Rotten Rice
Two crises. 25 million workers. One broken bet.
Late in the evening on March 6, 2026, Kuna Khuntia called his father from Doha. He was 25 years old, a pipe fitter from Odisha in eastern India, three months into a contract at an oil facility in Qatar. His family had borrowed 300,000 rupees, about $3,200, to pay the recruitment agent who placed him. He had been sending 15,000 rupees a month home.
The call lasted a few minutes. He told his father, Jaya, that he was safe. He said not to worry.
Hours later, on March 7, a missile launched from Iran came down near the worksite. Kuna did not die in the blast. He died of a heart attack, brought on by the noise, the flash, and the proximity. His father learned about it the next morning.
![Kuna Khuntia, a 25-year-old pipe fitter from India's Odisha, who died of a heart attack in Doha Qatar [Photo courtesy the Khuntia family]](https://substackcdn.com/image/fetch/$s_!M8PX!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fce038598-1baa-4d33-ba40-192709ebeb61_770x770.jpeg "Kuna Khuntia, a 25-year-old pipe fitter from India's Odisha, who died of a heart attack in Doha Qatar [Photo courtesy the Khuntia family]")
“That one call,” Jaya Khuntia told Al Jazeera, “finished us.”
Last year, major outsourcing and remittance hubs like India, Pakistan, Bangladesh, Nepal, and the Philippines received more than $236 billion in remittances from their citizens working abroad, alongside an outsourcing market worth more than $350 billion. In Nepal, remittances account for 26% of GDP. In Pakistan, close to 10%. Both economies are built on a simple bet: that by exporting people, they can one day have a better country for themselves.
These economies run on three income streams: remittances from Gulf labor, revenue from knowledge-labor services exports, and physical-goods exports. Two of the three are under attack right now, by completely unrelated forces, at the same time.
On the physical-labor side, the war in Iran has frozen hiring, closed worksites, and begun repatriating hundreds of thousands of workers.
On the knowledge-labor side, Anthropic’s Claude Cowork release in February 2026 sent India’s Nifty IT index down nearly 6% in a single trading day, part of a roughly $285 billion global software-stock selloff the press dubbed the “SaaSpocalypse.” India’s index recovered most of that single-session drop by early April, but the structural issues have not improved.
The Fan Sweeps East
I wrote Mushy Potatoes about America’s disruptions from tech and outsourcing. Manufacturing left America through NAFTA, and now, 30 years later, some forecasts estimate that AI is coming for 3.3 million medical coders, claims adjusters, and customer service reps.
In the years America was losing factory jobs, India and the Philippines were gaining office ones. When a mill closed in the US, someone in Dhaka got a shift on the sewing line. When the American radiologist started reading digital scans, a trainee in Bengaluru started interpreting preliminary cases.
The same logic that moved the work built entire national economies on the receiving end. India’s IT and business services sector today employs roughly 5.8 million people and is on track for nearly $300 billion in annual revenue. The Philippine BPO (Business Process Outsourcing) industry employs 1.8 million and generates $38 billion. On top of that, the Gulf labor pipeline moves another 20 million or so workers from South Asia into kitchens, cranes, construction scaffolding, security posts, and domestic households across six petro-states.
Bangladesh negotiated bilateral labor agreements the way other countries negotiate trade treaties. The Philippines has a cabinet-level Department of Migrant Workers. India’s prime minister shows up at the opening of tech campuses the way American presidents used to show up at auto plants.
The people holding those jobs are the Asian equivalents of the mushy potatoes. They did what their governments told them to do. They studied English. They took the vocational course on claims adjudication or medical coding. They paid the recruitment fee. They followed the recipe. And now, in a single spring, both versions of the recipe are going bad.
The Gulf Ledger
Iran started firing missiles into Gulf states in February 2026, part of its war with the US and Israel. By March, at least 17 civilians had been killed by counter-strikes. More than half of them were South Asian migrant workers, because that is who was on the scaffolding, in the kitchen, driving the truck, or standing on the construction site when the strikes came down. Indians, Nepalis, Pakistanis, Bangladeshis. Saleh Ahmed and SM Tareq, both from Bangladesh, were confirmed dead in HRW’s March 31 report. Kuna Khuntia was confirmed by Al Jazeera.
