What happened
OpenAI announced on Aug. 27 that it launched commercial operations in Brazil, establishing a local team in São Paulo. The company cited rapid use of ChatGPT and Codex, new partnerships with Brazilian educational, research, health, legal, small-business and public-sector organizations, and developer programs as the foundation of its expansion.
OpenAI said it launched commercial operations in Brazil on Aug. 27 with a São Paulo team for Brazilian businesses, developers, researchers and public institutions. The team will identify use cases, train people, establish governance and expand deployments. OpenAI calls this a long-term commitment, not a single product release, covering commercial expansion of ChatGPT, Codex and related services and local institutional capacity. The company says Brazil is one of ChatGPT’s three largest markets by weekly active users; users there nearly doubled over the past year and send about 215 million messages daily. It said 35% of classified messages from individual Brazilian accounts were work-related in June 2026, versus 30% globally, and 53% of those requested a task or output such as drafting proposals, analyzing information, writing code or producing marketing materials. Brazil ranks second globally by API developers, is Codex’s largest Latin American market, and weekly Codex users grew more than elevenfold since the start of 2026, while daily interactions rose nearly thirtyfold. OpenAI gives no São Paulo team size, investment, revenue targets or program schedule.
OpenAI listed partnerships with the Instituto Tecnológico de Aeronáutica, which will provide ChatGPT Edu accounts and advanced-capability credits to students, faculty and staff; and, through a global grants program, the Instituto de Matemática Pura e Aplicada, which will study research institutions’ needs for turning AI access into scientific progress. The Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo will evaluate a clinical-information infrastructure prototype that OpenAI says preserves physician judgment and human oversight. OpenAI is developing legal AI-literacy training with ENTER, mobile-first training for micro and small-business owners with Estímulo, and a memorandum of understanding with São Paulo through Prodam to make AI available to the city and potentially other Brazilian administrations. The source gives no terms for the MOU or partnerships.
A Portuguese-language OpenAI para Devs roadshow brought 285 developers, startup engineers, founders, technical creators and other builders to Recife, Rio de Janeiro, Florianópolis and São Paulo, followed by a São Paulo hackathon. The winning team used Codex to build a prototype for micro and small businesses seeking public-procurement opportunities. OpenAI says it uses public data, accepts voice descriptions of businesses, matches them with opportunities and eligibility requirements, and maps nearby public-sector demand. The company also says Brazil has the world’s highest ChatGPT Images use rate, is among the five largest markets for voice and dictation, and has seen weekly voice users double over the past year. These are company claims.
Why it matters
The announcement moves OpenAI’s Brazil strategy beyond product access toward local sales, training, research and institutional deployment. Its claims suggest growing use of AI in Brazilian work and software development, but the source does not independently verify OpenAI’s adoption figures or establish that the announced partnerships have produced measurable public benefits.
The announcement shifts OpenAI from selling access toward a country-specific adoption and support network linking commercial operations with education, developer assistance, research grants, legal training, small-business instruction and public-sector exploration. That matters because OpenAI says Brazilian users already apply ChatGPT to routine work, software development and business administration. If its work-use classification is accurate, Brazilian users request work outputs more often than users globally, with many asking for completed tasks rather than general experimentation; this could affect marketing, pricing, inventory, customer communication and software creation, while increasing dependence on one provider. Examples from a jewelry designer and food-business owner are OpenAI-supplied individual accounts, not evidence of representativeness.
The figures’ significance is limited by missing methodology: the source does not explain the classification method, sample size, anonymization, or distinction between productive and unreliable or low-quality output. Nor does it independently verify outcomes. Education and research access through ChatGPT Edu and Codex could give students and researchers new tools, while legal and small-business training could help people use AI more critically and effectively. The clinical project’s stated physician judgment and human oversight acknowledge that health applications require safeguards. The procurement prototype could help navigate government opportunities, but a memorandum, prototype or evaluation is not scaled deployment or evidence of improved services. The source does not describe privacy protections, procurement rules, clinical validation, accessibility, language coverage beyond Portuguese events, or error-resolution mechanisms.
OpenAI cited a RegLab study, funded by OpenAI, estimating that AI could add nearly R$1 trillion to Brazil’s economy by 2030; this is a forecast, not an observed result, and the funding relationship matters. The announcement presents Brazilian creators as contributors to international AI development through text, images, video, music and voice, which may matter for representation and cultural adaptation, but supplies no independent impact assessment. The defensible conclusion is that OpenAI is investing in local adoption infrastructure while benefits and risks remain to be measured.
What to watch next
Key questions include how OpenAI’s commercial operations will be structured, when the announced programs will begin, what data and safeguards will govern public-sector and clinical work, and whether the reported growth translates into durable benefits for Brazilian workers, small businesses, researchers and creators.
Execution is the first issue to watch. OpenAI announced multiple programs but gives no launch date for each, participant count, committed funding or success metrics. It broadly describes the São Paulo team’s responsibilities and discloses no staffing, local revenue, customer numbers or partner arrangements. Follow-up reporting should distinguish signed agreements from operational deployments, pilots from permanent services, and prototypes from tools businesses or governments actually use.
The Hospital das Clínicas initiative is an evaluation of a people-first clinical-information infrastructure prototype, not a deployed clinical system. The São Paulo arrangement is a memorandum of understanding; the source does not specify the public services, data processing or any city commitment to purchase or deploy OpenAI products. Key unknowns are data retention, access controls, audits, liability, human review, accessibility, and error reporting and correction. Legal-literacy and small-business programs should be assessed for independent evaluation and guidance on confidential or sensitive information.
OpenAI does not explain how it calculated daily messages, weekly users, work classifications, API-developer rankings, Codex growth, or the fivefold increase in enterprise seats, including whether figures cover free and paid accounts, repeated or automated activity, or competitors. The RegLab projection needs scrutiny of its assumptions and methods. Future coverage should measure productivity, earnings, public-opportunity access and research output alongside inaccurate outputs, unequal access, privacy exposure and provider concentration. Brazil’s first Portuguese-language roadshow surfaced requests for tools, technical resources and local support, while Brazilian cultural work was presented as input to global product development; it remains unknown whether localization will change model behavior, product design, safety policies or commercial terms, or deliver benefits beyond large customers and highly connected users. OpenAI’s long-term commitment will be more credible if later updates provide transparent evidence about those changes and benefits.
