115,000 Engineers. Most of Them Already Left — Without Leaving.

Human Becoming

The Developer Who Earns More Than His Dean

Martín graduated from the Universidad de Buenos Aires in 2019 with a degree in computer science. He was twenty-four. His first job paid 180,000 Argentine pesos per month — roughly $3,000 USD at the time, a reasonable salary by local standards, enough for a one-bedroom in Palermo and dinners out on weekends. He worked for a fintech startup in Microcentro. The office had exposed brick and a ping-pong table. The founders talked about Series A.

By 2021, the peso had moved. Not a little. Not the way currencies drift in developed economies. It moved the way tectonic plates move — in slow, grinding shifts that suddenly produce earthquakes. Martín's 180,000 pesos was now worth $1,200. Same salary. Same apartment. Sixty percent less purchasing power. His landlord raised the rent by forty percent. The startup froze hiring.

A recruiter from a San Francisco company contacted him on LinkedIn. They offered $4,500 per month — in dollars, deposited into a Wise account. Remote. Full-time. No relocation required. No visa paperwork. Just a laptop and a decent internet connection. Martín took the job in three days.

He did not leave Buenos Aires. He still takes the subte to meet friends in San Telmo. He still votes in Argentine elections. His mother still calls him every Sunday from Córdoba. But economically, Martín left Argentina the day he received his first wire transfer in a currency his own country cannot print. He now earns roughly $54,000 per year. The average Argentine salary is $1,400 per month.^[1] A tenured professor at UBA — the university that trained him — earns less than $800.^[2]

Martín is not unusual. He is the median case. There are approximately 115,000 software engineers in Argentina, and an increasing majority of them work primarily for foreign companies.^[3] Argentina now ranks third globally as a source of remote tech talent, behind only India and the Philippines — countries with populations twenty and four times larger, respectively.^[4] The engineers have not emigrated. They are physically present. They pay rent, buy groceries, ride the colectivo. But their productive output — the code they write, the systems they architect, the products they ship — accrues entirely to companies headquartered in San Francisco, New York, London, and Berlin. Every keystroke is an export. Every sprint review is a capital transfer.

Martín does not think of himself as part of a brain drain. He thinks of himself as someone who adapted. But adaptation at individual scale is extraction at national scale, and no one has reconciled these two truths.

Structural Read

The Invisible Export Economy

The numbers tell a story that the government's narrative does not want told.

Argentina's knowledge-based services exports hit $8.9 billion in 2025, a record.^[5] This figure includes software development, IT consulting, engineering services, and design work performed by Argentine professionals for foreign clients. It has grown every year for a decade, accelerating sharply after 2020 when remote work erased the last friction between Argentine talent and global demand. For context: Argentina's total goods exports were approximately $78 billion in 2025. Knowledge services now represent more than ten percent of the country's total export revenue — larger than the wine industry, larger than the fishing industry, approaching the scale of the automotive sector.

But here is the structural contradiction: the $8.9 billion flows into Argentina as dollars, which individual workers then convert to pesos (or, increasingly, hold in parallel dollar accounts). The government benefits from the tax revenue and the dollar inflow. What the government does not benefit from is the productive capacity itself. When Martín writes code for a San Francisco fintech, he is not building Argentine infrastructure, Argentine products, or Argentine intellectual property. He is building American intellectual property from a desk in Palermo. The value creation happens in Buenos Aires. The value capture happens in Delaware.

The compensation gap makes this structurally irreversible. A mid-level software engineer working for an Argentine company earns between $18,000 and $28,000 per year in total compensation. The same engineer, with the same skills, working remotely for a US company earns between $46,000 and $82,000.^[6] This is not a marginal difference that career development or equity packages can close. It is a 200-to-300 percent gap. No Argentine startup — not even the well-funded ones — can match it, because doing so would make their labor costs uncompetitive with their own product-market economics.

The result is a labor market that has bifurcated into two economies operating in the same physical space. Economy A is the peso economy: local companies, local clients, local wages, subject to inflation that has exceeded 200 percent annually in recent memory.^[7] Economy B is the dollar economy: foreign employers, USD compensation, shielded from peso depreciation by the simple mechanism of being paid in a currency that holds its value. Economy A and Economy B share the same cafes, the same apartment buildings, the same city buses. But they inhabit different financial realities.

The peso-dollar dynamic has undergone a 900 percent shift over the past decade.^[8] In 2015, one US dollar bought approximately 9 Argentine pesos on the official market. By mid-2025, one dollar bought approximately 1,100 pesos. For workers in Economy B, this is abstract — they earn in dollars and spend in pesos, which means their purchasing power inflates as the peso deflates. For workers in Economy A, this is existential. Their savings evaporate. Their wages lag. Their colleagues leave for Economy B, and the companies that remain in Economy A cannot replace them.

