OpenAI, Anthropic push new metrics to gauge AI’s true cost

OpenAI and Anthropic are urging enterprise customers to rethink how they measure the cost of artificial intelligence, according to Bloomberg, as the industry grapples with a fundamental accounting challenge: determining what companies actually get for their AI spending.bloomberg

The Measurement Gap

The push from the two leading AI developers comes as a broader field known as "tokenomics" takes shape. Tokens — the units used to price AI model usage — have become the currency of enterprise AI, but many organizations lack the tools to connect that spending to business outcomes. Howard Rubin, an economist who advises companies on technology spending, told IBTimes that firms are handling AI budgets with little grounding. "It's a currency where you have no instinct to know what you're using, and the accounting practices aren't even there for it," he said.ibtimes

The problem has grown more acute as AI spending accelerates. Goldman Sachs The Goldman Sachs Group, Inc. ↘ 1.38% reports AI-related spending is tracking to exceed $800 billion in 2026, yet only 23% of C-suite leaders surveyed by Accenture ↘ 0.11% said they can measure the return on that investment.accenture

Linux Foundation Enters the Fray

On August 4, the Linux Foundation formally launched the Tokenomics Foundation, backed by more than 30 founding members including IBM International Business Machines Corporation ↗ 0.60%, JPMorgan Chase ↘ 0.73%, Oracle ↘ 3.11%, SAP ↘ 0.03%, ServiceNow ↗ 2.76%, and Broadcom ↘ 2.62%. The initiative aims to build vendor-neutral standards so companies can compare AI providers the way they already compare cloud computing costs.linuxfoundation+2

J.R. Storment, the foundation's executive director, said the goal is to give organizations "better starting places" for decisions they are already making without reliable data. Notably, the founding membership does not include OpenAI, Anthropic, or Microsoft ↗ 0.52%, suggesting the AI developers and the open-source standards body may be pursuing parallel but distinct approaches to the same problem.constellationr+1

Companies Building Their Own Yardsticks

In the absence of industry standards, some firms have improvised. Cybersecurity company Elisity created a metric called "bionic head count," dividing total AI spending by average staff cost and comparing combined human-AI output against revenue. "Essentially, you're adding virtual head count," said Charlie Treadwell, the company's chief marketing officer. "Is that resulting in incremental revenue, or are you just eating at your margins?"ibtimes

Bain & Company analysts have modeled a scenario in which AI tokens eventually make up roughly a quarter of corporate operating expenses, which would force businesses to treat AI spending as part of the labor budget rather than a technology line item.ibtimes