The Hidden Cost of AI: Why Most Companies Are Flying Blind
Artificial intelligence has rapidly become an indispensable tool for businesses worldwide. From automating customer service to generating code, AI tools like Claude Code and OpenAI’s Codex are now woven into the fabric of daily operations. Yet beneath this wave of adoption lies a growing crisis: the cost of AI is spiraling out of control, and most organizations have no idea how much they are actually spending.
A recent report from KPMG has laid bare the scale of the problem. The findings are startling—nearly three-quarters of companies lack a clear understanding of how their AI bills are calculated. What was once seen as a competitive advantage is now quietly draining budgets, forcing even the biggest names in tech to rethink their strategies.
The Alarming Gap in AI Cost Awareness
According to the KPMG survey, only 26 percent of companies have a complete and transparent picture of their AI expenditures. About half of the firms surveyed have some sense of what they are spending, but the full scope remains elusive. More troubling still, 22 percent admitted they have almost no visibility into their AI costs until the bill arrives.
In other words, for every four companies using AI, only one truly understands what it is paying. This lack of awareness is not just an accounting oversight—it is a strategic blind spot that can derail budgets and erode profitability.
Steve Chase, KPMG’s global head of AI, noted that AI has become a resource whose consumption and costs are both accelerating rapidly. Many organizations are simply unable to keep pace with the speed at which expenses are mounting.
Why AI Bills Are So High
The pricing model used by leading AI providers is a key factor behind the soaring costs. Companies like OpenAI and Anthropic charge based on tokens—a unit of measurement that reflects how much work an AI system performs. The more a company uses AI, the more tokens it consumes.
Organizations typically purchase licenses for their employees that include a fixed number of tokens. Once that limit is exceeded, additional fees kick in. And as AI tools become more integrated into workflows, token usage is exploding.
KPMG reports that some companies have burned through their entire annual AI and cloud budget in just a few months. In many cases, token consumption has increased sixfold or more, catching finance teams completely off guard.
Token-Based Pricing: A New Kind of Meter
Unlike traditional software licensing, token-based pricing is dynamic and usage-driven. This means costs can fluctuate wildly from month to month. For businesses accustomed to predictable subscription fees, this model introduces a level of uncertainty that is difficult to manage.
Consider this: a company that deploys AI for customer support, content generation, and data analysis might see its token usage spike during peak seasons. Without real-time monitoring, the resulting bill can be a shock.
Even Industry Giants Are Struggling
The challenge is not limited to small or mid-sized firms. Some of the world’s largest companies are also feeling the pinch. Uber, for example, exhausted its entire annual AI budget in just four months. In response, the ride-hailing giant imposed a strict limit of 1,500 dollars per employee for AI usage.
In another case, a single company reportedly spent around 500 million dollars on Claude AI in just one month. Such figures highlight how quickly costs can escalate when AI usage is left unchecked.
Even Sam Altman, CEO of OpenAI, has acknowledged that AI costs have become a major concern. He noted that many companies are joking that their entire 2026 budget was spent in the first quarter alone. What was once a punchline is now a painful reality.
Walmart and Others Follow Suit
Retail giant Walmart has also taken steps to curb AI spending, joining a growing list of companies that are placing restrictions on usage. These moves signal a broader shift: the AI market is maturing, and with maturity comes a focus on cost control.
Investment Continues Despite Rising Costs
Despite the financial strain, investment in AI infrastructure shows no signs of slowing down. Google recently announced plans to raise 80 billion dollars, with a significant portion earmarked for AI projects. Meanwhile, Anthropic is preparing for an initial public offering, and some analysts predict its valuation could reach one trillion dollars. OpenAI itself is expected to go public in the future.
This paradox—rising costs alongside massive investment—underscores the high stakes involved. Companies are betting heavily on AI’s long-term potential, even as they struggle to manage its short-term expenses.
The ROI Question Looms Large
Experts warn that balancing the benefits of AI against its costs will be the defining challenge for businesses in the years ahead. As AI becomes more deeply embedded in operations, the pressure to demonstrate a clear return on investment will only intensify.
For now, the message from KPMG’s report is clear: companies cannot afford to ignore the cost side of the AI equation. Without better visibility and control, the very technology meant to drive efficiency could end up undermining financial stability.
- Only 26 percent of companies fully understand their AI costs.
- Token-based pricing leads to unpredictable and rapidly growing expenses.
- Major firms like Uber and Walmart are already imposing usage limits.
- Investment in AI infrastructure continues, but ROI remains uncertain.
