The global artificial intelligence investment boom is entering a more dangerous financial phase as technology companies and infrastructure developers pour enormous amounts of borrowed money into data centers, raising concerns that not every project will generate the returns needed to repay its lenders.
Now, one of the world’s largest banks is issuing a warning.
HSBC Chief Executive Georges Elhedery says artificial intelligence has extraordinary potential to transform businesses, but investors should not assume that every data center benefiting from the industry’s expansion represents a sound financial investment.
Speaking to CNBC’s Dan Murphy on October 20, Elhedery urged lenders and investors to examine the quality of data center tenancy agreements and the prices customers have agreed to pay.
His message comes as trillions of dollars are expected to flow into AI infrastructure and financial institutions face mounting questions over how much borrowing the industry can sustain.
HSBC CEO WARNS NOT ALL DATA CENTERS ARE GOOD INVESTMENTS
For Elhedery, the concern is not whether artificial intelligence will become an important part of the global economy.
He believes the technology will fundamentally change how businesses operate, including HSBC’s own banking operations.
But the financial viability of the facilities supporting that transformation is a different question.
The HSBC chief executive warned that investors must be selective when financing data centers, pointing particularly to customer contracts and the pricing arrangements underpinning individual projects.
A data center may be equipped with advanced processors and located in a rapidly growing market, but that does not automatically guarantee sufficient revenue to cover construction expenses, electricity consumption, maintenance and borrowing costs.
Facilities supported by financially strong tenants, favorable long-term contracts and sustainable operating economics could be better positioned than speculative developments that depend on future customers materializing.
The distinction is increasingly important as companies compete to build computing infrastructure at unprecedented speed.
JPMORGAN PROJECTS $4.1 TRILLION IN AI-RELATED DEBT ISSUANCE
The scale of the financing boom is staggering.
According to estimates cited by CNBC, JPMorgan Chase has projected that approximately $4.1 trillion in AI-related debt could be issued by the end of the decade.
This is an estimate of potential debt issuance across the AI investment ecosystem, not money that has already been borrowed or a prediction that those loans will default.
The figure nevertheless illustrates how heavily the next stage of the artificial intelligence expansion could depend on lenders and capital markets.
Companies building large-scale AI infrastructure frequently need to secure billions of dollars before their facilities can begin generating revenue.
That money goes toward land acquisition, construction, specialized cooling systems, electricity infrastructure, networking equipment and expensive processors used to train and operate advanced AI systems.
The challenge is that construction costs arrive long before the full financial benefits of the investment are realized.
For lenders, the central question is whether the future revenue generated by a project will be sufficient to service debt and produce acceptable returns.
DATA CENTER INVESTMENTS COULD REACH $7 TRILLION BY 2030
Industry projections cited in CNBC’s reporting suggest that global data center spending could reach approximately $7 trillion by 2030.
That estimate reflects the scale of the infrastructure investment being contemplated, rather than a guaranteed amount of expenditure.
The expansion is being driven by major technology companies and specialized cloud providers competing to secure enough computing power to support AI development and commercial deployment.
The infrastructure requirements extend far beyond computer chips.
Data centers need reliable electricity, advanced cooling technology, fiber-optic connections, secure facilities and access to land where construction can be completed.
These requirements have transformed artificial intelligence from a predominantly software-focused industry into a major driver of construction, energy and infrastructure investment.
However, the larger the spending commitments become, the more important it is for investors to distinguish between projects with sustainable demand and those based on overly optimistic growth assumptions.
AI BORROWING SPREE RAISES CONCERNS ON WALL STREET
Financial markets have become increasingly sensitive to the borrowing requirements of companies involved in artificial intelligence.
Investors are closely monitoring bond issuance, private credit transactions and specialized financing arrangements tied to AI infrastructure.
JPMorgan’s multitrillion-dollar projection comes amid forecasts of a sharp rise in bond borrowing by major technology companies seeking to expand their computing capacity.
Allianz adviser Mohamed El-Erian has pointed to the enormous amounts of financing being raised by large cloud and technology operators, highlighting the growing competition for capital.
The concern is that the financing demands of the AI industry are expanding at a time when borrowing costs remain significant.
If companies issue debt at expensive rates, their infrastructure projects must generate sufficiently strong earnings to justify those costs.
Higher interest expenses can also reduce the amount of cash available for additional investment.
The result could be increasing pressure on companies whose business models depend on continuous expansion and access to affordable financing.
IMF WARNS ABOUT INTERCONNECTED AI FINANCING
The International Monetary Fund has also highlighted risks arising from the complicated financial relationships within the artificial intelligence industry.
Its October 2026 Global Financial Stability Report examined major AI-related financing arrangements involving substantial debt, equity investment and future revenue commitments.
Analysis of 14 major transactions identified approximately $300 billion in equity, $650 billion in debt and $1.4 trillion in future revenue commitments.
These categories describe different financial exposures and should not be combined into a single measure of outstanding debt.
Some arrangements involve large technology companies investing in AI developers that subsequently commit to purchasing cloud infrastructure or computing services from their investors or partners.
Such relationships can provide funding and commercial demand simultaneously.
However, they may also create financial dependencies that become problematic if expected revenue fails to materialize.
A decline in AI spending could therefore affect multiple companies connected through investment agreements, infrastructure contracts and financing commitments.
The IMF’s findings do not establish that a financial crisis is inevitable.
They do, however, reinforce concerns about transparency, financial leverage and the concentration of risk across the rapidly expanding AI ecosystem.
COREWEAVE HIGHLIGHTS THE COST OF AGGRESSIVE EXPANSION
Recent developments involving Nvidia-backed cloud computing provider CoreWeave illustrate why investors are becoming more cautious.
CoreWeave supplies high-performance computing infrastructure to customers developing and deploying artificial intelligence.
The company has secured major commercial agreements and expanded rapidly to meet strong demand for AI computing.
But its growth has also required substantial borrowing.
Recent market reporting indicates that CoreWeave has accumulated approximately $52 billion in debt and financing obligations, depending on the obligations included in the calculation.
Yields on some of its unsecured bonds have moved toward distressed levels, reflecting investor concerns about credit risk and future cash generation.
JPMorgan analysts have examined potential banking-sector exposure to CoreWeave and concluded that the risks for major global banks could remain manageable because of collateral arrangements and structured lending protections.
Nevertheless, the situation demonstrates an increasingly important distinction.
Strong customer demand does not automatically mean an infrastructure company can finance its expansion cheaply or generate positive cash flow quickly.
That lesson is relevant to other highly leveraged AI infrastructure developers.
GOLDMAN SACHS IDENTIFIES A DIFFERENT AI BOTTLENECK: ELECTRICITY
The financial concerns surrounding data centers are being intensified by physical infrastructure challenges.
A Goldman Sachs analysis reported by Axios on October 20 projected that global data center electricity demand could grow approximately 170% between 2025 and 2030.
More than 60% of that projected increase is expected to occur in the United States.
However, electricity infrastructure is not expanding as quickly as the technology industry’s ambitions.
Power-grid connections, construction approvals, transmission capacity and local opposition can delay new facilities.
Axios reported that dozens of infrastructure projects had been canceled or stalled during 2026, underscoring the difficulties developers face in turning investment plans into operating facilities.
Such delays create financial risks.
Companies may already have committed capital to property, equipment and construction before a data center can begin generating revenue.
If electricity access is delayed, the time needed to recover those investments becomes longer.
For borrowers facing substantial interest expenses, the consequences can be serious.
The emerging challenge is therefore not simply obtaining enough AI chips.
It is financing and building the electricity and infrastructure systems required to use them profitably.
HSBC SEES AI AS A TRANSFORMATIONAL OPPORTUNITY
Despite his warning about data center financing, Elhedery remains optimistic about the broader potential of artificial intelligence.
HSBC is pursuing its own AI transformation, including initiatives designed to improve productivity, automate internal processes and strengthen customer services.
