Thato Ntseare, an investment manager at E-Squared Investments, has called for stronger governance guardrails around artificial intelligence investments in South Africa, warning that a potential AI bubble could pose risks to markets and investors.
In an interview with Business Times, Ntseare said growing enthusiasm around AI investments is increasingly resembling a speculative bubble, driven in part by companies branding their products as AI-driven without developing proprietary technologies.
“There is a lot of hype around these businesses, and as long as a company says it uses AI in its pitch, it can attract billions of dollars in valuation,” Ntseare said. “But many of these solutions are not, at their core, native AI products.”
Ntseare said the risk of a market correction has increased amid global economic uncertainty, raising questions about whether the scale of current AI-related spending is sustainable over the long term.
According to the Financial Times, global AI spending could exceed $660 billion in 2026, more than half of South Africa’s annual gross domestic product. At the same time, shares of major U.S. technology companies with large AI investments have come under pressure.
Companies including Amazon, Microsoft, Nvidia, Oracle, Meta Platforms and Alphabet lost more than $1 trillion in combined market capitalization in a single day last week, according to FactSet data.
Investor sentiment was further unsettled after Nvidia scaled back expectations around a widely reported investment linked to OpenAI. Nvidia Chief Executive Jensen Huang said reports of a planned investment of up to $100 billion to support new data centers and AI infrastructure were “never a commitment.”
Nvidia shares had surged following a letter of intent in September 2025, reaching a record high before retreating in recent months. Huang later dismissed reports of dissatisfaction with OpenAI’s business practices, saying he continued to support the company.
Some market analysts have drawn parallels between the current AI boom and the dot-com bubble of the late 1990s, pointing to the circular nature of some high-profile AI deals. Others argue that key differences make the current cycle more resilient.
JP Morgan investment specialist Stephanie Aliaga said much of today’s AI buildout is being financed through free cash flow from profitable companies with strong balance sheets, unlike the dot-com era, which relied heavily on external capital and unproven business models.
She also noted that AI revenues are scaling alongside infrastructure investment, with large cloud providers already seeing returns through increased demand and productivity gains. In addition, demand for AI processing power is exceeding supply, contrasting with the excess capacity that characterized the dot-com collapse.
“At the peak of the dot-com era, only a small portion of fiber-optic networks was utilized,” Aliaga said. “Today, data center utilization rates are high, and demand for compute continues to outpace supply.”
Despite these differences, JP Morgan has urged caution, citing the unprecedented scale of AI spending and uncertainties around long-term returns. Investor Michael Burry, known for predicting the U.S. subprime mortgage crisis, has also warned of risks tied to current AI valuations.
Aliaga said investors should focus on selectivity and active management to distinguish companies with sustainable, transformative AI capabilities from those driven primarily by hype.
