks what the other side wants. TR8S matches instead. Both sides say what they look for, and a stable-matching engine finds the pairs where neither side would rather have someone else. The theory won the Nobel Prize in Economics in 2012. Tim has built on it for almost ten years: dating, conferences, workshops.
Bekanta (events, launched October 2026), TrackEid (tree-ring data with the University of Gothenburg), Pejl (a weather app for relationships) and Eniga (Swedish law with sources). Not raising. Looking for organizers who want to run Bekanta at their next event.
Founded in 2026 by Tim Brandin.
Insurance distribution still runs on legacy systems stitched together by hand, policy administration in one place, claims in another, and partners managed over email. Verda's platform, VerdaCORE, collapses the stack into a single ecosystem built on AI: policy administration, claims, billing, CRM and partner management.
AI runs across the platform for risk analysis, fraud detection and claims triage, operational workflows that insurers and brokers still perform manually today. Built for insurers, brokers, MGAs, distributors and affinity partners.
Founded in 2024 by Niklas Dahlgren.
Voidleap is a Swedish software company that builds tools for agentic software development. AI should make skilled people more capable, not replace their judgment or turn their work into someone else's data asset.
Voidleap Code is designed around privacy, transparency and practical control. The model is becoming the commodity, the workflow is becoming the advantage. Create provider-independent workflows and own your development process. Built for developers and organisations that want to move fast and secure with AI, without giving up control of their data or development process.
Co-founded in 2025 by Conny Wallström and Daniel Nilebäck.
Recruitment still runs on CV signals, undocumented "culture fit" calls, and processes that leave most candidates without feedback. Wickla sits between employers and candidates as a neutral layer: AI conducts structured interviews at scale, candidates build portable profiles assessed on competencies and soft skills, employers receive shortlists with narrative summaries and documented decision trails.
The EU AI Act classifies recruitment AI as a high-risk system, rolling out compliance from August 2026. Wickla is built around those requirements from the start, human oversight, explainability, full audit trails. For HR teams, recruiters and headhunters at organisations with high hiring volume.
Founded in 2026 by Petr Chadraba.
building an AI company in Gothenburg?
The hub is open to AI founders building from Gothenburg. A desk here puts you next to the teams above. Apply, get a desk, ship.
apply as a founder