The Purchasing Power Problem AI Is Creating

The Warning Nobody Wants to Hear

Sridhar Vembu, founder of Zoho, put forward a thesis that made many in the technology industry uncomfortable: the rush to deploy artificial intelligence at scale is engineering a purchasing power crisis. His argument was not the usual hand-wringing over automation displacing workers. It was more structural. By eliminating entry-level roles - the very positions that historically built middle-class consumer bases - AI adoption threatens to hollow out the demand side of the economy. Fewer people earning wages means fewer people buying products, subscribing to services, and sustaining the growth that justifies the capital being poured into AI in the first place.

The response was predictable. Commentators pivoted to debating whether software engineers would lose their jobs or whether new roles would emerge to replace old ones. That misses the point. Vembu was not describing a labor market disruption that would self-correct over time. He was describing a feedback loop in which the productivity gains from AI undermine the conditions necessary for those gains to translate into broader prosperity.

This is not a hypothetical concern for Asia. The region's growth model over the past three decades has relied heavily on expanding workforces, rising incomes, and the creation of new consumer classes. From Bangalore to Jakarta, the pattern has been consistent: young people enter the workforce in entry-level roles, gain skills and income, and begin participating in the formal economy as consumers. That cycle is now under threat, and the investment decisions being made today - by venture capital firms, corporate treasuries, and sovereign wealth funds - will determine how severe the disruption becomes.

Why Productivity Alone Is Not Enough

The conventional wisdom in technology circles is that productivity improvements always generate net benefits. Automation makes processes cheaper and faster, freeing up resources for new ventures and creating space for higher-value work. This logic has held for decades, but it assumes that displaced workers transition smoothly into new roles and that aggregate demand remains stable or grows.

AI challenges both assumptions. Unlike previous waves of automation, which largely affected manufacturing and routine clerical work, AI is now encroaching on knowledge work, creative tasks, and even interpersonal services. The jobs being automated today are not just assembly-line positions. They are the entry points for educated workers in software development, customer service, content creation, financial analysis, and legal research. These are precisely the roles that have fueled the rise of Asia's urban middle class.

When a corporation replaces ten junior analysts with an AI tool, the immediate effect is a reduction in payroll costs and an increase in operational efficiency. The second-order effect is that ten fewer households have disposable income. Multiply that across thousands of companies and millions of roles, and the macroeconomic impact becomes visible. Consumer spending softens. Retail sales stagnate. Demand for housing, education, and discretionary goods weakens. The productivity gain at the firm level turns into a demand shortfall at the economy level.

This is not a new phenomenon. The 2008 financial crisis illustrated how income concentration and weak wage growth can undermine consumption, even when asset prices and corporate profits are rising. The difference now is that AI is accelerating the trend. Companies can scale revenue without proportionally scaling headcount. Investors reward firms that demonstrate margin expansion through automation. The incentive structure is clear: minimize labor costs, maximize returns to capital.

The Asia Angle: Where Growth Met Employment

Asia's economic transformation has been inseparable from job creation. China's manufacturing boom employed hundreds of millions. India's IT services sector created pathways into the global economy for an entire generation. Southeast Asia's digital economy has been a rare bright spot for youth employment in recent years. In each case, the expansion of opportunity was not a byproduct of growth - it was the mechanism of growth.

AI threatens to sever that link. Consider India's IT services industry, which employs roughly five million people and generates more than $250 billion in annual revenue. A significant portion of that workforce consists of entry-level developers, testers, and support engineers - roles that are increasingly being automated or augmented by AI tools. Companies like Infosys, TCS, and Wipro are already shifting hiring patterns, emphasizing niche skills over volume. The result is a tightening funnel: fewer entry points, higher barriers to entry, and a growing cohort of young graduates who cannot access the careers their predecessors took for granted.

The same dynamic is playing out in Southeast Asia's e-commerce and fintech sectors. Customer service roles, once a staple of the gig economy, are being replaced by chatbots and virtual assistants. Content moderation, data labeling, and routine compliance tasks are being automated. These jobs were never glamorous, but they provided income and experience. Their disappearance is not being offset by equivalent opportunities elsewhere.