Argentine startups raised $2.8 billion in venture capital between 2020 and 2025.^[9] This is a real number, driven by genuine innovation in fintech (Ualá, Pomelo), agtech (Kilimo, Auravant), and logistics (Nuvocargo). But the startups face an impossible hiring environment. They need engineers. The engineers are available — they live here. But they are not available at peso-denominated salaries. The startups must either raise in dollars and pay in dollars (becoming, functionally, American companies with Argentine headcount) or accept that their best candidates will always be lured away by a recruiter offering three times the salary for the same work.

MercadoLibre, Argentina's largest technology company, navigated this by becoming a regional giant with dollar-denominated economics.^[10] But MercadoLibre is the exception. The median Argentine tech company does not have $45 billion in market capitalization to cushion its hiring strategy. The median Argentine tech company is a fifteen-person shop in Recoleta trying to compete for talent against every company in the world that has discovered Argentine engineers are excellent and cheap.

Pattern Confirmation

The Milei Paradox

President Javier Milei took office in December 2023 promising to dollarize the economy, slash the state, and attract foreign investment. Two and a half years later, the results are a contradiction that his supporters and opponents interpret in irreconcilable ways.

The "super peso" narrative is real, if provisional. Milei's government implemented a crawling peg that appreciated the peso against the dollar in nominal terms through early 2026, supported by a $20 billion IMF program and aggressive fiscal austerity.^[11] Inflation dropped from 25 percent monthly in late 2023 to single digits monthly by mid-2025. The blue-dollar gap — the spread between the official exchange rate and the parallel market rate — narrowed dramatically. For the first time in years, Argentines holding pesos were not watching their savings dissolve in real time.

But the structural incentives for the invisible exodus have not changed. Even with a "strong" peso, the compensation gap between Argentine salaries and foreign remote salaries remains enormous. A peso that buys 1,100 to a dollar instead of 1,400 to a dollar does not change the fundamental calculus when the alternative is earning in a currency that has appreciated against every emerging-market currency on earth for twenty years. The invisible exodus does not require a currency crisis. It only requires a currency differential.

Simultaneously, Milei's government is pursuing what it calls an AI-hub strategy. In January 2026, OpenAI announced a $25 billion data center investment in partnership with Argentine energy companies, designed to leverage Patagonia's cold climate and cheap hydroelectric power for AI compute infrastructure.^[12] The announcement was treated as validation: Argentina would become a node in the global AI supply chain. Milei posted about it on X. The business press ran breathless profiles.

But a data center is not a talent ecosystem. A data center is a building full of GPUs that requires perhaps two hundred operational staff. The talent ecosystem — the 115,000 engineers, the university pipeline, the startup culture — requires something the Milei government is actively dismantling: public universities.

Argentina's public university system, anchored by UBA, is one of the most productive engineering pipelines in Latin America. It is free. It is open-access. It has produced Nobel laureates, built nuclear reactors, and trained the generation of developers that made Argentina the third-largest source of remote talent on earth. And it is being gutted.

University budgets were slashed in 2024 and again in 2025. Professor salaries declined by approximately sixty percent in real terms between 2023 and 2026.^[2] Full-time faculty positions are being replaced by part-time contracts. Research budgets have been eliminated or frozen. CONICET, the national scientific research council, saw its budget cut by over forty percent.^[13] The marches of 2024 — when hundreds of thousands filled the streets of Buenos Aires to defend public education — produced political pressure but not policy reversal.

The paradox is precise: the government is trying to attract a $25 billion AI data center while defunding the institutions that produce the engineers who would staff it. It is promising a knowledge economy while dismantling the knowledge infrastructure. The OpenAI data center will import its cooling systems, its chip architecture, its operational protocols, and, increasingly, its senior technical staff. What it will not import is the mid-level engineering talent that makes a technology ecosystem function — because that talent is already here, already trained, and already working for someone else.

The Emigration That Actually Happened

The invisible exodus — working remotely for foreign companies while remaining in Argentina — is one half of the story. The other half is the visible exodus, and it is larger than at any point since the 2001 economic collapse.

Emigration from Argentina has surpassed 2001-crisis levels.^[14] Spain is the top destination, driven by language, colonial-era citizenship pathways, and a booming Spanish tech sector that actively recruits in Buenos Aires. Uruguay is second — close enough to visit family on weekends, stable enough to park savings without existential anxiety. Portugal, Italy, and the United States round out the top five.

The profile of the emigrant has shifted. In 2001, Argentina exported working-class labor and middle-class professionals fleeing a banking collapse. In 2025-2026, the departures are concentrated in the knowledge class: engineers, designers, product managers, data scientists. These are the people who could stay and earn in dollars remotely — and some of them did, for a while. But remote work from Buenos Aires carries risks that a stable legal domicile in Madrid does not. Argentine tax law is ambiguous about foreign-sourced income. Currency controls, while relaxed under Milei, have a history of being reimposed without warning. The informal economy — holding dollars in crypto wallets, receiving payments through third-party platforms — works until it doesn't.

