Jones argues that the central risk in the AI buildout is not simply high technology valuations. It is the way enormous data-centre and chip commitments are being financed before long-term demand and revenue are certain. Large platform companies can fund much of this expansion from cash flow, but other participants increasingly rely on loans, bonds and outside capital.
He describes special-purpose financing structures that separate data-centre assets and their debt from the operating company. Long contracts and purchase commitments can make these projects appear predictable enough to finance, while still leaving open questions about utilisation, customer concentration and what happens if computing demand grows more slowly than expected.
The consequences can extend beyond technology shareholders because the debt may be held through banks, insurers, credit funds and retirement portfolios. Jones recommends watching credit spreads, refinancing conditions, construction commitments and actual utilisation rather than treating share prices alone as the measure of whether the AI investment cycle is healthy.
Watch on YouTube