Even in manufacturing, where automation has been a constant for decades, the shift is different this time. Advanced robotics and AI-driven quality control are reducing the need for human labor in ways that previous generations of machinery did not. Factories in Vietnam and Bangladesh that once employed thousands are now designing facilities that require hundreds. The jobs are not moving to another country. They are simply vanishing.

Investment Patterns That Reinforce the Problem

The venture capital and private equity communities in Asia are amplifying this trend. Investors are pouring capital into AI startups and automation platforms, often with the explicit goal of reducing labor costs for clients. Pitches emphasize margin improvement, scalability, and the ability to replace human functions with software. These are rational investment theses from a financial perspective. They generate attractive returns on paper.

But they also reflect a narrow definition of value creation. The focus is on efficiency within the firm, not resilience within the economy. A portfolio company that cuts its workforce by 30 percent while doubling revenue is celebrated as a success story. The broader question - what happens when every portfolio company pursues the same strategy - is rarely asked.

This is not an argument against automation or AI. It is an argument for a more complete accounting of the trade-offs involved. Investors and executives are optimizing for shareholder returns in a system that assumes demand will remain constant. That assumption is increasingly fragile. If large segments of the population are shut out of stable employment, they cannot sustain the consumption patterns that drive corporate revenue. The efficiency gains become self-defeating.

What a Different Approach Would Look Like

Rethinking AI investment does not mean rejecting the technology. It means expanding the criteria by which success is measured. Companies and investors should ask not only whether AI improves productivity but also whether it preserves or expands access to economic participation.

One model is to prioritize augmentation over replacement. AI tools that enhance the capabilities of existing workers - enabling them to do more complex or creative work - generate productivity gains without eliminating roles. This requires different product design choices and different incentives for management teams. It also requires investors to reward companies that maintain or grow employment alongside revenue.

Another approach is to channel capital into sectors where AI can create new forms of demand rather than simply reducing costs in existing markets. Healthcare, education, and environmental services are all areas where automation could expand access and improve outcomes, rather than displacing labor. These sectors are less attractive to venture investors because they often involve lower margins and longer timelines. But they are precisely the areas where AI could generate broad-based benefits.

Governments in Asia also have a role to play. Tax policy, labor regulations, and industrial strategy can all be adjusted to encourage investment patterns that balance efficiency with employment. Singapore, for instance, has experimented with wage subsidies and training programs designed to help workers transition into AI-adjacent roles. South Korea has floated proposals for taxing automation in sectors where displacement is acute. These are incremental steps, but they signal an awareness that market incentives alone will not solve the purchasing power problem.

The Risk of Ignoring the Feedback Loop

If the current trajectory continues, the likely outcome is not mass unemployment in the traditional sense. It is underemployment, wage stagnation, and a bifurcated economy in which a small cohort of high earners coexists with a large population of gig workers, informal laborers, and people cycling in and out of precarious roles. That structure is already visible in parts of Asia. AI will accelerate it.

The political and social consequences are harder to predict but easier to imagine. Populist movements, labor unrest, and demands for redistribution are all plausible responses to an economy that delivers productivity gains to capital while eroding stability for labor. These dynamics are not unique to Asia, but they are particularly acute in a region where social contracts are still being negotiated and where large youth populations have high expectations for upward mobility.

The irony is that the technology itself is not the problem. AI has the potential to solve enormous challenges, from climate modeling to drug discovery to infrastructure optimization. The issue is how investment decisions are shaping the deployment of that technology. When the primary use case is cost reduction through labor displacement, the benefits accrue narrowly and the risks accumulate broadly.

A Question of Priorities

Vembu's warning was not a call to halt AI development. It was a call to think more carefully about the second-order effects of the choices being made in boardrooms and investment committees across Asia. The purchasing power problem is not inevitable. It is the result of specific decisions about what to build, who to hire, and where to allocate capital.

The technology industry has spent years talking about scale, disruption, and exponential growth. It has spent far less time talking about sustainability, inclusion, and the distribution of gains. That imbalance is becoming harder to ignore. If AI investment continues to prioritize efficiency over access, the region risks engineering a crisis of its own making - one in which rising productivity coexists with falling demand, and in which the growth model that lifted hundreds of millions out of poverty begins to work in reverse.

The question is whether investors, executives, and policymakers will adjust course before that feedback loop becomes entrenched. The answer will shape not only the future of work in Asia but the viability of the economic model that has defined the region's rise.