Beyond Subsidized Growth: The Structural Fault Lines in AI's Computing Power Race

Paolo Ardoino, CEO of Tether, recently cast a critical eye on a dominant growth strategy within the artificial intelligence sector. He highlighted a widespread practice among major tech firms: heavily subsidizing computing power to rapidly onboard users and capture market share. While effective for scaling in the short term, this approach is fundamentally fueled by relentless and massive capital expenditure.

The Unsustainable Cycle: High Capex Meets Rapid Obsolescence

The crux of the issue lies in the financial lifecycle of the underlying infrastructure. The specialized hardware—like advanced AI chips and server farms—required to deliver this subsidized compute is extraordinarily costly. However, it faces accelerated technological obsolescence, leading to a brutally short financial depreciation schedule of just three to five years. Companies are thus pressured to generate sufficient returns from these assets within a narrow window to justify their cost and fund the next investment cycle.

Accumulating Risk: A Web of Structural Mismatches

Ardoino argues this model is fostering a dangerous convergence of several structural mismatches:

  • Token Price vs. Fundamental Value: Market valuations or token prices may inflate based on user growth metrics, while inadequately pricing in the realities of swift asset depreciation and long-term profit uncertainty.
  • Profit Cycles vs. Investment Cycles: There is a tension between the long timeline needed to monetize massive capital investments and the urgent financial demands created by fast-depreciating assets.
  • Cost of Capital vs. Debt Maturity: Financing huge capital outlays often requires short-term debt or funding, but project payback periods are long, creating potential liquidity crunches.
  • Open-Source Erosion of Revenue: The rising quality and accessibility of open-source AI models continuously compress the pricing power and monetization potential of user bases acquired through subsidies.

In Ardoino's view, the simultaneous pressure from these misalignments is increasing systemic vulnerability across the industry. His warning is succinct: “In a structure like this, something is likely to break.” This critique extends beyond a single business tactic, prompting a necessary conversation about how the AI industry can pursue growth without compromising long-term financial stability.