Bangladesh granted 92,460 emigration clearances to Gulf countries in March 2025. In March 2026, that number was 31,279. A 66% drop in a year. Not because demand structurally fell, but because the Bangladeshi government stopped letting new workers ship out. Nepal went further and suspended issuance of labor permits altogether. Gulf firms froze hiring on their end.
The workers already there did not get to choose whether to stay. Many were told to leave when their site closed, their employer disappeared, or their sponsor stopped paying. An unnamed Bangladeshi man in Bahrain told HRW: “Sometimes I do not even have enough food to eat. Even if it is risky, I have to go out and look for work just to arrange food, but I cannot find any work.” The airport was closed. He could not fly home. He had borrowed 20,000 taka (roughly $163) just to survive the week.
Sometimes I do not even have enough food to eat. Even if it is risky, I have to go out and look for work just to arrange food … but I cannot find any work.
The cost of all this is measurable in currency. Bangladesh received $32 billion in remittances in 2025, 6.5% of GDP. About $15 billion of that came from the Gulf. Pakistan got $38 billion, with 96% of its newly registered labor emigrants bound for Gulf Cooperation Council (GCC) countries. Nepal’s 26% GDP remittance share is heavily Gulf-fed, with Malaysia and India filling most of the rest. Capital Economics estimates a 1-2% GDP contraction in the GCC translates to a 5% drop in money sent home. Earlier this month, the World Bank warned of exactly that scale of contraction this year.
The World Bank cut its Pakistan growth forecast on April 10 to 3%, naming the Middle East war as the driver. The current account deficit could widen by $4.9 billion on fuel imports alone, because Brent crude has spiked above $100 a barrel since the conflict began, peaking near $112 in April 2026, up as much as 60% from pre-conflict levels.
So these countries are being hit twice by the same event. Energy prices up, because the Gulf is burning. Remittances down, because the Gulf is not hiring.
The World Bank calls it a “double shock.” Selim Raihan, the Dhaka University economist and head of the think tank SANEM, put it more directly: “This is not only a migration story. It is also an energy story, a trade story, and a macroeconomic story. That combination is what makes the present crisis more dangerous than a routine overseas labor shock.”
The recipe for escape is supposed to be labor diplomacy: opening new destinations to replace the closed ones. On April 9, 2026, Malaysia reopened its labor market to Bangladeshi workers after a two-year closure. The bilateral agreement promises anti-exploitation clauses and a ramp in the number of jobs available, but it is fundamentally a marginal patch on a Gulf-sized hole.
Japan and South Korea are two other potential destinations. Both have aging populations, labor shortages, and the theoretical capacity to absorb a lot of Asian migrants. But the political reality is hostile. Recent protests outside mosques in Japan highlight a rising nativist sentiment that makes large-scale South Asian migration politically toxic. Migrants make up around 3% of Japan’s population and 5% of South Korea’s. The OECD average is 15%. The GCC runs around 50%.
So the near-term answer for the Gulf collapse is: Malaysia for a fraction of the Bangladeshis, and a lot of people waiting it out at home.
The Bangalore Ledger
While the Gulf collapses with a bang, the white-collar labor export market is dissolving behind the scenes. The US press is obsessed with two domestic tech stories: how generative AI disrupts white-collar work, and the need for / impact of large numbers of H-1B visas.
In India and the Philipines, Western companies have spent the last twenty years outsourcing major corporate functions. Now that document formatting, preliminary accounting, and customer service can be automated, these same companies are considering is time to bring those functions back inhouse.
Ishan Talathi, co-founder of the Pune-based cloud firm CloudPe, told Rest of World: “The Indian IT model is built on man-day billing. We charge for bodies on projects. That model is now facing existential pressure.”
