Brazil's public sector leans into on-prem AI: Nutanix widens its footprint amid Dell and VMware competition
Brazilian agencies are pushing for on-prem, AI-ready infrastructure to keep sensitive data under tighter control. Nutanix is leaning into that shift, expanding contracts across government and public institutions while competing head-to-head with Dell (VxRail), VMware (vSphere and vSAN), Cisco (HyperFlex), Red Hat (OpenShift), and Scale Computing (HC3).
The company's bet is clear: AI workloads need fast, local processing, predictable performance, and strong compliance. Cloud stays in the mix, but mission-critical datasets often remain on site.
Local assembly, tax strategy, and why it matters to agencies
In Brazil, Nutanix partnered with Positivo to locally assemble Supermicro servers optimized for its hybrid cloud and AI stack. The move targets tax incentives tied to the Basic Production Process (PPB) and mitigates the impact of a complex tax environment.
"Operating in this environment isn't simple due to the tax burden. Tax changes have a significant impact on businesses, making the production chain more critical and making acquisition decisions more difficult," said Leonel Oliveira, Nutanix's general director in Brazil.
For public buyers, local assembly can mean shorter lead times, more predictable costs, and clearer compliance paths-especially where data sovereignty and auditability are non-negotiable.
Enterprise AI, on your terms
Nutanix Enterprise AI (NAI), launched last November, aims to simplify deployment and management of AI and large language models across on-prem, cloud, or edge. It offers secure endpoints for generative AI APIs and is already in production with Brazilian customers.
If you're evaluating options, this approach addresses a common blocker: how to run models close to your data without adding operational overhead. More on the platform can be found on the official page: Nutanix AI.
Public sector use cases in Brazil
- Court of Auditors of the State of Mato Grosso: Kept data in its own data center and used the Nutanix platform to run Meta's Llama model. The team is now exploring electronic analysis for bids as a next step in its digitalization. Learn about the model family here: Meta Llama.
- Acre State Finance Department: Testing Llama 3.2 and Gema 2 models in a Nutanix environment. The pilot, supported by IBM Granite and integrated with a Nutanix NAI cluster, has been running for three months.
Procurement reality: VMware, Dell, and Red Hat
Market shifts are driving migrations. After VMware's sale to Broadcom and uncertainty around Dell licensing renewals, some agencies are reviewing their options. A Brazilian Secretariat that spent more than a decade with VMware and Dell started moving to Nutanix for tighter security and convergence, according to senior analyst Rafael Castro.
"We operate within the market that already has VMware, but we're also serving a lot of new customers. Only 12% of our customers work with VMware," said Oliveira. He also noted that while Red Hat remains a strong competitor, some customers report faster implementation on Nutanix. As he put it: "If you don't have application portability, you don't have the freedom to come and go."
What the numbers say
- ARR: $2.2B in Q4 2025, up 17% year over year.
- Customers: 29,000 globally across finance, government, and healthcare, including Wells Fargo, Tata, Airbus, Caterpillar, NTT Data, Total Energies, Hitachi, and ConocoPhillips.
- Team and R&D: 7,500 employees; R&D at 22% of sales.
- Brazil footprint: 12 years in-country, ~500 clients across public, financial, industrial, education, and retail sectors; presence in all states.
- Channel: 100+ active partners; partner ecosystem grew 35% in Q4 FY25. All local sales run through partners.
- Adoption: AHV virtualization is in use at over 70% of local customers.
- Alliances: Collaboration with Pure Storage; recent public-sector contract with the government of Cundinamarca, Colombia.
Why this matters for government IT leaders
- Data residency and confidentiality: On-prem AI keeps sensitive workloads close to their source and simplifies audit trails.
- Predictable performance and costs: Critical for budget planning and service-level agreements.
- Tax and procurement: Local assembly and PPB incentives can improve TCO and delivery timelines.
- Vendor flexibility: Prioritize portability to avoid lock-in-especially for AI models and containerized apps.
- Operational simplicity: Consolidated stacks reduce integration risk and speed up delivery.
Practical checklist before you commit
- Workload mapping: Identify datasets and apps that must remain on-prem. Flag anything with legal or regulatory constraints.
- Model strategy: Confirm support for your short-list of models (e.g., Llama) and your preferred inference approach (GPU/CPU, online/batch).
- Security controls: Validate network segmentation, key management, API security, and audit logging across clusters.
- Portability: Test migration flows across on-prem and cloud. Prove rollback and failover.
- TCO and funding: Include tax incentives, local assembly, support tiers, and training in your total cost view.
- Partner capacity: Check the availability and certifications of local partners for deployment and support.
- Pilots with purpose: Run a 60-90 day pilot tied to a specific service outcome (e.g., bid analysis, fraud detection, or citizen support).
2026 outlook
Despite potential shifts in the government sector with October elections, Oliveira expects growth to hold, driven by modernization and more disciplined internal investment. Agencies are moving from proof-of-concepts to production workloads-especially where AI can speed audits, service delivery, and back-office operations.
Next steps for public teams
- Shortlist platforms that support on-prem, cloud, and edge with the same management model.
- Align legal, security, and procurement early-especially on data handling and audit requirements.
- Pick one high-impact use case and ship it in a controlled pilot. Measure service outcomes, not just benchmarks.
- Plan staff enablement. If your team needs a quick path to upskill on AI for public roles, see focused training options: AI courses by job.
Bottom line: Brazil's public sector is getting serious about AI on infrastructure it can control. Nutanix is positioning itself as a go-to option for that path-competing hard on simplicity, portability, and local execution.
Your membership also unlocks:
