the end of the model

in the last twelve months, more than 1,500 people lost their jobs in uruguay's tech sector. pedidosya, ukg, basf, sabre, verizon/alorica. different names, same pattern: announcement without notice, operations moved to india or to cheaper markets, workers finding out through the press.

the government reacted by calling for mandatory advance notice (). the industry asked for exchange rate competitiveness. nobody talked about what is really at stake.

what is at stake is the entire model.

uruguay spent twenty years building a reputation as a software exporter. the largest per capita in south america, according to some rankings. an industry that billed us$3,681 million in 2024, equivalent to 4.5% of gdp. figures that generate pride and that, looked at closely, also raise uncomfortable questions.

because behind those numbers there is a fragile architecture. roughly 70% of the sector, according to data from cpa ferrere and cuti, consists of software factories: companies that sell technical capacity on demand, not a product of their own. the model works as long as local talent is cheaper than talent in developed markets and more reliable than talent in cheaper markets. for years, uruguay occupied that middle ground comfortably.

that middle ground no longer exists.

the first fracture is the exchange rate. the dollar at 39 pesos puts uruguayan labor costs at parity with germany, in relative terms. not in absolute value, a uruguayan developer does not earn the same as one in frankfurt, but in what it costs a multinational to operate here versus operating there. the basf executive who said it publicly was not being cruel. he was being precise.

service exporters charge in dollars and pay in pesos. when the peso appreciates, their margins compress without any productivity variable changing. the result is that uruguay became more expensive without becoming more productive. and that has consequences that no wage negotiation can reverse.

the second fracture is structural. india is not just cheaper. india has scale. while uruguay produces a few thousand developers per year, india produces hundreds of thousands. ukg tripled its headcount in montevideo in three years and then shut down. not because the work was bad. but because in pune they can do the same thing, with the same technical quality, at a fraction of the cost and with ten times the available volume.

uruguayan quality is real. but it is not different enough to justify the price differential when the work is interchangeable.

the third fracture is technological. artificial intelligence is not going to eliminate software work all at once. it is going to eliminate first exactly the kind of work the factories do: mid-level code, functional support, systems integration, technical documentation. the layer uruguay exports the most is the layer most exposed. globant already acknowledged it implicitly when it cut a thousand positions under the cosmetic name of "vision 2030". mercadolibre said it without euphemisms: the ux employees who trained the systems were replaced by them.

there is a fourth fracture that almost nobody mentions because it is uncomfortable: the educational base.

out of every ten young people who start secondary school in uruguay, three finish it. of those three, a smaller fraction completes a university degree. the tech sector operates on top of that funnel. high-level talent exists and is genuinely good, but it is scarce by structural definition, not by bad luck.

the problem is not just that technicians are lacking. it is that the system does not produce in sufficient quantity the kind of profile the high-value model requires: product thinking, systemic design capacity, business understanding on top of code. that profile is formed slowly, in a system that has gone decades without solving its foundational problems.

and here is the most brutal irony: artificial intelligence could be an acceleration tool for that scarce talent. but only for those who already have a base. for those who don't, ai does not compensate, it amplifies the gap.

what no political actor is willing to discuss is productivity.

uruguay has a solid tradition of debate about distribution. the wage councils, the 2007 tax reform, universal health coverage are real achievements of a society that learned to negotiate the split. but in that process, the real wage became the proxy for everything. if wages rise in real terms, the political narrative reads it as progress. period.

the problem is that the real wage does not measure productivity. it measures purchasing power. they are different things. and confusing them for twenty years has consequences: today uruguay has wages that grew above its relative productivity, which in service-exporting sectors translates directly into lost competitiveness. the bill arrived in the form of corporate announcements without notice.

discussing systemic productivity would mean touching collective bargaining, the rigidity of the formal labor market, and the real measurement of public spending by results. no political actor is willing to pay that cost. so the debate does not happen, and the model erodes in silence.

where is the sector headed, then?

the most likely trajectory is not collapse. it is bifurcation.

on one side, a tiny, sophisticated segment oriented toward its own product, ip, regional fintech and ai specialization. dlocal is the paradigmatic case: its own product, global market, unicorn valuation, a record us$746 million in revenue in 2024. tryolabs, pyxis, genexus in its niche, companies that built something that cannot simply be relocated to pune or vietnam because it is not a generic service, it is accumulated knowledge with identity.

on the other side, a void where today there are 15,000 people in factories. that void does not get filled with good intentions or with statements of support for the sector. it gets filled with a transition that the education system is not equipped to accelerate, that politics is not discussing and that the market on its own will not solve in time.

the question nobody is asking out loud is what happens to those 15,000 people (or more) in the next two years. not the ones who were already laid off. the ones still working in a model that is being gradually dismantled.

uruguay has real talent. it has institutions that work. it has a reputation for seriousness that took decades to build.

but the software export model that was built on that base is no longer sustainable in its current form. and the transition toward something better, more sophisticated, more product-based, more resilient, requires exactly the kind of conversation uruguayan politics is not having: an honest conversation about productivity, about the real cost of developing talent, and about what kind of economy we want to be when generic work is no longer available.

that conversation has not started yet. and the time to have it is shrinking.