Artificial intelligence (AI) has long been heralded as a transformative force, promising a new era of productivity growth unburdened by the traditional constraints of labor costs. While the full realization of this promise is still unfolding, its impact is becoming increasingly evident across various sectors. In the realm of software development, for instance, AI tools are demonstrating a remarkable capacity to generate code, significantly enhancing individual worker output. This surge in AI-driven efficiency has already begun to ripple through equity markets, as evidenced by recent sell-offs in some software stocks. The pervasive adoption of AI is no longer a hypothetical scenario; a comprehensive study by KPMG released in late 2025 revealed that a staggering 93% of Canadian businesses are actively exploring AI integration, with 31% having already deployed AI tools within their core operations. As AI solidifies its position as a primary engine of productivity, a critical question emerges for investors and economists alike: how do we accurately measure its costs and economic contributions?
The escalating investment in AI services, aimed squarely at boosting productivity, suggests that the cost of these services may soon rival, if not surpass, traditional economic indicators like the price of oil or the cost of labor. In this evolving economic landscape, "tokens" are emerging as a potential unit of account for the cost of AI compute power, a fundamental input for AI operations. Companies at the forefront of AI service provision, such as Anthropic and OpenAI, primarily bill their clients using tokens. Essentially, a token represents a unit of computational power, encompassing all the necessary resources and inputs required to process a user’s prompt or execute a given task. While tokens, in their current form, may present an imperfect metric, the global economic implications are already being recognized. Notably, Reuters has reported that China is actively establishing the framework for a token futures exchange in Shanghai, signaling a strategic move to formalize and potentially leverage this nascent market.
This development has prompted considerable discussion among financial professionals. Craig Jerusalim, Senior Portfolio Manager at CIBC Global Asset Management in Toronto, anticipates a significant financialization of this emerging market. "They’re absolutely going to become a financial asset and there’ll be futures traded on it and options and gamemanship and everything in between. The good, the bad and the ugly of finance," Jerusalim stated. He further commented on the potential for innovation, provided it remains within reasonable bounds. "And I think that also spurs innovation as long as it doesn’t go crazy or overboard. But I don’t think that the financialization of token infrastructure is an important development because what you’ve seen is that for other technologies the ramp up curve from early adopters to the mass markets can take decades at times." His perspective highlights the potential for a robust derivatives market to develop around AI tokens, mirroring the evolution of other technological commodities.
The Imperfect Measure: Challenges in Token Valuation
Despite the growing interest, significant hurdles remain in establishing AI tokens as a reliable and universally applicable economic metric, akin to a commodity like oil. Elliot Johnson, Chief Investment Officer at Evolve ETFs in Toronto, points out several key limitations. "Unlike oil consumption, token usage is unpredictable and only calculated after a task has been executed," Johnson explained. He elaborated that tokens currently lack the inherent physical scarcity that underpins the valuation of traditional commodities. For instance, a user might unexpectedly exhaust their daily token allocation by mid-morning when undertaking complex tasks, leading to unforeseen costs. Furthermore, tokens are not fungible in their current state. An individual with unused tokens from an Anthropic Claude account, for example, cannot resell them back to Anthropic or utilize them to run tasks on a competing platform like OpenAI’s ChatGPT. This lack of interoperability and fungibility creates a fragmented market and complicates their adoption as a standardized economic input.
However, efforts are underway to address these very issues, aiming to bring greater tangibility and predictability to token usage. The integration of what are sometimes referred to as "harness" or "orchestrator layers" into AI models is a significant step in this direction. These layers act as intermediaries, communicating with users and intelligently distributing tasks to various AI tools. Johnson suggests that while these orchestrator layers themselves might not incur additional costs if run on existing computing infrastructure, they hold the potential to implement metered billing models. Such models would enable users to precisely factor in token usage before initiating a task, thereby enhancing cost transparency and predictability.
Analogies for the AI Economy: Beyond Crude Oil
Given these complexities, Johnson proposes a more fitting analogy for AI tokens than crude oil. He suggests comparing them to utilities like natural gas or electricity, measured in cubic meters or kilowatt-hours, respectively. While the pricing of these utility inputs is also influenced by market forces, they may not possess the same speculative investment case as raw commodities. This analogy implies that investors seeking to capitalize on AI’s productivity gains should focus on the underlying infrastructure and essential services that enable AI compute, rather than solely on the abstract token itself. The crucial question for investors, therefore, becomes identifying the key inputs that will drive AI productivity and determining the most effective avenues for investment.
"If I could answer that question, I’d probably be on a beach somewhere because I think it is the question," Johnson admitted when asked about identifying these core AI inputs. "It’s confounded by so many challenges. How much venture money is being spent to keep the cost of compute at a certain level rather than another level? How much is debt fueled? Where is the business case coming, coming from? But some of these questions are starting to get answered." His candid assessment underscores the intricate and multifaceted nature of valuing and investing in the AI ecosystem. The current landscape is influenced by substantial venture capital funding, debt financing, and the evolving business models of AI providers, all of which contribute to the complexity of identifying predictable investment opportunities.
Navigating the Investment Landscape: Key Inputs for AI’s Productivity Engine
For retail investors and financial advisors, Johnson maintains that identifying and investing in AI infrastructure bottlenecks remains the most prudent strategy for gaining exposure to the current stage of the AI cycle. He highlights three critical areas that are consistently in high demand and short supply: electricity, data centers, and semiconductors. Within these broad categories, significant diversification has occurred. Johnson notes that the initial scarcity of AI-specific chips was largely confined to Nvidia GPUs. However, this scarcity has since expanded to include central processing units (CPUs) and memory chips, indicating a broadening demand across the entire semiconductor supply chain.
As the AI landscape continues to evolve at an unprecedented pace, Johnson emphasizes the paramount importance of diversification as a strategy for accessing AI-related investment opportunities. He argues that a broad exposure to companies operating within the AI ecosystem is likely to yield exposure to emerging winners, particularly as many prominent private AI firms are poised for initial public offerings (IPOs). As financial markets gradually, and with a degree of deliberation, begin to price AI as a fundamental economic input, Johnson believes that advisors and their clients are keen to maintain a presence in this dynamic sector.
"I think if you’re going to say, ‘I’m going to skip this as a major main allocation of my portfolio,’ I think you have to have strong conviction that this is not going to work out, that AI is going to go away or become less relevant in the future. I certainly don’t feel that way," Johnson asserted. He elaborated on his continued astonishment at the rapid advancements in AI capabilities, stating, "I just feel like every few months I have this other moment of amazement because the AI tools I’m using do something that I didn’t know they could do before. I just feel like that hasn’t run out yet. We’re not yet at the point where it’s mundane." This sentiment reflects a broader industry view that the AI revolution is still in its early stages, with substantial room for continued innovation and growth, making strategic investment essential for those looking to benefit from its transformative potential. The ongoing integration of AI into business processes and everyday life suggests that its economic impact will only continue to grow, necessitating a deeper understanding of its underlying infrastructure and the evolving metrics used to quantify its value.
