Funding Brief｜Magentic Raises $18M Series A for Procurement Agents

🌏 中文版

Funding Details

What the company does

Magentic automates procurement for manufacturers — it hands the entire buy-or-build decision, supplier selection, contract negotiation, order placement, and invoice reconciliation loop to an AI agent called Mages, rather than just assisting a human through each step.

Mages plugs directly into a manufacturer's email, Microsoft Teams, and internal systems, functioning more like a team member than a piece of software. It handles operations spanning billions of data rows, tens of billions of dollars in spend, and legacy ERPs that still run on Excel. One manufacturer uses it to process more than a million purchase orders a year; another used it to uncover $4 million in savings its human team had missed.

Co-founders Robin Van Aeken (CEO, a former McKinsey supply-chain consultant) and Odhran O'Donoghue (CTO, a machine-learning PhD with stints at OpenAI, NASA, and the Crick Institute) met at Oxford hackathons and reunited seven years later around a problem McKinsey itself had sized: supplier-compliance "leaks" cost enterprises roughly 2% of total spend — about $40 million on a $2 billion budget.

What this round signals

What it means for the agent ecosystem

Procurement automation is already a contested field: Estonia's Pactum, focused on automating supplier negotiations, raised a $54M Series C in mid-2025 (over $100M total); New York's Freehand raised $75M and already runs agents at Meta, Unilever, Johnson & Johnson, and Pfizer; Didero, founded by another former McKinsey procurement expert, raised $30M less than a year ago. What sets Magentic apart is its aim to own the entire process — from the buy-or-build decision through direct and indirect spend to invoice reconciliation — rather than a single slice of it. That's a sign investors have stopped asking whether procurement can be automated and started betting on who can stitch the whole chain together instead of just one link.

What investors are betting on

Felicis partner Feyza Haskaraman put it directly: "Supply chains are the least glamorous part of the economy, yet the most consequential, deciding what gets built and what does not. That's also what makes them so hard to automate. Getting an agent to understand a manufacturer's complex systems well enough to take action inside them is no small feat, which is why we haven't seen anyone else build autonomous AI workers for the physical economy." Felicis has backed over 50 unicorns including Shopify, Canva, and Notion; its logic for Magentic is that as AI-driven demand fuels the largest physical capex cycle in history, whoever embeds intelligence into every procurement decision compounds a durable edge.

Numbers worth watching

• Total funding went from a $5.5M seed to $23.5M in 14 months; seed-round angels included people from SAP, Airbus, McKinsey, and Hugging Face, an early signal of industry-insider conviction
• One customer already processes over a million purchase orders a year through the platform, indicating production-scale usage rather than a proof of concept
• The global procurement software market was about $10.1B in 2025 and is projected to reach $21.3B by 2033 — a modest market on its own, but Goldman Sachs estimates roughly $8 trillion in AI-related capex by 2031, most of it going into physical infrastructure that all has to pass through procurement first

Watchlist status

Magentic is not yet tracked in the watchlist. Recommend adding it under section D9 (Industrial / Manufacturing / Supply Chain), with tracking focus on: end-to-end procurement agents for manufacturers, an $18M Series A led by Felicis, and 3x funding growth within 14 months.

Today's takeaway

The "agents replacing human work" narrative has mostly played out in digitally native jobs so far — writing code, sending sales emails. Magentic is a reminder that the most tedious, least-wanted work in the physical economy — supplier compliance, invoice reconciliation, Excel-based ERPs — may be exactly where agents gain their first real foothold, because the pain points are concrete, measurable, and chronically under-resourced by human teams.

References

• London's Magentic raises $18M to build AI agents that diagnose, plan, and act across terabytes of procurement data
• Magentic raises $18M to put AI agents to work in manufacturers' supply chains