TL;DR
Arrakis raised $38M led by Blossom Capital and Accel to build an AI operating system for industrial sectors like aerospace and logistics
Arrakis, a London and Paris startup building what it calls an AI operating system for industrial companies, has emerged from stealth with $38 million in total funding and a thesis that the biggest returns from artificial intelligence will come from factories and supply chains, not office software. The company raised a $30 million Series A led by Blossom Capital, on top of a seven and a half million dollar seed round led by Accel that closed in March. The Series A values Arrakis at $140 million post-money, according to Fortune, which first reported the round.
CEO Rafael Quintanilla, a former vice president at Accel, co-founded Arrakis in January 2026 with Haroun Beltaifa and Romain Fouilland, both formerly of Palantir, and Mikhail Galkov, previously of Delivery Hero. The founding team’s Palantir pedigree is deliberate, Quintanilla has said the company sees Palantir as a legacy player ripe for disruption. Arrakis positions itself as a faster, more flexible alternative to Palantir’s entrenched government and enterprise contracts, as well as to traditional consulting firms that charge for headcount rather than outcomes.
The startup deploys what it describes as forward-deployed AI engineers who embed directly at customer sites in sectors like aerospace, energy, logistics, and manufacturing. Rather than locking customers into a single model provider, Arrakis takes a model-agnostic approach, starting with commercial models from OpenAI and Anthropic before shifting workloads to open-source alternatives from vendors like Mistral, which recently acquired industrial AI firm Emmi to strengthen its own pitch to manufacturers. The company claims this flexibility delivers a two-to-four-fold improvement in output quality while cutting token costs by roughly 70 percent.
Arrakis says it already has five paying customers, including NYSE-listed enterprises, and that one client cut procurement cycle times by 90 percent using its platform. The company ties roughly half its fees to performance targets, an unusual structure in enterprise software that Quintanilla has framed as proof the system delivers measurable results. The startup says it has found its best traction with family-controlled businesses, which tend to make faster purchasing decisions than publicly traded corporations with layered approval processes.
The angel investor list signals credibility in both AI and industrial circles. Datadog CEO Olivier Pomel, OpenAI head of business products Olivier Godement, and Junaid Hussein, founder of Cambridge Aerospace, all participated in the round. Additional backing came from GFC, MainObject, and Rerail.
Arrakis enters a crowded but fast-growing market for industrial AI. PhysicsX raised $300 million at nearly two and a half billion dollars in valuation in June for AI-powered engineering simulation, while Jeff Bezos-backed Prometheus is pursuing a similar thesis in heavy industry. The difference, Quintanilla argues, is speed, Arrakis can deploy within weeks rather than the months or years that legacy vendors require.
The company plans to triple its headcount from about 15 employees and open offices in New York and the Middle East, using the fresh capital to expand its presence in markets where industrial transformation is a government priority. Whether Arrakis can scale that forward-deployed model without ballooning costs will be the central test of whether its approach works beyond its first handful of clients.