The Indian IT model is built on man-day billing. We charge for bodies on projects. That model is now facing existential pressure. - Ishan Talathi
Wall Street has been repricing Indian IT for about a year. Morgan Stanley downgraded Infosys in March 2025, the first-mover call on the sector. Others cut price targets by up to 44% and calling AI’s effect on IT services revenue “deflationary.” India’s own planning body, NITI Aayog, is projecting tech-services headcount falling from 7.5 to 8 million today to 6 million by 2031.
In September 2025, Accenture’s CEO Julie Sweet announced the company would cut 11,000 jobs, backed by an $865 million restructuring, and told CNBC they plan to exit people unable to up-skill. TCS, the largest Indian IT firm, reduced its headcount by 19,755 in a single quarter, Q2 FY26, through a mix of contract non-renewals and layoffs. The NITES union, representing Indian IT workers, filed a complaint with India’s Labour Minister calling the cuts “inhumane.”
In the Philippines, the International Labour Organization has estimated that 89% of the country’s BPO workforce is at “high risk of automation.” The IMF’s own working paper from February 2025 on Philippine labor market AI exposure reaches a similar conclusion with more measured hedges. The IT and Business Process Association of the Philippines (IBPAP), the industry body, maintains officially that the sector is still growing. They report 1.9 million workers and $40 billion in revenue by end-2025, both records. Both can be true. The sector is adding non-voice and AI-augmentation roles. It is also staring down the barrel of voice-services automation, which is where the majority of Filipino BPO workers sit.
Renso Bajala was a Concentrix agent in the Philippines. He spoke to Rest of World about AI-driven productivity monitoring at his call center. He said, on the record, “Someday, AI could just replace us.” A few days after that quote was published, Concentrix fired him for violating the “code of discipline and ethics.” A new Philippine worker coalition, CODE AI, formed around his firing. Its convenor, Lean Porquia, put it plainly: “We can’t say anything bad. BPOs are prepping for AI implementation, hiding exploitation.”
Meanwhile, Teleperformance’s deputy CEO, Thomas Mackenbrock, whose firm employs 60,000 people in the Philippines alone, was asked directly in November 2025 whether AI would reduce headcount. He refused to answer. He said it would happen “step by step, process by process, structure by structure.”
The automation pipe has also reached creative exports. On March 5, 2026, Netflix acquired InterPositive, an AI post-production startup that automates color grading, relighting, and frame-by-frame effects. The exposed workers are not only in Hollywood. They are rotoscoping crews in India, where more than 90% of Hollywood rotoscoping is done.
Why 2008 Held and 2026 Won’t
The historical counterargument is straightforward: the Gulf labor market has survived shocks before. The 1990 Gulf War, the 2008 financial crisis, the 2014-16 oil collapse (Brent $115 to $27), COVID in 2020. Each time, hundreds of thousands of South Asian workers came home. Each time, they went back.
The difference is that in every prior shock, one of the two income streams kept flowing while the other was hit. In 1990 and 2014-16, Gulf remittances tanked and services exports stayed strong. In 2008, both wobbled, but India’s IT sector held up because services demand is less cyclical than goods demand. The labor-export countries always had one working stream.
That is not the case in 2026. The Gulf is disrupted by a war with no clear end. The IT stream is being repriced by a technology that has been improving on a clear curve for three years.
The academic work by S. Irudaya Rajan and D. Narayana on the 2008 GFC and South Asian labor export made this point before anyone was thinking about AI: the resilience of the system depended on having a counter-cyclical backstop. When Gulf construction collapsed in 2008, Bangalore kept coding. When Bangalore gets hit in 2026, there is no construction boom in Dubai to absorb the redirected workers.
In 2008, Indian IT services absorbed labor because they were selling human hours, and human hours scaled with demand. When a US bank needed 500 more people to clean up loan paperwork, TCS hired them. Services were counter-cyclical because they were human: labor-intensive, available on demand. In 2026, that reverses. When a US bank needs 500 more people to clean up loan paperwork, it deploys Claude Cowork. The companies still have demand, but they no longer rely on people to get the work done.
The question these labor exporting countries have to ask themselves is how to survive two simultaneous assaults on the same revenue model.
