With frontier models costing over 100x more than competent alternatives ($56 vs. 50¢ per million tokens), companies are burning cash. An estimated 98% of tasks sent to top-tier models don't require that power, an inefficiency driven by engineers who are disconnected from cost implications.

Frontier AI labs are knowingly losing vast sums on token-based services, a classic "J Curve" strategy to achieve mass adoption first and profit later, mirroring Uber's early ride subsidies. However, tokens may ultimately become a commodity like bandwidth, making this a risky long-term bet.

Newer AI models may have low per-token prices but are often "token hungry," requiring more tokens to complete a task. This can make them more expensive overall. The true measure of economic viability is the final cost-per-task, not the misleading per-token price.

Unlike traditional SaaS, achieving product-market fit in AI is not enough for survival. The high and variable costs of model inference mean that as usage grows, companies can scale directly into unprofitability. This makes developing cost-efficient infrastructure a critical moat and survival strategy, not just an optimization.

The paradoxical financial state of AI labs: individual models can generate healthy gross margins from inference, but the parent company operates at a loss. This is due to the massive, exponentially increasing R&D costs required to train the next, more powerful model.

As AI token consumption becomes a major budget item, companies are moving beyond using a single frontier model. Every organization will need a portfolio of models, including cheaper options for less complex tasks, to manage the "madness" of runaway costs.

The hedge fund Citadel Securities observes that the AI market is splitting. After initial enthusiasm, companies are now facing the reality of high token costs and compute constraints, causing a shift away from expensive frontier models toward simpler, more cost-effective AI that offers clearer ROI.

Despite billions in funding, large AI models face a difficult path to profitability. The immense training cost is undercut by competitors creating similar models for a fraction of the price and, more critically, the ability for others to reverse-engineer and extract the weights from existing models, eroding any competitive moat.

While hardware gets cheaper (Moore's Law), the competitive pressure to release superior AI models leads to exponentially larger and more complex systems. This results in a higher number of "tokens burned" per query, making the cost of delivering a useful answer actually increase with each new generation.