
CARUPI is the world's 1st managed P2P marketplace for used cars
Turn this teardown into a decision-ready prompt for ChatGPT, Claude, or your agent.
If you only have a few minutes to spare, here’s what investors, operators, and founders should know about CARUPI (W20).
Carupi tried to bring trust to Brazil's chaotic used-car market, and ran into a familiar Latin American trap: an under-capitalized, operations-heavy marketplace in a low-frequency category, fighting a capital war against a mega-funded incumbent, just as the funding winter arrived. Founded in 2019 by Diego Fischer and part of Y Combinator's Winter 2020 batch, Carupi billed itself as the world's first managed peer-to-peer marketplace for used cars — connecting private buyers and sellers while managing inspection, paperwork, and payments to make risky informal car sales safe.[1]
It gained early traction, raising over $2 million (a $125K YC seed plus Streamlined Ventures, Liquid2, and K50 Ventures) and scaling to around 200 employees across ten Brazilian cities at its peak.[3] But records indicate Carupi ceased operations around October 2022.[4] The structural problem: used-car transactions are extremely infrequent (weak retention, high acquisition cost), the managed service is operations-heavy on thin margins, and Carupi's ~$2 million was tiny against the billions flowing to used-car giants like Kavak in the same market.
Diego Fischer founded Carupi in 2019 to fix a genuine, painful problem in Brazil.[6] The Brazilian used-car market is enormous but plagued by distrust and fraud — buying a car from a stranger risks hidden mechanical problems, title issues, and scams, while selling one means dealing with unreliable buyers and complex paperwork. This friction pushed many people toward dealers who took large margins, or toward risky informal transactions. Carupi's "managed P2P" model aimed to give private sales the safety of a trusted intermediary: it would inspect the car, verify the paperwork, hold funds in escrow, and manage the transaction, so buyers and sellers could transact directly with confidence.
The idea was compelling enough to earn a spot in Y Combinator and early investment, and Carupi scaled quickly, reaching roughly 200 employees and ten cities.[2] But the model carried heavy structural costs. Adding trust to a used-car sale means real operational work per transaction — inspections, paperwork, escrow, logistics — which is labor-intensive and hard to scale profitably. And the transactions themselves are rare: a person buys a car every several years, so there's no repeat-purchase flywheel to lower the cost of acquiring each customer. Carupi was building an expensive, high-touch service for an infrequent purchase, in a market where a far-better-funded competitor was already spending heavily.
Carupi was a managed marketplace for private used-car sales. A seller listed their car, and Carupi handled the parts that make P2P car sales scary: inspecting the vehicle, verifying its history and paperwork, facilitating financing where needed, holding payment in escrow, and managing the transfer of ownership.[1] For buyers, this meant confidence that the car was as described and the transaction safe; for sellers, it meant a trustworthy buyer and less paperwork hassle.
The trust layer was genuinely valuable in a fraud-prone market, but it was expensive to provide.[5] Each transaction required inspections, document verification, and coordination — operational work that scaled with volume rather than amortizing. Carupi's ~200-person headcount reflected this labor intensity. Meanwhile the revenue per transaction was constrained by the thin margins of used-car sales and by competition. Building a high-touch, operations-heavy service for an infrequent, price-sensitive purchase is a difficult economic proposition, and it requires either exceptional efficiency or deep capital to reach the scale where it might work.
Read the complete post-mortem, the rebuild playbook, and the exact reasons CARUPI is still worth studying now.