The initial projection for artificial intelligence (AI) suggested that enterprises would significantly reduce operational expenses by implementing AI solutions. This was largely based on two assumptions: first, that AI would offer a lower cost alternative to human labour, and second, that it would potentially enhance productivity by tenfold. However, by 2026, the scenario has shifted quite drastically, revealing that the expenses associated with AI deployment have escalated instead of the anticipated productivity benefits.

Reports indicate that some industry experts are now advocating for the employment of skilled human workers over AI solutions, such as Claude, due to the increased costs of AI tools. Major corporations like Microsoft and Uber are struggling to justify their financial commitments towards AI technologies amidst these rising costs.

Last week, various reports highlighted the increasing disconnect between the costs incurred from AI and the productivity outcomes. At Microsoft, thousands of engineers have been instructed to discontinue the use of Claude Code, transitioning instead to an internal AI platform. The deadline for this switch is set for June 30. Although Microsoft has yet to officially state the reasons for this abrupt change, insiders suggest it relates to the excessive expenditure on Claude tokens that has not translated into visible productivity improvements.

Uber’s Experience with AI Expenses

Similarly, Uber allocated a budget specifically for AI-related expenses, initially introducing access to Claude Code for its 5,000 engineers in January. However, within just five months, the company reported that it had entirely depleted its annual AI budget. This was confirmed by Uber’s Chief Operating Officer, Andrew Macdonald, who elucidated that the utilisation of AI has not yielded any increases in productivity thus far.

Macdonald raised concerns regarding the rising costs, stating during the Rapid Response podcast, “We’re going to have to start talking about token consumption and the associated costs versus headcount.” He emphasised the challenge of correlating AI expenditures with the functional benefits delivered to users, leading to difficulty in justifying continued investment in these tools.

Former Infosys CEO Vishal Sikka has voiced similar concerns about the situation, noting on social media that the issue of token costs is becoming increasingly prominent and that greater scrutiny of AI expenditure is likely in the future. The original hope that AI would serve as a financial remedy is now seen as misguided as businesses grapple with unforeseen costs without a corresponding rise in productivity.

The Shift to Token-Based Pricing Models

As organisations begin to recognise the financial implications of AI, the term ‘tokens’ has emerged as a notable point of discussion. Tokens serve as the units by which AI technologies measure both input and output, akin to how electricity is metered. In 2026, companies like Google, Anthropic, and OpenAI face pressure to produce AI-related revenues, leading them to adopt token-based pricing structures for clients. Previously, these organisations offered subsidised usage; however, this model is shifting as they seek to establish a more sustainable revenue framework.

Reports indicate that the operational costs for platforms like Claude and ChatGPT can range between $1,500 to $5,000, while clients only pay around $200 per month. The transition toward token-based pricing is likely to escalate expenses beyond those of maintaining human staff for many businesses. For example, Nvidia’s Vice President, Bryan Catanzaro, has recently indicated that for his team, the cost of computing exceeds that of employing human workers.

Industry insights suggest that the hike in AI costs does not signify a downturn in AI enthusiasm, but rather calls for a reassessment of AI strategies among companies. Experts advocate for a shift from a narrow focus on cost to a broader evaluation of the value that AI can provide. They highlight the importance of carefully managing the volume of AI interactions and scrutinising resource consumption to ensure that AI adoption is sustainable and justified.

Long or Short, get news the way you like. No ads. No redirections. Download Newspin and Stay Alert, The CSR Journal Mobile app, for fast, crisp, clean updates!

App Store –  https://apps.apple.com/in/app/newspin/id6746449540 

Google Play Store – https://play.google.com/store/apps/details?id=com.inventifweb.newspin&pcampaignid=web_share