The bank has announced a global AI center of excellence in Singapore and expanded its work with Google Cloud.
These initiatives illustrate how financial institutions can simultaneously embrace AI technology while remaining cautious about lending to infrastructure developers.
HSBC’s October 19 investment commentary also argued that high AI spending does not automatically imply an impending debt crisis.
Instead, the bank emphasized assessing corporate leverage, realistic cash-flow expectations and the valuations attached to AI investments.
That distinction is important.
Some of the world’s largest technology companies possess strong balance sheets and established sources of recurring revenue.
Smaller infrastructure developers may face greater financial pressure because they depend more heavily on external financing and concentrated customer relationships.
Investors should therefore avoid treating every company connected to artificial intelligence as having the same risk profile.
FINANCIAL TIMES AND REUTERS HIGHLIGHT GROWING CONCERNS
Other major financial news organizations have also examined the vulnerabilities emerging in the AI investment boom.
The Financial Times reported on October 20 that investors were increasingly using credit-default swaps to protect themselves against potential problems involving AI-related companies.
Such instruments can provide a form of insurance against credit events, although their pricing also reflects market sentiment and assumptions about potential recoveries.
Reuters reported on October 14 that major Wall Street banks were balancing opportunities to finance AI expansion against the need for disciplined lending standards.
The reporting highlighted concerns that substantial capital expenditure may not immediately produce the economic returns investors expect.
Meanwhile, The Wall Street Journal has examined growing differences between the financing costs of AI cloud providers and those of data center property and infrastructure projects.
The distinction suggests that credit markets are becoming more selective rather than treating all AI-related debt equally.
Projects supported by secured assets and strong contracts may attract financing on more favorable terms than companies relying on less-protected borrowing structures.
Together, these developments reinforce HSBC’s central warning: The growth of artificial intelligence does not guarantee that every investment connected to it will succeed.
WHAT THIS MEANS FOR ASIA AND THE PHILIPPINES
The global financing race also has implications for Asian markets, where technology companies, property developers, energy providers and infrastructure investors are exploring AI-related opportunities.
Regional economies such as Singapore, Malaysia, Japan and the Philippines could benefit from rising demand for data storage, cloud computing, telecommunications and electricity infrastructure.
However, developers must consider more than the possibility of attracting international technology customers.
Access to dependable electricity, competitive energy prices, data connectivity, land availability and regulatory approvals will influence the economics of individual projects.
For the Philippines, stronger digital infrastructure could support cloud adoption, business process outsourcing, financial technology and the growing use of AI by local businesses.
But large data center investments also require careful assessment of electricity reliability, construction costs and long-term customer demand.
The lesson from HSBC’s warning is particularly relevant for emerging markets seeking a share of the AI infrastructure boom.
Attracting investment is only the first step.
Ensuring that projects generate sustainable financial and economic benefits is the much more difficult challenge.
THE BIGGER QUESTION: WHO WILL PAY FOR THE AI BOOM?
The global artificial intelligence industry is entering a period when financial discipline could become just as important as technological innovation.
For years, investors have focused on the race to produce more powerful AI models, secure advanced semiconductors and build enormous computing facilities.
Now they are increasingly asking how those investments will be funded and when they will generate sufficient returns.
HSBC’s warning does not mean the bank expects artificial intelligence to fail.
Instead, it highlights the possibility that a transformative technology can produce both extraordinary business opportunities and costly investment mistakes.
The winners could be companies that combine strong customer demand with reliable infrastructure, sound contracts and manageable debt.
The losers could be developers that overestimate future revenues, underestimate construction costs or rely too heavily on continued borrowing.
As the AI industry’s financing requirements move toward the trillions of dollars, these differences could become more pronounced.
Artificial intelligence may transform the world economy, but that does not mean every company financing the revolution will emerge as a winner.
And with banks, private lenders and bond investors committing increasingly large amounts of capital to the infrastructure race, one uncomfortable question is becoming harder to ignore:
If the AI boom continues growing but some data centers fail to generate enough revenue, who will ultimately absorb the financial losses?
