Dave Clark ran Amazon's worldwide operations machine, then had a short and messy turn at Flexport. Now Auger has $150 million behind a harder bet: AI agents that do real supply chain work, not just talk about it.
Auger just closed a $50 million Series B. The Bellevue, Washington startup, founded by Clark in 2024, landed the round from Eclipse, with Oak HC/FT returning for a second round. That brings the company's total funding to $150 million in under two years, according to GeekWire, which reported the deal on July 9. That's fast. For a startup barely 18 months removed from its Series A, it's a quick pace of capital even by AI-era standards.
You've heard plenty about AI agents that draft emails, summarize meetings, and sit politely inside chat windows. Auger is chasing something less showy and much harder: software that can make the thousands of small operational calls that keep a warehouse, a truck fleet, and a demand forecast from drifting apart.
That is real work.
The product sits on top of a company's existing ERP, warehouse management, transportation management, and demand planning systems rather than replacing them. It pulls that data into one layer, then uses AI agents alongside more traditional optimization models to make routine execution decisions and carry them out. When something falls outside the pattern, it kicks the exception back to a person. Automate the routine, escalate the strange cases, leave the humans for judgment calls. That's the pitch.
Auger already counts Meta's Reality Labs division, sports merchandise giant Fanatics, and consumer products maker Kimberly-Clark as customers, and Clark told GeekWire that another eight to ten companies are in pilots or contract negotiations. The company has grown to roughly 130 employees. Those are production relationships, not press-release logos, and you should pay attention to that distinction because supply chain software has a brutal habit of dying in pilot purgatory.
The résumé behind the pitch
Clark's own case for why now is worth taking seriously starts with the fact that he's one of the few people alive who's run this at true scale. He spent 23 years at Amazon, eventually leading its worldwide operations and then its entire worldwide consumer business, the unit responsible for essentially everything that gets ordered, stocked, and shipped. He left in 2022 and took the CEO job at Flexport. That lasted less than a year.
According to Fortune's reporting on Auger's $100 million Series A in October 2024, Clark moved back to the Seattle area, partly to draw on the region's logistics and cloud talent, and launched Auger with his wife, Leigh Anne Clark, who runs the company's fashion and beauty division as co-founder and president.
His argument, as he's put it, is that general-purpose AI wasn't built for the level of expertise supply chain execution actually requires. That's a pointed line. It is aimed squarely at the wave of startups slapping a chat interface on a logistics dashboard and calling it transformation. Auger isn't selling insight. It's selling execution.
This is not a chatbot story
Here's the thing that makes this round more interesting than the average AI-agent raise. Investors have chased agent pitches relentlessly over the past year, and too many of them are thin: a wrapper around a language model, aimed at a task nobody's proven agents can do reliably yet. Supply chain execution is different. It has decades of structured data, clear rules, and a steep cost of error. If an agent is ever going to earn the right to act rather than suggest, this is one of the places where it has to prove itself.
Eclipse and Oak HC/FT aren't betting on another assistant that tells you what a dashboard already knows. They're betting on Clark's Amazon years. He knows what a routine warehouse decision actually looks like, and which decisions still need a person in the loop. That distinction matters: enterprise buyers don't hand operational authority to software because the demo looks clever. They do it when the software can take work off the floor without creating new messes somewhere else.
It is also a bet on Clark himself, and that cuts both ways. His Flexport tenure ended fast, and operators with big résumés don't always translate that experience into a startup that survives its own growing pains. Auger's early customer list, spanning retail, consumer goods, and Meta's hardware division, suggests the product is landing outside Clark's old Amazon orbit. That is the harder test for any operator turned founder: proving the playbook works somewhere your old badge doesn't open doors.
The next stretch will tell the real story. Eight to ten pilots converting into paying contracts would validate the model quickly. If they stall instead, Auger becomes another reminder that enterprises like automation in theory, then slow down when software asks for permission to make real decisions.
Also read: OpenAI Researcher Miles Wang Wants $200 Million to Bet on Failed Drugs; Hinge's Founder Just Bet $18 Million That Swiping Is Broken; The Market Has Priced In Africa's Risk But Not Its Trillion Dollar Upside
