AI Spending Tops KSh94 Trillion as Investors Question ReturnsAI Spending Tops KSh94Trillion as InvestorsQuestion Returns

Combined AI spending by major technology companies is expected to exceed KSh94.9 trillion (US$730 billion) in 2026, up from about KSh91 trillion (US$700 billion), Reuters reported on 3 November. Financial leaders say the scale of demand will require more large fundraising deals, even as investors question how quickly the spending will produce returns.

The issue is no longer only how much technology companies are investing. It is who will provide the money, what risks lenders and shareholders are taking, and whether the revenue generated by AI services will justify the cost of the infrastructure.

The AI build-out needs more than company cash

Data centres require land, buildings, electricity, cooling systems, network connections and specialised computer chips. These facilities are expensive to build and operate, and the largest AI companies are racing to secure capacity before rivals do.

Reuters reported that Blackstone President Jon Gray expects more major fundraising deals as the industry seeks capital. Blackstone is itself an active investor in AI infrastructure. The comments came as global financial leaders gathered in Hong Kong and discussed the growing demand for funding.

The spending estimate of KSh94.9 trillion (US$730 billion) is a forecast reported by Reuters for major technology companies. It is not a single audited industry bill, nor does it mean every shilling is being spent directly on AI models. The exact companies and spending categories matter when comparing the figure with revenue or profits.

Investment announcements, capital expenditure, financing commitments and actual cash payments are not the same thing. A company can announce a large infrastructure plan and spread the spending across several years. Investors need to know when the money leaves the balance sheet and what the company expects to earn from it.

Amazon shows the strain on cash flow

Amazon's latest results offer one example of how infrastructure spending can affect a company's finances. In its third-quarter results released on 29 October the company said free cash flow over the previous 12 months was negative KSh5.5 trillion (US$42.4 billion).

Amazon said the change was driven mainly by a KSh12.3 trillion (US$94.8 billion) year-on-year increase in purchases of property and equipment, net of sales and incentives. The company said those investments primarily reflected spending on artificial intelligence.

At the same time, Amazon Web Services (AWS) sales rose 39% year on year to KSh6 trillion (US$46 billion) in the third quarter. That contrast matters. Demand for cloud services is growing quickly, but the company is also spending heavily to build the capacity it expects customers to need.

Free cash flow is the money left after a company pays for its normal operations and investment. A negative figure does not automatically mean a business is in trouble, particularly when it is investing for future growth. It does mean investors need to judge whether the new facilities will generate enough cash later to justify the current outlay.

Investors are asking who carries the risk

Large technology companies can fund some investment from cash generated by existing businesses. But the scale and speed of the AI build-out are creating demand for debt, new share issues, private investment and specialist infrastructure financing.

Borrowing can help a company build capacity sooner, but it also creates interest and repayment obligations. If demand for computing power grows more slowly than expected, a data-centre operator could be left with expensive facilities and debt payments that continue even when customers spend less.

Selling new shares brings in money without adding loan repayments, but it can reduce the ownership percentage of existing shareholders. The market response depends on whether investors believe the new capital will create enough future value to justify the dilution.

Reuters' reporting has also highlighted circular financing arrangements, where investment flows between AI companies and infrastructure providers that are also customers or commercial partners. Such deals are not automatically improper. The concern is whether reported demand reflects independent customers paying for services, or whether funding and spending are partly supporting each other.

This is one reason investors are looking beyond headline revenue growth. They want to see whether customers renew contracts, whether computing capacity is being used, and whether prices cover electricity, depreciation, financing and operating costs.

The companies are not all taking the same route

The largest technology groups have different ways of funding their AI plans. Microsoft, Alphabet, Amazon, Meta and Oracle have large existing businesses that can support investment, but their spending priorities and financing choices differ.

Some infrastructure specialists have a more concentrated risk. Companies whose business depends heavily on renting computing capacity need to keep raising money as they expand. If lenders demand higher interest rates or investors become less willing to provide fresh capital, those firms can face pressure before the wider AI market slows.

The market has shown that investors can support AI growth and still punish a company for the way it finances expansion. The question is not whether AI will be used. It is whether each company's spending pace, debt burden and expected customer revenue make financial sense.

Mwenendo has previously examined the large spending plans of technology groups in its report on AI investment by the world's biggest technology companies. This analysis focuses on the next stage: how the build-out is funded and who bears the risk if returns arrive later than expected.

What this means for Africa

The financing race matters to Africa in practical ways, even though most of the largest data-centre investments are taking place outside the continent. African businesses increasingly rely on cloud services, digital platforms and AI tools whose costs are shaped by global computing supply.

If data-centre operators face higher borrowing costs, they may delay projects or charge more for capacity. If chip supply remains tight, smaller cloud providers and African technology companies may struggle to secure computing resources on competitive terms.

There is also an opportunity. Africa needs reliable electricity, fibre connections and data-centre capacity to support local digital businesses. Investors looking at the continent will want evidence that power supply, connectivity, regulation and customer demand can support facilities that are expensive to build and maintain.

The risk is assuming that global AI spending automatically translates into African investment or jobs. That outcome depends on where infrastructure is built, who owns it, whether local power and network systems can support it, and whether companies can attract enough paying customers.

What investors should watch next

The next signals will come from company results, debt deals, new share issues and disclosures about data-centre use. Investors should compare capital spending with operating cash flow, borrowing and the revenue earned from AI products and cloud services.

A rise in spending is not proof of a bubble. Nor is a rise in AI revenue proof that every infrastructure project will earn an adequate return. The evidence will emerge company by company, as facilities open, customers pay and financing costs become clearer.

The central question for markets is whether AI companies can turn a very large infrastructure bill into lasting cash generation. Until that becomes clearer, fundraising itself will be part of the story, not just a footnote to the technology race.