Three years have passed since OpenAI ignited the artificial intelligence boom with the launch of ChatGPT. While capital continues to flow heavily into the sector, doubts are growing over whether this golden era can be sustained.

Signs of skepticism are becoming more evident, ranging from massive sell-offs in Nvidia Corp. shares and a sharp drop in Oracle Corp.’s stock following reports of rising AI-related spending, to weakening sentiment around companies closely tied to OpenAI.

As 2026 approaches, investors face a critical debate: should they scale back AI exposure ahead of a potential bubble burst, or double down to capitalize on what many view as a transformative technology?

“We are now entering the phase of the cycle where the real test begins,” said Jim Morrow, CEO of Callodine Capital Management. “It has been a compelling story, but capital is now at stake to determine whether returns will justify the investment.”

Concerns surrounding AI trades include the actual adoption of the technology, soaring development costs, and whether consumers will ultimately pay for AI-based services. The answers to these questions will have major implications for the future direction of equity markets.

Much of the S&P 500’s US$30 trillion market value increase over the past three years has been driven by major tech firms such as Alphabet and Microsoft, along with beneficiaries of AI infrastructure spending like Nvidia, Broadcom, and energy providers such as Constellation Energy. If their momentum stalls, broader equity indices are likely to follow.

“These stocks don’t correct because growth slows,” said Sameer Bhasin of Value Point Capital. “They correct when growth stops accelerating.”

Still, there are solid reasons for optimism. The tech giants leading AI investment possess deep financial resources and have pledged to continue heavy spending in the years ahead. Meanwhile, AI developers such as Google are making steady progress with new models, fueling ongoing debate about the sector’s trajectory.

Several key trends stand out amid current market volatility.

Access to Capital
OpenAI plans to spend roughly US$1.4 trillion over the coming years, despite generating revenue that remains far below its operating costs. The world’s most valuable startup is expected to burn US$115 billion through 2029 before turning cash-flow positive in 2030.

So far, funding has not been an issue, with OpenAI raising US$40 billion from SoftBank and other investors. Nvidia has also pledged up to US$100 billion in investments, raising concerns about circular financing within the AI ecosystem.

However, a pullback in investor appetite could pose serious challenges for OpenAI and related firms such as CoreWeave. Heavy reliance on debt, as seen with Oracle, has also heightened pressure following capital spending overruns and delays in AI data center projects.

Aggressive Spending by Big Tech
Alphabet, Microsoft, Amazon, and Meta are expected to spend more than US$400 billion over the next 12 months, largely on data centers. Revenue growth from AI-related cloud and advertising businesses has yet to match these expenditures.

Earnings growth for the seven largest tech firms is projected at 18% in 2026—the weakest pace in four years. Rising depreciation costs from data center expansion may also weigh on share buybacks and dividends.

The most significant concern lies in the strategic shift itself. Big Tech valuations have historically been built on rapid revenue growth with low costs, generating strong free cash flow. AI investments are reversing that equation.

A More Rational Optimism
While Big Tech valuations remain elevated, they are far below levels seen during past market manias. The Nasdaq 100 currently trades at about 26 times forward earnings, compared with more than 80 times at the peak of the dot-com bubble.

Some AI stocks, such as Palantir and Snowflake, carry extremely high valuations. However, major players like Nvidia, Alphabet, and Microsoft trade below 30 times earnings—relatively modest given the surrounding enthusiasm.

As a result, investors remain conflicted. Risks are clearly visible, yet valuations have not reached panic-inducing levels. The key question now is where AI-driven trading heads next.

“This kind of collective mindset eventually breaks,” said Bhasin. “It may not collapse like in 2000, but a significant rotation is inevitable.”