‘The cost of compute is far beyond the costs of the employee’: Nvidia executive says right now AI is more expensive than paying human workers
Recent layoffs in the technology sector might suggest that a significant shift from human labor to artificial intelligence is underway. For instance, Meta disclosed plans to reduce its workforce by 10%, equating to approximately 8,000 employees, and to halt recruitment for 6,000 open positions. This was framed as a move toward greater efficiency and to balance other investment priorities. Similarly, Microsoft launched its largest voluntary buyout program to date, aiming to cut thousands of jobs. However, despite these reductions, key industry figures indicate that AI currently does not reduce labor costs; on the contrary, it often entails higher expenses than retaining human staff.
Executives like Nvidia’s vice president of applied deep learning emphasize that the computational expenses associated with AI substantially outweigh employee costs. Supporting this perspective, a 2024 MIT study evaluated AI’s capability to match human performance, particularly in vision-based roles, and concluded that AI automation would be cost-effective in only 23% of such cases, leaving the majority more economically viable when performed by humans. Additionally, practical setbacks exist; one engineer recounted how excessive AI usage led to catastrophic system failures, underscoring the technology’s current limitations.
Despite the absence of convincing data showing AI improves productivity or displaces labor at scale, major technology corporations continue to heavily invest in this area. Spending on AI capital expenditures has surged by 69% compared to the previous year, reaching $740 billion so far in 2026. This rapid outflow of capital is prompting companies to reassess budgets, as illustrated by Uber’s CTO, who noted the company’s entire AI coding tools budget for 2026 was depleted by April due to aggressive adoption incentives. Microsoft likewise shifted its AI software strategy by discontinuing many Claude Code licenses in favor of GitHub Copilot CLI, a move driven by the unexpectedly fast uptake among staff.
This escalating investment in AI occurs simultaneously with intensified layoffs within the tech industry. Data tracking over 118,000 job cuts in nearly 100 companies for 2026 shows that workforce reductions are accelerating beyond the levels seen in the previous year. This trend reveals a complicated economic picture: while human labor is currently less expensive, the commitment to AI continues to grow, indicating a temporal imbalance in cost dynamics. Experts describe this as a short-term mismatch, shaped by the high operating costs connected to AI’s hardware and energy demands, which are not yet offset by efficiency gains.
Looking ahead, the costs associated with AI are expected to remain substantial in the near term, with projections estimating expenditures could reach $5.2 trillion by 2030 and possibly escalate to $7.9 trillion if adoption accelerates. Meanwhile, increases in AI software fees and fixed subscription models raise additional financial challenges for companies with heavy AI usage. Consequently, some organizations are shifting their perception, viewing AI more as a complementary tool rather than a direct labor replacement, pending more favorable cost structures.
Future signs pointing to a financial tipping point include significant reductions in AI operational expenses, particularly the cost of AI inference in large language models, which analysts predict could drop by over 90% within the next four years. Advancements in AI infrastructure, hardware availability, and pricing strategies are expected to contribute to this shift. Adoption models may transition from flat fees to usage-based pricing to better align with actual consumption. Nevertheless, the economic viability of AI will also hinge on its reliability, reduced error rates, minimized human supervision, and seamless integration into existing workflows. As of late 2025, nearly 18% of businesses had implemented AI tools, marking a large increase within months, although the focus remains on balancing cost reductions with predictability and performance at scale.