The Rebranding Era
Historically, every shocked economy rebrands. TCS’s CEO, K. Krithivasan, has publicly urged “bold AI adoption,” which is a polite way of saying the same firm that is cutting 20,000 jobs a quarter is going to be the one selling you the AI that replaces the next 20,000. HCLTech’s CEO has publicly acknowledged that AI will cause “2 to 3% annual deflation” in IT services, modeled into company projections.
On the Gulf side, the rebranding is labor diplomacy: Bangladesh to Malaysia, Pakistan to Europe, India to Japan and South Korea. The math is off by an order of magnitude. Germany’s 400,000-a-year skilled-labor shortfall is the largest single European opening, and it does not replace 5% of what Saudi Arabia and the UAE used to absorb. The Gulf placed a million workers annually at its peak.
Malaysia’s April 9 agreement with Bangladesh places tens of thousands of workers in year one. The credible absorption ceiling across all non-Gulf destinations is in the low millions over five years, against a displaced Gulf worker base closer to 20 million. There’s no clear plan for what to do for the millions of workers who no longer have any hope of having a job.
Option one is to stop betting on labor. India has a five-year-old semiconductor story that needs infrastructure still decades out. It has a potential pharmaceutical-manufacturing story, but that has been bogged down in shady practices and corruption.
The Philippines has financial services and creative industries, both still too small to absorb much of the potential BPO layoffs.
Option two is to bet on labor differently: move Bangladesh’s blue-collar Gulf pipeline into white-collar East Asian professional migration. This is the bet Bangladesh is making. But exporting people needs a willing host, and Japan and South Korea are resisting.
Option three is to let the workers find their own way and hope nothing breaks. Labor-sending governments have been quietly running this for years: tolerating recruitment fraud, under-regulating the abusive sponsorship systems, allowing informal migration. It is the cheapest option and the default. It is also fraught with danger for the governments, as both Bangladesh and Nepal suffered major riots over young people’s frustration over their lack of a future.
A country that built 6.5% of its GDP on money sent home from the Gulf cannot replace that flow with labor diplomacy in 18 months. A country that built nearly $300 billion of its services economy on English-speaking white-collar arbitrage cannot replace it with semiconductor plants in five years.
The strongest version of the bull case goes the other way: AI expands the services market rather than shrinking it. Nandan Nilekani, co-founder of Infosys, has been the public face of this argument. His version: cheaper agentic work lets Western firms afford more outsourcing, not less, and India’s scale and engineer density make it the default beneficiary. The pie grows, India’s slice grows with it, and the whole industry moves up the value chain into agent orchestration, AI model customization, and enterprise AI deployment.
There is evidence. TCS’s AI revenue is growing 16% quarter-on-quarter off a small base. Infosys and Wipro are winning AI-platform implementation deals that did not exist five years ago. Enterprise generative-AI spending is rising fast, and India has the largest skills pool at the lowest price for the systems work.
The problem is distribution. The expansion is selling a different product to a different buyer. Enterprise AI platform deployment is sold to a Fortune 500 CTO by a two-hundred-person team of principal architects, not by the 5.8 million contract workers currently doing claims adjudication, document processing, and customer voice support. Nilekani’s bull case is correct about the market. It is silent about what happens to the workers the old market built.
The old services business had economics working in its favor: human labor could not be sped up without hiring more humans, so pricing stayed sticky and headcount stayed sticky. AI breaks that. The same contract gets served by a fraction of the people at a lower unit price. Revenue grows, margins grow, shareholder returns grow, and nothing in that sequence requires the workers to keep their seats.
What the Rotten Rice Crowd Actually Faces
Market expansion does not pay off recruitment debt. Under these macro numbers are people, and mass displacement of millions of them sparks regional instability, forced migration, and border crises. Wars in the Middle East and North Africa pushed waves of refugees outward over the last decade. The economic collapse of the world’s outsourcing hubs will trigger its own wave, of informal migration chasing income.
Behind every one of those numbers is a household that borrowed money to place the worker. This is not a formal bank loan. The recruitment fees often come from local informal lenders at staggering interest rates, sometimes 3% to 5% a month. When the remittance check stops, the debt balloons. If the family cannot pay, the lender seizes whatever collateral they put up: the family home, the farm, the land. In the most vulnerable communities, debt collection turns to physical violence.
