Artificial intelligence companies are now planning to spend huge sums of money at a time when governments are taking out large loans and interest rates are rising. The chief executive of JPMorgan believes there could be a financial strain that reaches well beyond Silicon Valley.
Artificial intelligence could transform the global economy, but the funds needed for that transformation are becoming more and more costly. Although technology companies are planning to spend hundreds of billions of dollars on data centres, semiconductors and computing infrastructure, governments are at the same time taking out huge loans. This has led to an odd competition for capital which might eventually push up financing costs throughout the wider economy, even if artificial intelligence turns out as successful as its backers expect.
In his interview with Bloomberg Television on October 6 at the JPMorgan technology conference in London, chief executive Jamie Dimon recognised the tremendous potential of artificial intelligence while cauting about the state of the financial environment in which it is being developed. The worry he expressed was not that the technology would inevitably fail, but that the large sums of money being asked for by technology companies and governments could influence the cost of borrowing for everybody else.
“Rates are going up. A lot of demand for capital, a lot of government financing,” Dimon said.
America’s $2 Trillion Borrowing Habit Is Reshaping the Treasury Market →
The extent of the competition can be better understood by looking at the figures. At JPMorgan’s second-quarter earnings call, Dimon stated that artificial intelligence capital expenditure would be about $700 billion in 2026, the forecasts going above $1 trillion in 2027. Once again, these amounts refer to investment spending and not just borrowing, yet they show how much financing the technology industry might need.
At the same time, the Congressional Budget Office estimated on October 8 that the United States had a federal budget deficit of about $2 trillion in fiscal year 2026, this representing an increase of $218 billion over the previous year. The government meets its deficit by issuing debt, which in turn places extra pressure on financial markets that are already providing funds for corporate investment, mortgages and other types of borrowing.
During the interview Dimon referred to the wider fiscal problem and stated, “We can’t borrow endlessly and spend endlessly.” His worry was that investors might eventually require ever more attractive yields in order to keep financing the governments. Once this situation arises, the consequences can extend into the corporate credit markets, causing it to become more expensive for businesses which have nothing at all to do with artificial intelligence.
Wall Street Is Watching AI Capex. The Real Bill Is Still Coming →
The financial effects are already appearing in the market for corporate bonds. As a September Reuters analysis has shown, the five big technology companies, namely Alphabet, Amazon, Meta, Microsoft and Oracle, raised about $220 billion through the issuance of bonds in the previous year. This huge amount of borrowing is now having an impact on how investors set prices for corporate debt, especially in cases where several technology companies need large amounts of financing at the same time.
A credit spread represents the extra interest that investors require when lending to a company as compared to holding a government bond that has the same maturity. When government yields rise at the same time as investors are demanding larger corporate spreads, borrowing becomes more expensive regardless of which option is chosen. A technology firm which is making huge profits may be able to withstand this kind of pressure, but a smaller manufacturer, retailer or infrastructure developer could end up putting off investment since financing has become too expensive.
This could lead to a second-order effect of the artificial intelligence boom. Although the technology might result in substantial gains in productivity, the funding needed to develop it could place a strain on other industries. Investors will thus have to decide not only whether or not artificial intelligence companies will eventually be able to justify their spending, but also how their enormous demand for capital alters the financial situation of companies in sectors other than technology.
AI Is Coming for Your Electric Bill as Power Costs Surge →
There is also a significant counterpoint. Interest rates which result from productive investment do not have to indicate that the economy is in difficulty. For example, the building of new data centres will need construction workers, electrical equipment, power generation and sophisticated manufacturing. Such investments might eventually lead to higher productivity, cause economic output to expand and generate a sufficient amount of extra wealth to make it worthwhile to pay the financing costs. Dimon himself stayed optimistic regarding artificial intelligence, pointing out that technological progress could bring about huge benefits in the longer term.
Yet even when investments are successful there are uncertainties involved. Data centres may suffer from delays in construction, from shortages of electricity and from legal opposition, while new semiconductor designs or more efficient AI models could alter the economics of projects which were planned many years earlier. The financial risks increase if the expected returns are delayed but the interest payments still have to be made. Unlike shareholders, lenders usually expect to be paid no matter how quickly a technology becomes profitable.
The difference between technological success and financial success could end up being one of the major investment questions over the coming several years, with artificial intelligence possibly fulfilling its remarkable promises even though certain companies, creditors or shareholders continue to suffer losses since too much money was invested at the wrong price.
Investors ought therefore to keep an eye on Treasury yields, corporate credit spreads, the amount of debt being issued relating to AI, and the profitability of new computing infrastructure. The most important warning contained in Dimon’s comments is not that the artificial intelligence revolution will fail. It is that the revolution has to compete for financing in an economy in which money is becoming increasingly costly, and the final cost could turn out to be much greater than that incurred by the companies which are building it.