The AI transition is hitting a double bottleneck of cash and capability
While Morgan Stanley projects strong long-term returns for tech infrastructure, employers are struggling with an unprepared global workforce.

The multi-billion dollar bet on artificial intelligence is entering a cold, calculating new phase. As tech giants brace for a temporary cash-flow squeeze to build out massive infrastructure, employers are facing an equally urgent challenge: a workforce that is largely unprepared to use the tools being built.
Can tech giants justify their massive AI spending?
A new analysis by Morgan Stanley suggests that while the eye-watering capital expenditure by tech giants will temporarily depress their free cash flow, the underlying returns remain highly profitable. The bank tracked the "return on incremental invested capital" (ROIIC)—which measures the profit generated by new investments over a rolling three-year period—for five major tech giants: Amazon, Alphabet, Microsoft, Meta, and Oracle.
Morgan Stanley found that their combined ROIIC peaked at nearly 40 per cent in the first quarter of 2026, easily outpacing their estimated 8 per cent cost of capital. Though the bank projects this return will dip to around 23 per cent by late 2027 as heavy infrastructure spending peaks, it is expected to recover to approximately 35 per cent by 2030.
The short-term pain will be felt in free cash flow. Morgan Stanley estimates that the combined free cash flow of these five companies will bottom out at around -$265 billion in the third quarter of 2027, before rebounding to nearly $505 billion by the end of the decade.
Amazon Chief Executive Andy Jassy recently defended the heavy upfront spending on chips, land, power, and servers, noting that while early cash flow is inevitably squeezed, the returns become highly attractive once those assets are fully deployed and monetised.
The growing AI skills gap in the workforce
Building the infrastructure is only half the battle. The other half is finding people who can run the software. A severe shortage of practical AI talent is threatening to stall these corporate ambitions.
In India, which serves as a critical talent pool for the global technology industry, the skills gap is especially acute. According to data from workforce solutions provider Adecco, roughly 65 per cent of India’s overall skills gap is now concentrated in AI and emerging technologies. Demand for talent in AI, cybersecurity, and digital engineering is growing by 12 to 15 per cent quarter-on-quarter.
Yet a recent study by IBM found that only about 30 per cent of employees currently possess the level of AI literacy that businesses require—a figure that needs to rise to nearly 57 per cent by 2030.
"AI certification and AI capability are two different signals," says Karishma Parikh, vice president of human resources at Adecco India. She argues that the future-ready worker is not necessarily the one who has memorised the most tools, but the one who can apply technology to solve actual workplace problems with sound judgment.
How employers are rushing to retrain workers
To bridge the gap, major technology and consulting firms are rushing to build their own pipelines of qualified workers, often bypassing traditional academic timelines.
IBM has launched a massive initiative to skill five million learners across India in AI, cybersecurity, and quantum computing by 2030, partnering with institutions like Christ (Deemed to be University) in Bengaluru. Similarly, EY India has trained more than 50,000 of its own employees, delivering over two million hours of AI learning, and has partnered with Delhi's Shri Ram College of Commerce to launch a 12-month career-preparation programme called AI VIBE.
Tech companies are also restructuring how they view roles. Ramachandran Sundararajan, Chief People Officer at HCLTech, says the firm has organized its workforce into three distinct cohorts: "AI Builders" who develop the platforms, "AI Super Users" who use the tools to enhance everyday productivity, and "Human-in-the-Loop AI Decision Makers" who apply human judgment to AI-generated results.
Ultimately, the success of the massive financial investments in AI will depend heavily on whether these skilling initiatives can keep pace with the infrastructure being built.
Key numbers
- Nearly 40%
- -$265 billion in Q3 2027
- About 30%
- Nearly 57%
- 65%



