- Arrakis has exited stealth with $38 mn in funding, including a $30 mn Series A led by Blossom Capital, to build agentic AI software for industrial companies.
- The startup targets aerospace, energy, logistics and manufacturing, arguing that most AI investment has focused on office workers while larger ROI sits in industrial operations.
- Arrakis faces competition from Palantir, consulting firms and well-funded industrial AI startups, but says its advantage is narrow operational deployment, model-agnostic design and performance-linked pricing.
Arrakis, a seven-month-old London and Paris startup building an AI operating system for industrial companies, has come out of stealth with $38 mn in venture funding.
The company says it wants to bring agentic AI into aerospace, energy, logistics and manufacturing, sectors where software work sits close to physical operations, machinery and supply chains.
The new round values Arrakis at $140 mn post-money, cofounder and CEO Rafael Quintanilla.
Quintanilla previously worked as a vice president at Accel. During his time there, he spent much of a year travelling across the US, Europe and the Middle East while developing the firm’s thesis on defence and industrial durability. What he saw pushed him out of venture capital and into company-building.
He said he found a wide gap between AI investment activity at Accel and in Silicon Valley, including companies such as Anthropic and Lovable in Europe, and the needs he saw across more industrial parts of the economy.
Most AI software has targeted knowledge workers whose jobs run through screens and applications. Most AI investment to date has targeted the 30% of workers behind a desk. The real ROI lies in the 70% running industrial operations.
Sonali de Rycker, the Accel partner who backed Arrakis at seed stage, said she is backing the founder as much as the market. She described Quintanilla as unusually curious, persistent and driven, adding that she worked closely with him during his Accel years and now backs him again as an entrepreneur.
Arrakis enters a crowded industrial AI race. Accenture and Boston Consulting Group are moving aggressively into the sector. Palantir already sells operational software to large industrial and government clients.
Prometheus, backed by Jeff Bezos and now valued in the tens of bn of dollars, is putting capital into automating physical product engineering. Frontier AI labs are also watching the same market.
Arrakis occupies a different lane. If Prometheus worked with Airbus it would build AI for the engineering work behind an aircraft.
Arrakis would handle the operational work around it. The company wants to become the AI layer for industrial operations rather than the system designing the product itself.
Quintanilla said he respects Palantir and noted that about half of his team comes from there, including a former head of Palantir’s procurement and supply-chain team. His critique sits with the product architecture and pricing. Palantir, he said, is a 20-year-old company with a heavy price point and technology starting to look legacy.
Consulting firms face a different issue, according to Quintanilla. Their model still depends on consultant hours or outsourced labour, even as AI begins to reshape the business. He said consultants solve strategy problems with people as the main enabler, while Arrakis wants to solve company problems through software, with people still involved where needed.
Arrakis engineers rebuilt the spreadsheet operators already used, then had AI populate the data while the system learned from operator corrections and codified their working knowledge.
The company’s rollout model starts at headquarters, proves value, then moves into field operations once internal demand appears. Arrakis usually links about half of its fees to hitting a specific performance target.
Winning over conservative European industrial companies brings its own friction. Quintanilla said his strongest traction has come from family-controlled businesses. He described it as an open secret. Those owners think long term, push top-down initiatives through the organisation and allow relationships less driven by single transactions.
Arrakis remains model-agnostic by design. Executives want to avoid dependence on one AI model provider and worry about high token costs.
The executive wanted a system able to route work to the best provider instead of locking the business into one stack.
