Loading organizations...

§ Private Profile · San Francisco, CA, USA
AI Broker for Tariff Refunds
Pax has raised $5.0M across 2 funding rounds.
Key people at Pax.
Pax was founded in 2024 by Penny Chen (Founder & CEO) and Chris Le (Founder).
Pax has raised $5.0M in total across 2 funding rounds.
Pax uses AI to automate tariff refunds (aka duty drawback) for retailers and manufacturers. We launched 4 weeks ago and have ~$200K in contracts.
Each year, 80% of eligible refunds:equivalent to $15B:go unclaimed. Pax is the first AI-powered broker helping brands under $50M reclaim 3-5% of their revenue. Our algorithms generate 15% more refunds than the industry leader and reduce processing time by 99% with AI.
We are two technical founders: Penny got her PhD from MIT and, and was a former research scientist at Amazon & Flexport, where I encountered the problem firsthand. Chris is a second-time founder and former software engineer at Amazon, Brex, and TikTok.
Key people at Pax.
Pax was founded in 2024 by Penny Chen (Founder & CEO) and Chris Le (Founder).
Pax has raised $5.0M in total across 2 funding rounds.
Pax's investors include Y Combinator.
Pax (Pax AI) is an AI-powered broker that automates U.S. duty drawback claims, helping importers reclaim import tariffs and duties they’ve overpaid. The company’s platform turns the complex, manual customs refund process into a fast, self-serve experience—like “TurboTax for duty drawback”—using AI and optimization algorithms to collect, validate, and submit claims while maximizing refund amounts. Pax serves small and mid-sized businesses (SMBs) and mid-market brands, especially in e-commerce and manufacturing, that previously couldn’t access drawback due to high costs and complexity.
Pax solves a massive, underutilized opportunity: U.S. companies pay over $100B annually in import duties, but an estimated $10B in eligible drawback refunds go unclaimed each year. By automating the end-to-end process, Pax reduces processing time by ~99% and typically generates 15% higher refunds than incumbent software. The company has already crossed $1M in annualized booking revenue with a small team, demonstrating strong early product-market fit and traction in a niche but high-value regulatory domain.
---
Pax was founded by Penny Chen and Chris Le (often referred to as Chris Lee in public materials), who met on Y Combinator’s co-founder matching platform. Penny, a research scientist with a PhD from MIT and prior experience at Amazon and Flexport, worked on supply chain and logistics algorithms. During her time at Flexport, she gained firsthand exposure to duty drawback and recognized that while the program offered substantial refunds, it was only accessible to large enterprises due to its complexity and cost.
Chris, a software engineer with experience at Amazon, Brex, and TikTok, brought deep expertise in building financial and supply chain systems. Together, they saw an opportunity to use AI and software to automate the drawback process, making it affordable and accessible to SMBs. The idea crystallized around building a fully automated, algorithm-driven platform that could handle data extraction, matching, optimization, and claim submission—without requiring manual intervention or large teams.
Pax launched as a Y Combinator–backed startup and quickly gained momentum, raising a $4.5M seed round led by Initialized Capital. Within less than a year of fundraising, the seven-person team hit $1M in booking revenue, validating the demand for an AI-first approach to tariff refunds.
---
---
Pax sits at the intersection of three powerful trends: the resurgence of trade protectionism, the digitization of global trade, and the rise of AI in vertical SaaS. With ongoing trade tensions and tariffs like Section 301 (25–100% duties on Chinese goods), import costs have surged, making duty drawback more valuable than ever. At the same time, legacy customs and logistics systems remain highly manual and fragmented, creating a ripe opportunity for automation.
Pax is part of a new wave of “AI brokers”—vertical AI applications that automate complex, rules-heavy, high-value processes in regulated domains (like customs, tax, compliance). By making drawback accessible to SMBs, Pax is democratizing a financial benefit that was previously reserved for multinationals, effectively redistributing value back to smaller players in global trade.
The company also reflects a broader shift in B2B software: from generic tools to deeply specialized, domain-specific platforms that combine regulatory expertise with modern AI/ML. As more companies look to optimize supply chain costs and improve cash flow, Pax’s model could expand beyond drawback into related customs and trade finance use cases, influencing how businesses interact with government programs and cross-border regulations.
---
Pax is well-positioned to become the default platform for tariff refunds in the U.S., especially as trade barriers persist and SMBs seek every dollar of margin. In the near term, the company is likely to deepen its footprint among e-commerce brands and manufacturers while expanding its channel partnerships with brokers and logistics providers. Over time, Pax could evolve into a broader “trade optimization layer,” layering on capabilities like bonded warehouse management, foreign trade zone (FTZ) planning, and even export incentives.
The long-term vision is clear: turn Pax into the central AI engine for global trade cost recovery, where any company with a global footprint can automatically identify and claim every dollar they’re owed across tariffs, duties, and incentives. As AI becomes more embedded in enterprise workflows and regulatory tech matures, Pax’s blend of deep domain expertise and algorithmic optimization could set a new standard for how companies interact with customs and tax authorities.
Just as TurboTax made tax filing accessible to individuals, Pax is doing the same for a critical but overlooked piece of global trade finance—turning a bureaucratic headache into a source of effortless, automated cash flow.
Pax has raised $5.0M across 2 funding rounds. Most recently, it raised $4.5M Seed in April 2025.
| Date | Round | Lead Investors | Other Investors | Status |
|---|---|---|---|---|
| Apr 9, 2025 | $4.5M Seed | — | — | Announced |
| Sep 1, 2024 | $500K Seed | — | Y Combinator | Announced |