Brazil's AI adoption boom in public numbers: what IBGE, Bain, Gartner, and ANPD actually say

In two years, the share of Brazilian industrial companies using artificial intelligence went from 16.9% to 41.9%. The IBGE published that figure in September 2025 in the PINTEC Semestral. It is a 163% jump in twenty-four months, the kind of curve that makes consultants and software vendors very happy.

In the same two years, almost nothing changed in how the typical Brazilian company treats the data those AI systems are reading.

This piece is an attempt to look at what the public statistics actually say about Brazilian AI adoption, where the gap between ambition and capability sits, and what that gap is going to cost over the next two years. Sources are listed at the bottom; everything cited here is public.

The adoption numbers

Six independent measurements, run by different methodologies, all converged through 2025 on the same finding: Brazilian companies adopted AI faster than almost anywhere outside the US tech corridor.

The IBGE PINTEC Semestral 2024, released in September 2025, found that 41.9% of industrial companies with 100 or more employees were using AI by 2024, up from 16.9% in 2022. The absolute count rose from 1,619 firms to 4,261. Areas of heaviest application were Administration (87.9% of those firms), Commercialization (75.2%), and Product, Process, and Service Development (73.1%).

Bain and Company's South America study from late 2025 placed the share of Brazilian companies with at least one AI use case in production at 25%, more than double the 12% recorded a year earlier. Sixty-seven percent of executives surveyed called AI a strategic priority for the year. Microsoft's 2025 Brazil SMB report said that 74% of Brazilian small and medium-sized businesses use AI daily in some form. Sebrae's 2025 digital maturity research showed 35% of micro and small businesses using AI specifically to analyze data and improve operational efficiency. IBM and Morning Consult, in a separate survey of more than a thousand executives, found that 78% of Brazilian companies plan to expand their AI budgets through the end of 2025.

The headline interpretation is straightforward: real AI adoption is happening, at a pace ahead of the global average. Brazil is not behind on AI in the way it was behind on cloud or on warehouse modernization a decade ago. The interesting question, which the same studies do not answer, is what those companies are running AI on top of.

What the foundation looks like

The Brazilian IT services market reached US$17.31 billion in 2025 and is projected to reach US$29.94 billion by 2030 at an 11.57% compound annual growth rate, according to Mordor Intelligence. The full IT market sits at US$67.8 billion, putting Brazil in 10th place globally and at the unambiguous top of Latin America with 38.4% regional share, per ABES and IDC. Cloud computing alone grew 25% in 2025, ahead of the 18% global average.

Healthy macro numbers, and they obscure the real distribution of maturity. The same IBGE PINTEC dataset that showed AI adoption climbing also reported that the overall industrial innovation rate dropped to 64.4% in 2024, the third consecutive annual decline since 2021, when it was 70.5%. Brazilian industry as a whole is innovating less, not more, even as the AI subgroup grows.

The Bain survey is more direct about the maturity gap inside the AI cohort itself. It reported that 72% of Brazilian companies that have adopted AI sit at beginner or experimental levels of maturity. Adoption is wide. Maturity is shallow. Those are the two findings that should worry the buyers and sellers in this market.

The failure rate, in numbers

International research on AI project outcomes does not get gentler when you read more of it.

Gartner, in a survey of 782 infrastructure and operations leaders run between November and December 2025, reported that only 28% of AI use cases fully succeed and meet ROI expectations. Another 20% fail outright. The middle 52% deliver partial value, which in practice means they consume budget and produce inconclusive results. A separate Gartner finding, repeated in their 2025 data management research, places the share of AI projects that fail due to poor data quality or absence of relevant data at 85%. A 2024 Gartner survey of 248 data management leaders found that 63% of organizations either lack the right data management practices for AI or are unsure whether they have them.

By the end of 2025, Gartner predicted that 30% of generative AI projects would be abandoned after proof of concept due to poor data quality, inadequate risk controls, escalating costs, or unclear business value. Secondary analyst sources in 2026 reported the actual post-PoC abandonment rate above 50%.

Take those numbers and apply them to the Brazilian adoption curve. If 4,261 Brazilian industrial firms were running AI by 2024, and the international failure rate sits somewhere between 50% and 85% depending on how you count, the implied volume of stalled or abandoned AI initiatives inside Brazilian industry alone is in the low thousands. That is not a forecast. It is an inventory.

The LGPD sleeping giant

The other piece of context that public data makes clear is the regulatory exposure most companies are not pricing in.

The Brazilian Lei Geral de Proteção de Dados has been in force since 2020. ANPD enforcement has been notably gentle: through August 2024, only 18 administrative sanctions were applied, of which only two were fines, both against the same micro-enterprise (Telekall Infoservice), totaling R$14,400. No private company was fined for an LGPD violation in all of 2024. Five sanction processes were closed in 2024, all against public agencies (the Ministry of Health, INSS, the Federal District Education Secretariat, and the Pernambuco Social Assistance Secretariat).

