Lula's 2034 tech push: reactor, supercomputer, Portuguese AI, 2% of GDP - and scientists coming home

Brazil set a 10-year push to lift R&D to 2% of GDP to cut dependence, with Portuguese AI, a reactor, compute, and talent return. Success hinges on steady funding and delivery.

Categorized in: AI News Science and Research
Published on: Dec 08, 2025
Lula's 2034 tech push: reactor, supercomputer, Portuguese AI, 2% of GDP - and scientists coming home

Brazil's 10-year bet on scientific autonomy: reactor, AI in Portuguese, supercomputing, and talent return

Brazil put a clear stake in the ground: lift science and technology from 1.2% to 2% of GDP by 2034 and use that momentum to cut strategic dependencies. The plan wraps AI in Portuguese, a multipurpose nuclear reactor, expanded compute, and a large-scale scientist repatriation program into one national playbook.

It was presented to President Luiz Inácio Lula da Silva at the 6th meeting of the National Council for Science and Technology as part of a broader reindustrialization effort. The message was direct: long-term industrial capacity follows consistent investment in science and technology.

The four pillars

  • Expand and consolidate the national research system
  • Incentivize business innovation
  • Execute sovereignty projects in critical technologies
  • Drive science for social development

The strategy reflects guidelines from the 5th National Conference on Science, Technology and Innovation (2024) and goes to public consultation until December 20. If you lead a lab or company R&D unit, this is the right window to shape priorities and funding mechanisms.

Why this matters for autonomy

The government diagnosed heavy external dependence in semiconductors, strategic minerals, fertilizers, pharmaceuticals, clean energy, digital technologies, and climate solutions. These are now priority targets.

Lula acknowledged weak private appetite for risk in research and signaled more decisive state action where capital is absent. The intent is to rebuild scientific and productive capacity as a sovereignty issue.

Money and mechanisms

  • Investment target: 2% of GDP in R&D by 2034
  • Scientist repatriation: R$1 billion over five years (R$200 million/year)
  • Industrial policy link: R$180 billion projected for the automotive sector under New Industry Brazil

Expect blended instruments: research grants, mission-driven programs, and incentives tied to private co-investment. The emphasis is national capacity, not single-project wins.

Infrastructure: reactor, beamlines, and risk labs

Sirius, Latin America's largest synchrotron light source, will add three new beamlines and connect to NB4, Brazil's first maximum biological containment laboratory. This pairing enables advanced studies on high-risk pathogens and faster responses to future outbreaks.

The Brazilian Multipurpose Reactor, in cooperation with Argentina, is set to reduce reliance on radiopharmaceutical imports, currently around 80%. Cemaden's disaster monitoring will expand from coverage in roughly a thousand municipalities to about 1,900 of the most vulnerable, tightening early-warning loops for floods and landslides.

For context on large-scale light sources, see the facility overview at Sirius (CNPEM).

AI in Portuguese and more compute

The Brazilian Artificial Intelligence Plan lines up infrastructure, training, and solution development. Eight National Institutes of Science and Technology are dedicated to AI, with focus areas spanning industry, public policy, and language models.

The Santos Dumont supercomputer was upgraded and returned to the TOP500 list, entering the top 100 for energy efficiency. On the model front, the government highlighted Parakeet (seen as a leading Portuguese LLM) and the Piauí model Sovereignty, aimed at digital sovereignty and public policy use cases. For benchmarking context, check the TOP500 ranking.

Talent return: numbers and direction

The repatriation program coordinated by CNPq earmarks R$1 billion over five years. More than 2,500 researchers signaled interest; 599 projects were approved; 248 are ready for call-up; and 251 are in return planning - about 34% with concrete movements.

Projects prioritize networks with institutions in the North, Northeast, and Midwest to counter regional inequality in scientific output. CNPq also updated rules to avoid career setbacks after maternity leave and launched six programs for women in science, including a R$100 million initiative with six beneficiaries.

Defense as a catalyst

The Ministry of Defense pointed to complex programs as engines for innovation with civilian spillovers. Examples include the Navy Nuclear Program, the KC-390 freighter, and cyber defense initiatives using AI and quantum computing.

Energy and climate context

Lula tied the strategy to COP30 and Brazil's position in the energy transition. Brazil operates with roughly 53% renewable energy in its total energy mix, and about 90% of electricity generation comes from clean sources.

For reference on national energy indicators, see the IEA's country data. IEA: Brazil

What R&D leaders can do now

  • Map upcoming calls linked to the 2% GDP target and prepare consortium proposals that include institutions in the North, Northeast, and Midwest.
  • Position projects to use the new beamlines at Sirius and coordinate with NB4-capable teams for pathogen work and biosafety readiness.
  • Plan compute needs and energy budgets for Santos Dumont access; align AI projects with national priorities (industry, public services, and Portuguese LLMs).
  • For AI teams, set up data governance, evaluation suites for Portuguese, and domain-specific corpora that reflect local context.
  • Engage with the repatriation program to attract or reintegrate team members; offer joint appointments and clear lab onboarding plans.
  • In health, anticipate domestic radioisotope supply from the reactor and formalize clinic-research pipelines for diagnostics and trials.
  • Partner with municipalities on Cemaden-linked projects: sensor networks, modeling, and decision-support tools for early warnings.

If you're building AI capability inside your lab or company and need structured upskilling, see practical course paths by role at Complete AI Training.

The open question

Will this be enough to meaningfully reduce technological dependence by 2034? The investment target is significant, but delivery hinges on continuity across administrations, procurement that speeds translation, semiconductor and materials strategies that go beyond pilots, and real co-investment from industry.

Watch for execution signals: yearly R&D-to-GDP movement, share of domestic inputs in health and energy, open access to compute and beamlines, the retention rate of repatriated scientists, and measurable industry adoption. If those move in the right direction, the plan has teeth - and researchers will feel it in labs, budgets, and publication-to-product cycles.


Get Daily AI News

Your membership also unlocks:

700+ AI Courses
700+ Certifications
Personalized AI Learning Plan
6500+ AI Tools (no Ads)
Daily AI News by job industry (no Ads)
Advertisement
Stream Watch Guide