Stability AI boss warns India's outsourced coders could vanish in 2 years

The global IT outsourcing empire might be heading toward a rapid reckoning.

Speaking on a call with UBS analysts, Stability AI CEO Emad Mostaque dropped a blunt prediction: generative AI is poised to wipe out most outsourced programmers in India within the next one to two years. As modern models make it possible to build software with far fewer human hands, the traditional developer army is finding itself on the front lines of automation.

The screen barrier and the regulatory divide

Mostaque didn't mince words about who gets hit first. If your entire job consists of sitting behind a computer screen where nobody ever physically sees you, your vulnerability is sky-high. He compared today's advanced models to a crop of exceptionally talented university graduates.

In India, Mostaque claimed that outsourced coders up to level three programmers will disappear in a matter of months. When you look at the software pipeline—from raw ideation and unit testing all the way to squashing bugs—AI systems can already handle those tasks at comparable or superior quality.

Yet geography and local law will dictate how unevenly this wave crashes. While Indian outsourcing contracts can be scaled down almost overnight, Mostaque pointed out that strict labor rules in countries like France mean firing a software engineer remains virtually impossible.

India is home to more than 5 million software programmers, according to Bloomberg data, powering offshore operations for Wall Street banks, Silicon Valley giants, airlines, and retailers. That concentration makes the region an epicenter for whatever disruption hits enterprise tech next.

Outsourcing giants scramble to adapt

Local IT juggernauts aren't simply waiting around to be replaced. Tata Consultancy Services (TCS), India's largest outsourcing firm alongside Infosys and Wipro, is actively retraining its workforce.

TCS recently committed to training more than 25,000 engineers on generative AI via Microsoft's Azure OpenAI service. In an interview with CNBC, TCS executive N. Ganapathy Subramaniam explained that the company adopted a machine-first delivery philosophy four years ago, noting that generative AI merely accelerated that transition by several years.

Mostaque also revisited his earlier claim that there will be no traditional programmers left in five years. Software creation won't become completely driverless; instead, human engineers will operate alongside AI co-pilots. The real question is whether the millions of people previously tasked with classical coding can pivot into new roles fast enough.

Why it matters

Mostaque has a reputation for provocative soundbites, but the underlying pressure point he is highlighting is very real.

For individual engineers, routine coding is no longer a defensible career moat. If your day-to-day workflow consists of translating rigid specifications into boilerplate code or running routine unit tests, an AI assistant can already match your output at a fraction of the cost. The practical playbook right now isn't avoiding code, but mastering AI workflows so you can deliver the output of an entire small team solo.

For the broader IT industry, the traditional "time-and-materials" outsourcing model is under siege. For decades, offshore firms grew revenues by adding headcount and billing hourly rates to Western corporations. If an enterprise client realizes an internal team equipped with AI copilots can handle the work of dozens of offshore contractors, the economic incentive to outsource disappears.

We think Mostaque's one-to-two-year timeline is overly aggressive, given how slow large enterprises move with legacy codebases and compliance guardrails. But the structural shift away from junior, seat-filling programming is unmistakable.