Databricks is putting R$1.5 billion into Brazil. What it says about enterprise AI demand outside the US
Databricks plans to invest about US$300 million in Brazil over three years and train 150,000 people. Why Brazil is a core enterprise AI market, and what still depends on enterprises.
Listen to this article · 5 min
AI-generated narration of the full article.
Global software vendors have often treated Brazil as a localization project: translate the product, find a reseller, revisit later. Databricks’ latest announcement points the other way.
On September 16, Databricks said it will invest more than R$1.5 billion (about US$300 million) in Brazil over the next three years to help companies advance enterprise AI adoption. For a US audience, the interesting part is not the number. It is what a data platform vendor is betting on, and what that implies for anyone running AI programs that include Brazilian operations.
What Databricks announced
According to the company’s release:
- The investment covers three years and includes expanding the local team.
- Databricks commits to training more than 150,000 people in data and AI skills over three and a half years, through academic partnerships, free online training and its Free Edition.
- It cites an ecosystem of more than 300 partners in the country.
- It says its business in Brazil nearly tripled in the last two years, and that at least eight of the ten largest companies in education, retail, health and financial services in Brazil, by the Valor 1000 ranking, use its platform. The release names customers such as Banco do Brasil, Bradesco, iFood, Natura, Nubank, Petrobras and Vale.
The product focus is telling. Alongside Agent Bricks and Unity Catalog, the release highlights Lakebase, Genie and Unity Gateway, which Databricks describes as helping companies manage AI usage, security and costs across models, agents, MCP servers and tools.
These adoption figures are Databricks’ own and have not been independently verified.
Why this matters beyond Brazil
Enterprise AI demand follows data platforms. Vendors invest where companies already have governed data at scale and are moving to agents on top of it. Brazil’s large banks, retailers and digital natives fit that profile. An investment of this size is a statement that Brazilian enterprises are buying production AI, not just experimenting.
Talent is the constraint everyone shares. A commitment to train 150,000 people is as much about the vendor’s ecosystem as about the market. For global companies, it means a deeper local talent pool for data and AI roles, mostly trained on one platform.
Governance is now part of the pitch. The emphasis on a gateway for AI usage, security and cost across models and tools mirrors what enterprises in every market are asking for. Brazilian buyers, operating under the LGPD, are asking for the same controls US buyers are.
- Local team and support
- Training at scale
- Partner ecosystem
- Which use cases reach production
- Data governance under local law
- Owners and evaluation in production
- The vendor invests: Local team and support · Training at scale · Partner ecosystem
- The enterprise still decides: Which use cases reach production · Data governance under local law · Owners and evaluation in production
Responsibilities no platform investment transfers
What it means for companies with Brazilian operations
If your organization runs AI programs across the US and Brazil, three practical implications follow:
- Plan for local delivery capacity. More certified partners and trained engineers in Brazil make it more realistic to build and operate systems there, close to the business and the data, instead of shipping everything from a central US team.
- Keep the architecture portable where it matters. Platform-specific training is valuable, but identity, evaluation assets and business definitions should not be locked to one vendor. We discussed which layers to own once across vendors in our analysis of reusable AI harnesses.
- Design governance for both jurisdictions. A gateway that controls models, tools and costs is a mechanism. The rules it enforces, including how Brazilian personal data is handled under the LGPD, still have to be written by your teams.
The bottom line
Databricks’ investment is a strong signal that Brazil is a primary enterprise AI market. For global companies, that means better local capacity to build and run AI where the business operates. It does not change the fundamentals: which use cases deserve production, who owns them and how they are governed. The 2026 enterprise AI stack guide lays out those layers.