For comparison, the European Union, the United Kingdom, the United States, Argentina, and Australia together applied US$1.7 billion in data protection fines in 2024. Brazil applied none.

That gap is closing. ANPD's Deliberação CD-10/2025 introduced daily fines for non-compliance with cautionary measures. Telecom and fintech sectors are under increasing scrutiny. Legal commentary across 2025 converges on the expectation that the next two years will see the first serious financial sanctions against private Brazilian companies. The companies running AI on top of warehouses they cannot fully audit are the ones most exposed when this enforcement cycle begins.

The agribusiness slice

Worth a separate section because the numbers are unusual.

Brazilian agribusiness contributed R$3.20 trillion to GDP in 2025, equivalent to 25.13% of the national economy, up from 22.9% in 2024, according to CEPEA-Esalq/USP and CNA. The pecuária (livestock) ramp grew 32.55% over the year. The sector is large, growing fast, and has more capital available than most Brazilian industries.

On the digital adoption side, 84% of interviewed Brazilian farmers report using at least one digital technology in their production system, according to a 504-farmer survey published in MDPI's Agriculture journal. AI-equipped farm machinery is operational on roughly 10 to 15% of commercial farms in Brazil, per Farmonaut's 2025 review. The smart agriculture subsector is projected to grow from US$404 million in 2024 to US$795 million by 2033, per IMARC Group. BNDES has estimated that fuller IoT integration in Brazilian agriculture could generate between US$50 and US$200 billion in annual economic impact.

The peculiarity is that most of this digital data, despite the high adoption percentage, is not consolidated. A producer running John Deere telemetry, satellite NDVI from a paid weather subscription, ERP data from TOTVS, and futures positions from B3 has four parallel data streams that nobody on the team is qualified to merge into a single decision-making layer. The adoption is wide. The integration is missing. I wrote about this specific case in detail in why Brazilian agribusiness needs a data concierge.

What this points to over 2026 and 2027

The synthesis of the public data is consistent. Brazilian companies are buying AI and digital tools faster than they are building the foundations needed to use them well. The mismatch will resolve in one of two ways over the next two years.

In the first scenario, foundations get built. Some fraction of Brazilian companies invest in the warehouse, the catalog, the metric definitions, and the governance work that turns AI initiatives into measurable returns. The Gartner-style 28% success rate might rise to 35 or 40% inside that cohort. They compound advantage over the rest of the market.

In the second scenario, the foundations stay broken and the AI initiatives stall. The Gartner abandonment rate plays out in Brazilian boardrooms over 2026 and 2027 as the post-PoC budget cycle ends and CFOs look for results that are not there. The mid-market segment is the most exposed because it lacks the in-house seniority to diagnose the problem before the budget runs out.

Both scenarios are happening in parallel right now in different companies. The question for any given executive is which scenario is happening in their company, and they probably do not know yet, because the failure mode of AI initiatives is silence, not crisis.

Closing thought

The public data is unambiguous about the adoption boom and unambiguous about the failure rate. The data nobody is publishing is what the foundation looks like inside any individual company, because that data is hard to produce. It requires somebody walking through the warehouse, mapping the metrics, checking the lineage, and writing down what is actually true.

That is the work that determines which scenario plays out. The numbers above are the inventory. The work is everything else.

Related reading

For the strategic framing of the foundation gap, see fix your data before adopting generative AI. For the Brazilian-market context, see data consulting in Brazil. For the practical case of companies that have committed to an AI rollout but have not started the foundation work, see companies transitioning to AI need a data concierge.

Sources

IBGE PINTEC Semestral 2024 (September 2025 release): IBGE Agência de Notícias; industrial innovation rate decline: IBGE 2024 innovation rate report.

Bain and Company South America AI study 2025: Bain press release.

Microsoft 2025 Brazil SMB report: Microsoft News Center Brazil.

Sebrae 2025 digital maturity research: Sebrae Agency.

IBM and Morning Consult AI investment survey: Exame coverage.

Gartner infrastructure and operations AI ROI survey, April 2026 release: Gartner press release; Gartner data readiness for AI: Gartner data readiness research; Gartner GenAI abandonment forecast: Gartner GenAI press release.

ABES and IDC Brazilian IT market 2025 study: ABES report; Mordor Intelligence Brazil IT services 2025-2030: Mordor Intelligence.

ANPD enforcement data: TI Inside (five years of LGPD); Mayer Brown 2024 retrospective.

Brazilian agribusiness GDP and digital adoption: CNA Brasil 2025 GDP report; MDPI Agriculture digital adoption study; IMARC Brazil smart agriculture market; Farmonaut AI agriculture statistics.