The US wants what Latam has

Three. Full. Weeks.

That’s how much time an HR team can save every year by managing employee benefits with iFood Benefícios.

For context, that’s 110 hours.

Not spent recruiting. Not improving culture. Not retaining your best people.

Spent adding employees, managing balances and cards, processing payments, preparing reports and dealing with bureaucracy.

Employee benefits are supposed to help your people. They probably shouldn’t consume three weeks of the HR team’s time.

iFood Benefícios puts all of that administrative plumbing into one management platform—with zero fees for the company.

Your HR team has better things to do.

Request a demo here.

Fresh off the U.S. midterm election results, I have a pretty radical idea: this outcome offers a generational opportunity for Latin America to get a seat at the AI table by building mission-critical infrastructure.

Voters in states where companies planned major data-center expansions elected politicians promising to block or limit construction. Close to half of the country’s planned capacity is now at risk of never getting built. And the market’s been on a wild ride since then:

But we still need more compute. The AI race shows no signs of slowing down. Companies like Higgsfield, now adding over $10 million in new ARR per day, can’t leave their growth dependent on whether enough computing capacity will be available.

Now it’s a question of where that capacity will come from.

Latin America has three ways to capture that demand: export energy, use that energy to export computing, and supply the equipment and materials the build-out requires.

Brazil combines enormous energy resources with a large domestic market and an established technology ecosystem. It can build computing capacity for its own businesses and sell to customers abroad. Paraguay and Uruguay show why a country’s domestic market doesn’t have to determine the size of its computing industry. Paraguay has abundant hydroelectricity; Uruguay has a largely renewable power system. Both can use that electricity at home to produce computing for customers elsewhere.

Argentina has its own opportunity in Vaca Muerta. It can export gas through pipelines and LNG, or use that gas domestically to generate power for data centers and export the computing. Hosting those facilities creates additional businesses around the energy: construction, connectivity and operations. Long-term computing customers can also help finance new power generation.

Chile can use its solar resources to power computing and build the storage and transmission needed to supply electricity around the clock. Alongside Peru, it also supplies copper for the grids, substations and electrical equipment being built internationally.

Mexico can participate on both sides of the U.S. border. It has a large domestic market for computing and a manufacturing base connected to American supply chains. It can host data centers and supply the cooling systems, power-distribution equipment and electrical components that facilities elsewhere need. A data center built in Texas can still mean business for a Mexican manufacturer.

Those advantages are already attracting U.S. companies.

Crusoe, the U.S. AI infrastructure company backed by Founders Fund and Nvidia, partnered with Paraguay’s Araico to build hydro-powered AI infrastructure near Itaipú after announcing a $3.9 billion Series F at a $30.9 billion valuation. Operations are planned to begin in mid-2027, with a path to 106 megawatts of IT capacity serving U.S. and Latin American customers.

At comparable U.S. infrastructure leasing rates, those 106 megawatts could generate roughly $200 million a year. One project, with revenue equivalent to roughly one-sixth of Paraguay’s entire electricity exports in 2025.

And this isn’t an isolated example. The hyperscalers have already committed billions to data centers across Latin America. AWS is putting over $5 billion into building and operating data centers in Mexico and another $4 billion into a similar project in Chile. Google is investing just south of $1 billion in a data center in Uruguay. These investments were already heading south before the midterms. Now the midterms could put that trend on steroids.

Latin America can turn its energy advantage into a new export industry, with investment upfront and recurring income from computing afterward.

Think about it: a fully equipped, one-gigawatt AI factory can cost $50–60 billion. Just four gigawatts would represent roughly $200 billion in investment. And the economic activity continues long after construction. Power has to be generated, facilities operated, and computing sold, year after year.

Inference is what makes this demand recurring. Training a model is only the beginning. Once businesses use it to review insurance claims, collect payments, answer customers or write software, it has to keep running. Latin America can supply the computing behind that work and get paid by customers across the Americas.

AWS alone estimates that its Mexican project will contribute roughly $10 billion to the country’s GDP over 15 years. That is one company’s project in one country. Now imagine that activity across Latin America, with local companies building and owning the infrastructure, supplying the power and selling computing internationally.

That is why I think demand for inference and compute could change the growth trajectory of entire economies in the region. But countries have to go after it.

Brazil is actively trying. REdata, signed into law in September, offers tax relief on qualifying data-center equipment, tied to requirements around clean energy, water efficiency, research and capacity for Brazilian customers. While parts of the U.S. are questioning whether they want more data centers, Brazil is creating incentives to attract them.

But attracting the investments is only part of the opportunity. Another one is how much of that business Latin America gets to own.

If the region supplies the land and electricity while everyone else owns the infrastructure and customer relationships, it captures only part of the value it makes possible. There is room for regional companies to own facilities, operate computing platforms, manufacture equipment and earn recurring revenue from global customers.

Some founders are already going after it. Take Colombia-based Panandina. Forrest Heath, the American entrepreneur behind Medellín’s Somos Internet, is taking on the next infrastructure bottleneck: power and computing. With backing from Ribbit, USV, Lowercarbon and Kaszek, and more than $200 million raised by my estimate, Panandina is building energy and data-center infrastructure around the region’s abundant power. The bet is straightforward: build from Latin America and sell computing to the world.

