•
•
•
•
•
•
•
Bonds, loans and Gulf money are financing the race for AI chips and data centres
The AI boom is increasingly being paid for with borrowed money. In early October 2026, reports said SpaceX was pursuing about $40 billion to buy Nvidia AI chips (roughly $30 billion in bonds and $10 billion in bank loans). Broadcom and Oracle were also in large financing deals, and SoftBank was seeking up to $100 billion from Gulf investors for a new AI-focused vehicle. Morgan Stanley estimates AI infrastructure could need about $1.5 trillion in external financing by 2028. Supporters see an investment cycle like the building of railways or the internet. Sceptics warn of a possible bubble.
| Company | Reported financing | Purpose |
|---|---|---|
| SpaceX | ~$40B (bonds + bank loans) | Nvidia AI chips |
| SoftBank | Up to $100B sought from Gulf (mainly UAE) investors; earlier $11.1B bond sale | AI-focused vehicle to buy and transform businesses |
| Oracle | Large debt deals | Cloud and AI data centres |
| Broadcom | Large financing | AI chip and networking business |
SoftBank's plan follows its $30 billion investment in OpenAI. The Financial Times reported it, and Reuters said talks were preliminary and could not independently verify the report.
Several stories from the same week show the strain:
Australian AI infrastructure company Firmus, backed by Nvidia and Blackstone, withdrew a planned $5 billion IPO at a valuation above $30 billion. Reuters reported that only 42 MW of its planned 1 gigawatt of capacity was operational.
Axios reported that OpenAI's annualised revenue was about $50 billion, not the $70 billion previously circulated. The difference reportedly reflects how revenue from cloud partners is counted. It was not confirmed as a decline, but it shows how hard AI revenues are to measure.
The US Federal Reserve raised rates in September 2026, and more hikes are expected. Higher rates make debt-funded projects more expensive.
AI investment is now a major driver of growth. Fed Vice Chair Philip Jefferson said US growth of 2.4% in the first half of 2026 was supported by AI investment. He also noted that core goods prices have risen, driven by AI-related demand, so the AI boom is also adding to inflation.
Other developments:
Arguments that it is sustainable:
Arguments for caution:
Building AI data centres and buying advanced chips costs hundreds of billions of dollars, more than most companies can fund from profits alone.
Morgan Stanley estimates about $1.5 trillion in external financing may be needed by 2028.
Opinions differ. Demand is strong, but heavy debt, uncertain revenues and rising interest rates have raised concerns.
It is boosting GDP growth and investment, but also adding to demand for goods and energy, which can push up prices.
The AI race has moved from software labs to bond markets. Debt is speeding up the building of chips and data centres, but it also raises the stakes. If AI revenues grow as hoped, the investment could pay off. If not, borrowers and lenders could face significant losses.