Research on Filipino migrants by Gaurav Khanna at UC San Diego and co-authors found that remittances are often used to pay for the next generation’s schooling, which is how families climb out of low-productivity farm work into higher-paying jobs. When remittances drop, kids pull out of school. That damage to the next generation’s earning power lasts decades, long after the crisis that caused it ends.
On the IT side, the Indian “big fat middle layer” of employees with 13 to 25 years of experience, is the group AI coding agents wipe out first. These are people who came up in the BPO era, who have mortgages on homes they bought when the sector was booming, who have kids they put through school on the assumption that the industry they worked in would be there in 2035. As a group they are not all going to retrain into AI implementation at 45.
If you are 40 in Bangalore, the news reads differently than it does from a West Coast AI office. Your parents sent you to engineering school because it was the path out. You took it. You bought a house in a gated community with a good school nearby and sent your kids to it. Now a CEO in San Francisco who has never met you is publicly explaining how your job will be automated, while your own CEO is promising shareholders the same thing on an earnings call.
Below them, a new generation of Indian IT graduates enters the job market every year. The sector used to absorb them. This year, it is shedding the experienced workers they were supposed to learn from.
The mushy potatoes in Danville did everything the system told them to do. The rotten rice in Dhaka, Kathmandu, Manila, and Mumbai did too. The American version got NAFTA, Trade Adjustment Assistance, and a promise of retraining that never materialized. The Asian version got the Gulf and the BPO boom, and a promise of upward mobility now running out on two ends.
The difference is scale and speed. NAFTA displaced 900,000 certified workers over decades. The combined Gulf pullback and IT/BPO AI compression potentially touches 25 million people over a handful of years. There is no Trade Adjustment Assistance for a Nepali construction worker in Dubai or a Filipino call center agent in Cebu. The governments have budgets for a few hundred thousand of their own, not the entire diaspora.
The Question the West Won’t Ask
There is no global safety net for them, and the architects of the automation are not factoring them into the equation. When Jamie Dimon and Larry Fink talk about AI displacement at Davos, they are thinking about the 3.3 million American office workers in the Mushy Potatoes piece. Their proposed solutions (phased retraining, pension investments, vocational pivots) already read like a modern “let them eat cake” to the domestic workers they name. For the 25 million on the other side of the world, these executives are not even pretending to have a plan.
India is not having a Davos conversation about AI displacement. It is having an earnings call conversation about AI transformation, which sounds like the same thing and is not. The Philippines is not having a national plan for the risk automation poses to their country. Bangladesh is not having a macroeconomic transition plan for a 6.5% GDP contraction in remittance-led consumption.
So the version of the question Mushy Potatoes ended on (what does it mean to be useful when the machines don’t need you?) has a harder version for the rotten rice crowd.
What does economic purpose look like for a country that built itself on exporting people, when the two markets it exported them to are both closing at the same time?
Bangladesh, Nepal, Pakistan, India, and the Philippines do not get to answer this the way Denmark or Germany might. They do not have fiscal room for a ten-year UBI experiment. They do not have the industrial base to absorb returnees into domestic manufacturing. They do not have the political stability to survive years of falling real wages and rising youth unemployment. They have cities full of people whose parents bet the family savings on a recruitment fee or a BPO-track engineering degree, and those bets are going bad.
Renso Bajala’s union organizers in Manila do not have a retraining budget. They have a Philippine Congress that has not yet agreed to debate an AI Regulation Act. The Bangladeshi man in Bahrain without enough food has a closed airport and a broken contract.
Multiply those by 25 million.
The mushy potatoes got 25 years of deaths of despair and Trade Adjustment Assistance that mostly did not work. The architects of that system sat in Washington, comfortably insulated from the consequences in Danville. Today, Sam Altman, Dario Amodei, and Sundar Pichai are building tools that are vastly more capable, with vastly worse instructions for the humans they displace.
Back in Odisha, Jaya Khuntia still owes 300,000 rupees. The debt he took on to send his son to Doha keeps compounding. His son will never call again.