My guess is that many more companies built on the same logic can be started and funded in the region. That is the opportunity I see after the midterms: deeper integration between U.S. demand and capital and Latin America’s energy, infrastructure and entrepreneurial talent. Now it’s a matter of going after it.

For companies operating across countries, moving money gets expensive and complicated fast. Every new market can mean another bank, another currency, another payment rail and another layer of reconciliation.

Jeeves is built to remove that complexity. Its stablecoin-native financial operating system brings payments, FX, corporate cards and treasury into one stack — allowing global companies to move and manage money across markets without rebuilding their financial infrastructure country by country.

For Orinter, one of Brazil’s largest B2B travel operators, the impact was 87% lower international transfer costs.

The less it costs to move money globally, the more capital you have to grow globally.

So if you’re a founder, CFO or finance leader building across markets, stop building financial infrastructure country by country. Go to tryJeeves.com to see what a global financial OS can do for your company.

ByteDance is considering expanding its data-center project in northeastern Brazil.

TikTok’s parent company is looking at a larger build-out in the region, where Brazilian companies are already developing infrastructure for it. Omnia, Pátria’s data-center platform, signed a $2B, 20-year power agreement with Casa dos Ventos earlier this year to supply the facility it is building for ByteDance.

→ This reinforces the argument in this week’s essay: Latin America can turn its energy advantage into long-term income from computing infrastructure. ByteDance brings the demand, Pátria’s Omnia builds the facility, and Casa dos Ventos supplies the power. A global technology company’s expansion becomes business for local infrastructure investors and energy developers, with contracts running for decades. And the demand is coming from China as well as the U.S.

MercadoLibre grew quarterly revenue 45% to $10.75B, but operating profit fell 6.6% to $676M.

More free shipping, a nearly 20% quarterly expansion in fulfillment space and higher spending on affiliates in Brazil weighed on profitability. Customers are buying more frequently, but cheaper products and smaller baskets mean less revenue per delivery. MercadoLibre is also spending more to reach shoppers through creators and social media, where affiliate orders grew 240% year over year.

→ This reinforces my point about Chinese competition: companies don’t have to lose market share to feel pressure on profits. MercadoLibre is still growing aggressively, but competing with the likes of Shopee and TikTok Shop means spending on cheaper purchases, free delivery and customer acquisition. The company can keep winning customers while having to spend more to win them. That is exactly what makes this competition expensive.

Rivio closes a ~$33M Series A and passes ~$180M in valuation just over a year after founding.

The Brazilian healthcare AI company raised a ~$15M extension led by Valor Capital, bringing its Series A to ~$33M, with participation from Monashees, Positive Ventures and APL. Founded by isaac co-founder Ricardo Sales and former Serasa Consumidor CEO Silvio Frison, Rivio already serves 13 hospital groups, including Einstein, Sírio-Libanês and Oswaldo Cruz. Its AI agents work alongside healthcare specialists to review medical bills, identify errors and help hospitals collect payments from insurers.

→ Hospital administration is a massive market for AI. Hospitals can spend 4–6% of their revenue on billing and collections and still lose another 5–8% to services they delivered but never billed correctly. Rivio can get paid to reduce that administrative cost and recover money hospitals are already owed. And billing is just the entry point: the same customer needs help managing staff, beds and supplies. That’s why this deal is exciting. There are enormous budgets behind these problems, and AI is making it possible to build companies that take on the work.

Revolut is investing $1B to expand in Latin America.

The commitment covers Brazil, Mexico, Colombia, Peru and Argentina, and Revolut is really gaining momentum across the region. Mexico reached 1M customers in eight months. The company secured banking authorization in Colombia, is working toward a license in Peru and completed its acquisition of Argentina’s Banco Cetelem. Revolut expects the Americas, including the U.S., to contribute one in four new customers globally in 2027.

→ Revolut is using $1B to build locally regulated banks on top of its existing global platform, following its Mexico playbook (launching with a full banking license and broad range of products from day one). Interestingly, lending isn’t at the core of its growth strategy. Credit is meant to add profitability without becoming the business’s main source of income.

Paraguay’s largest digital bank, Ueno, collapsed, leaving nearly three million customers temporarily locked out of their accounts.

Controlled by Grupo Vázquez, Ueno offered accounts, cards, payments and loans. Hungary’s OTP Bank invested $47 million through its parent company, and Ueno raised $350 million through an international bond issuance. The central bank intervened after adjustments to its financial statements revealed a severe capital shortfall.

→ What went wrong? Prosecutors allege Ueno made large loans to connected companies without sufficient collateral or evidence they could repay. They are also investigating a $74.4 million loan-portfolio sale in which the buyer alleged paid using money indirectly supplied by Ueno itself. In plain English: the bank appears to have financed the purchase of its own loans, keeping the risk inside the same circle. When regulators required additional provisions for potential loan losses, its cushion fell below the legal minimum.

Earlier this week, we dropped a special TJC Debrief with Paulo Passoni, Managing Partner at Valor Capital, on Brazil after the elections: the bull, bear and base cases for the economy, what they mean for the real, interest rates and IPOs, and where we see the biggest opportunities to invest and build over the next four years. Check it out here.

Thanks for reading,

Olga 

🎙 The J Curve  is where LATAM's boldest founders & investors come to talk real strategy, opportunity and leadership.