The engineers who leave take more than their labor. They take their networks, their institutional knowledge, their mentorship capacity. A senior developer who emigrates to Barcelona is not just one fewer engineer in Buenos Aires. She is one fewer person who would have trained the next generation, co-founded the next startup, or contributed to the open-source ecosystem that makes Argentine tech visible globally. Network effects work in both directions. When the most connected nodes leave, the remaining network degrades faster than headcount alone would suggest.

The government's response has been to celebrate the dollar inflow from knowledge exports while ignoring the structural hollowing that produces it. The $8.9 billion is real. It shows up in balance-of-payments data. It helps close the fiscal gap. But it represents a country trading its most valuable long-term asset — educated human capital — for short-term current-account relief. This is the definition of an extractive economy, applied not to lithium or soybeans but to people.

The Counter-Narrative and Its Limits

There is a credible counter-narrative, and it deserves honest engagement.

Argentina's tech ecosystem is not dying. It is transforming. The $2.8 billion in venture funding is real. Companies like Ualá (digital banking), Tiendanube (e-commerce), Auth0 (identity management, acquired by Okta for $6.5 billion), and Globant (a $7 billion public company) demonstrate that Argentine founders can build globally relevant companies. Globant alone employs over 29,000 people across Latin America and has expanded aggressively into AI services.^[15]

The remote-work economy, under this reading, is not extraction. It is integration. Argentine engineers earning USD are participating in the global economy on equal terms for the first time. They bring those skills, those networks, and those dollar earnings back into the local economy. The cafés of Palermo are full because remote workers spending dollars can afford to eat out. The coworking spaces are expanding because demand for professional infrastructure is growing. The talent is staying, the dollars are flowing in, and the ecosystem is evolving into something new — a distributed node in the global tech supply chain rather than a self-contained national industry.

This narrative is not wrong. But it is incomplete. It describes the condition of 2026 without projecting it forward. The question is not whether Argentine engineers can earn good salaries today. They can. The question is whether Argentina can continue producing those engineers in five years, ten years, twenty years, if the universities that train them are defunded, the professors who teach them are earning poverty wages, and the research institutions that push the frontier are being stripped for parts.

A country can export talent for a generation on the strength of its existing educational infrastructure. But educational infrastructure, once degraded, takes decades to rebuild. The professors who leave do not come back. The research programs that close do not reopen. The students who would have enrolled in computational physics at UBA choose to study in Madrid instead, because the program in Madrid still has funding and the program in Buenos Aires does not. This is not a problem that shows up in next quarter's export figures. It shows up in 2035, when the pipeline of junior engineers thins, and the recruiters from San Francisco start looking at Colombia instead.

Alternative Explanations

It is possible that the Milei government's university budget cuts are a transitional measure rather than a permanent policy — that fiscal consolidation required short-term pain across all government sectors, and university funding will be restored once the macro stabilization is complete. Under this reading, the alarm about educational decline is premature. This explanation would be more convincing if the government had articulated a timeline for restoration, or if the cuts were proportional across all sectors. They are not. Military spending has been maintained. University spending has been disproportionately targeted. The austerity is selective, which suggests ideology rather than necessity.^[16]

A second counter-argument holds that private universities and coding bootcamps will fill the gap left by defunded public institutions — that market mechanisms will produce the engineers that the state no longer subsidizes. Argentina has a growing ecosystem of private tech education (Digital House, Plataforma 5, Henry). But these institutions produce job-ready developers, not computer scientists. They teach React frameworks, not distributed systems theory. The distinction matters at scale: a country of bootcamp graduates can staff call centers and build CRUD applications. A country of university-trained engineers can architect the systems that the call centers and CRUD applications depend on. The pipeline that produced Auth0 and MercadoLibre was not a twelve-week bootcamp.

What is not known: The exact percentage of Argentina's 115,000 engineers who work primarily for foreign clients versus domestic companies. Industry estimates range from 45 to 70 percent, but no comprehensive survey exists. The tax compliance rate for dollar-denominated remote income is also unknown — government figures suggest significant under-reporting, but the magnitude is disputed.

What is not confirmed: Whether the OpenAI data center deal will proceed as announced. Large-scale infrastructure announcements in Argentina have a history of scaling back or stalling. The $25 billion figure is a commitment, not a disbursement.

What would change the signal: If university budgets were restored to 2022 real levels, the brain-drain acceleration would slow. If Milei's macro stabilization produces sustained peso stability below 50% annual inflation for three consecutive years, the compensation gap between local and foreign employment would begin to narrow organically. If OpenAI's data center generates a local engineering hiring wave at competitive salaries, the narrative shifts from extraction to ecosystem growth. None of these have occurred as of June 2026.

Monitoring indicators: Track UBA computer science enrollment annually. Track CONICET researcher count quarterly. Track Argentina's position in Deel/Turing global remote talent rankings. Monitor peso-dollar rate trajectory monthly. Track venture capital deployed to Argentine startups by quarter. Monitor emigration data from Spain's INE for Argentine arrivals. Track knowledge-services export volume in INDEC data.