AI's Reckoning: Big Promises, Bigger Bills, and a Growing Backlash

In recent times, artificial intelligence has dominated the conversation in global markets. Trillion-dollar valuations, record venture capital inflows, and breathless headlines about machines that could do everything from write code to generate Hollywood films made AI feel like an unstoppable force. In 2026, the bill is arriving and the industry is starting to buckle under its weight.

The Spending Trap

The scale of investment flowing into AI is staggering and the question of who captures the returns is becoming impossible to ignore. Advances in AI have set off an arms race among the largest technology companies in the world - Meta, Microsoft, Alphabet and Amazon, each racing to build models capable of competing with OpenAI and Anthropic. The capital being deployed is historic, yet the commercial payoff remains elusive for many.

The companies pouring the most money into AI infrastructure are not necessarily the ones best positioned to monetise it. OpenAI itself, the company that sparked the entire boom, has watched ChatGPT's weekly users grow to nearly one billion, yet has struggled to convert that extraordinary reach into sustainable revenue streams, abandoning expensive initiatives precisely because there was no viable path to profit.

Meanwhile, both OpenAI and Anthropic are now plotting initial public offerings as they seek new sources of financing, having exhausted much of the runway provided by private funding rounds from institutional investors.

OpenAI's Strategic Retreat

No story better captures AI's monetisation crisis than OpenAI's recent pivot. The company that triggered the generative AI boom has spent recent months quietly dismantling several of its most high-profile initiatives. Sora, its video generation app launched with enormous hype, has been scrapped. A $1 billion content deal with Disney, announced less than four months ago, has been abandoned. Plans for an erotic chatbot have been shelved indefinitely, and all of this comes barely three months after CEO Sam Altman declared an internal "Code Red," urging engineers to redirect all effort toward countering advances by Google.

The FT has framed this generously as OpenAI "showing its mettle", a start-up pivoting quickly when plans don't work. Sora was a serious drain on computing power with no viable monetisation strategy.

OpenAI, riding the wave of nearly one billion weekly ChatGPT users, believed it could do everything at once. It is now learning that focus is a competitive advantage it cannot afford to ignore, particularly as Google and Meta lean aggressively into their vast existing user bases to promote rival AI services. As the FT notes, there is still plenty of room for changes of leadership before the AI market coalesces around a clear winner.

The Hardware Crisis

Beneath the software layer, the physical infrastructure of the AI boom is facing its own reckoning. The entire industry runs on advanced semiconductors and those chips are now at the centre of a serious geopolitical crisis.

Following the indictment of a Supermicro co-founder accused of conspiring to ship large amounts of advanced Nvidia chips to China via south-east Asian intermediaries including Malaysia, Thailand, Vietnam and Singapore, US senators have demanded the immediate suspension of Nvidia's export licences.

The scale of the alleged diversion scheme has raised serious questions about whether the industry's compliance processes are adequate to protect cutting-edge American technology from reaching foreign adversaries. Republican Senator Jim Banks and Democrat Elizabeth Warren wrote jointly that evidence of large-scale smuggling suggested licences should not be issued based on assurances from executives with a financial stake in the outcome. For an industry already grappling with questions about returns on investment, supply chain vulnerability is the last thing it needs.

Who Actually Benefits?

Perhaps the most pointed intervention has come from an unexpected voice. Larry Fink, CEO of BlackRock, the world's largest asset manager managing $14 trillion used his annual shareholder letter to warn that AI risks repeating historical patterns of wealth concentration at an even larger scale. "The massive wealth created over the past several generations flowed mostly to people who already owned financial assets," he wrote, adding that AI threatens to repeat that pattern on an even grander scale.

The companies with the data, infrastructure and capital to deploy AI at scale are positioned to benefit disproportionately, and ownership of those companies remains narrow. BlackRock has leaned into this dynamic, partnering with Microsoft, Nvidia and Abu Dhabi fund MGX on a $30 billion vehicle to invest in the AI industry, and agreeing a $40 billion takeover of data centre operator Aligned Data Centers. When even the financiers of the AI boom are warning about its social consequences, it is worth paying attention.

The Political Dimension

The industry's troubles are not purely financial. Anthropic, developer of the Claude AI model and valued at $380 billion, found itself at the centre of a political confrontation after refusing to allow its technology to be used for lethal autonomous weapons and mass surveillance.

The Trump administration responded by designating Anthropic a supply-chain risk, a national security measure previously reserved for businesses linked to foreign adversaries, and barred civilian agencies from doing business with the company. A federal judge has since temporarily blocked the designation, ruling it was likely "arbitrary and capricious", but the episode illustrates the increasingly tense relationship between AI companies and government power.

The Road Ahead

None of this means AI is finished. The technology is real, the long-term productivity gains remain compelling, and the appetite for investment shows no sign of disappearing. However, the industry is entering periods of volatility. The hype cycle has peaked, geopolitics are intensifying, the monetisation gap is widening, and the question of who actually captures the gains remains unresolved. For the companies and investors still pouring money in, 2026 may prove to be the year the easy optimism finally ran out.