Published by
Published on
July 16, 2026
How Amazon Brazil built a distribution network across one of the world's most complex markets — and how Optilogic is powering what comes next.
Brazil presents logistics challenges that most markets never encounter. The fifth largest country in the world, it spans 8.5 million square kilometers — larger than the contiguous United States — with 26 states, each operating under distinct tax rules. Forty-five percent of its territory sits in the northern region, where few paved highways exist, and goods must move by a combination of transport modes. Twenty-two million people live in favelas that most traditional carriers won't serve. And the infrastructure that exists is concentrated in a handful of major urban corridors.
Amazon Brazil built its original network around two-day delivery. That model worked for the markets it served. But as customer expectations shifted and the business expanded its reach, the ask changed — not incrementally, but categorically. Quick commerce demanded delivery in 15 minutes. That meant rethinking not just where inventory lived, but the entire structure of how the network was designed.
The scale of Brazil's geography made this a genuine optimization problem. Standard playbooks from mature markets — the U.S., Western Europe — didn't translate. The infrastructure, the tax environment, the geography, and the demand patterns were all different. Building a network that could serve Brazil at speed required a modeling capability Amazon Brazil didn't yet have — and that Amazon's internal tools, optimized for individual workstreams, weren't designed to provide.
Over six years, Amazon Brazil completed three full redesigns of its distribution model — the first two without dedicated network design software, relying instead on ad hoc analysis and internal tools that optimized individual workstreams but couldn't model the network as a whole. When the shift to quick commerce demanded a level of precision and speed that reactive modeling couldn't support, the team made a deliberate decision: rather than building another internal tool, they brought in Optilogic.
“Brazil isn’t a market you can model with someone else’s playbook. We had to build something designed specifically for this environment — and Optilogic gave us the platform to do that.” — Felipe, Amazon Brazil
Working with Optilogic's Cosmic Frog and DataStar, the team moved from reactive analysis to a predictive and prescriptive continuous modeling capability: one that could evaluate node placement, replenishment strategy, and carrier decisions at the level of precision Brazil's geography and speed requirements demand.
As of early 2026, Amazon Brazil has launched quick commerce operations in eight main cities for Brazil, with an expected growth by three times targeted by year end. Sub-same-day deliveries, which were launched in two cities (Sao Paulo and Rio de Janeiro), are also expanding to main cities. In Southeast Brazil, 55 to 60% of total demand is now served by the localized fulfillment model the team designed and has continued to optimize using Optilogic.
The broader capability that Optilogic enables — rapid scenario modeling, continuous optimization, and the ability to respond to new questions without starting over — is what allows Amazon Brazil to operate at the frontier of quick commerce in one of the world's most logistically complex markets.
“The speed and complexity of our network decisions has to match the speed of our network changes. With Optilogic, we’re not rebuilding the model every time conditions change; we’re already running the next scenario.” — Felipe, Amazon Brazil
How Amazon Brazil built a distribution network across one of the world's most complex markets — and how Optilogic is powering what comes next.
Brazil presents logistics challenges that most markets never encounter. The fifth largest country in the world, it spans 8.5 million square kilometers — larger than the contiguous United States — with 26 states, each operating under distinct tax rules. Forty-five percent of its territory sits in the northern region, where few paved highways exist, and goods must move by a combination of transport modes. Twenty-two million people live in favelas that most traditional carriers won't serve. And the infrastructure that exists is concentrated in a handful of major urban corridors.
Amazon Brazil built its original network around two-day delivery. That model worked for the markets it served. But as customer expectations shifted and the business expanded its reach, the ask changed — not incrementally, but categorically. Quick commerce demanded delivery in 15 minutes. That meant rethinking not just where inventory lived, but the entire structure of how the network was designed.
The scale of Brazil's geography made this a genuine optimization problem. Standard playbooks from mature markets — the U.S., Western Europe — didn't translate. The infrastructure, the tax environment, the geography, and the demand patterns were all different. Building a network that could serve Brazil at speed required a modeling capability Amazon Brazil didn't yet have — and that Amazon's internal tools, optimized for individual workstreams, weren't designed to provide.
Over six years, Amazon Brazil completed three full redesigns of its distribution model — the first two without dedicated network design software, relying instead on ad hoc analysis and internal tools that optimized individual workstreams but couldn't model the network as a whole. When the shift to quick commerce demanded a level of precision and speed that reactive modeling couldn't support, the team made a deliberate decision: rather than building another internal tool, they brought in Optilogic.
“Brazil isn’t a market you can model with someone else’s playbook. We had to build something designed specifically for this environment — and Optilogic gave us the platform to do that.” — Felipe, Amazon Brazil
Working with Optilogic's Cosmic Frog and DataStar, the team moved from reactive analysis to a predictive and prescriptive continuous modeling capability: one that could evaluate node placement, replenishment strategy, and carrier decisions at the level of precision Brazil's geography and speed requirements demand.
As of early 2026, Amazon Brazil has launched quick commerce operations in eight main cities for Brazil, with an expected growth by three times targeted by year end. Sub-same-day deliveries, which were launched in two cities (Sao Paulo and Rio de Janeiro), are also expanding to main cities. In Southeast Brazil, 55 to 60% of total demand is now served by the localized fulfillment model the team designed and has continued to optimize using Optilogic.
The broader capability that Optilogic enables — rapid scenario modeling, continuous optimization, and the ability to respond to new questions without starting over — is what allows Amazon Brazil to operate at the frontier of quick commerce in one of the world's most logistically complex markets.
“The speed and complexity of our network decisions has to match the speed of our network changes. With Optilogic, we’re not rebuilding the model every time conditions change; we’re already running the next scenario.” — Felipe, Amazon Brazil
Fill out the form to unlock the full content
.jpg)
.png)
.png)
.png)