Speculation runs the modern world. It wasn’t always this way. But somewhere between the 1980s and the 2008 financial crash, a shift occurred. Investors discovered that hype was far more lucrative than building anything real. In the information age, narratives are cheap to manufacture, easy to spread, and highly profitable.

The result has been a parade of bubbles: the dot-com crash, the 2008 credit crisis, the 2016–17 crypto boom, the 2020–21 crypto resurgence, the 2022 NFT implosion—and now, the AI bubble. Today, nearly half of global private investment is pouring into AI, and speculation around it is one of the primary forces propping up the S&P 500. But just like every bubble before it, AI is beginning to show unmistakable signs of collapse.

This time, however, the tech and finance crowd have learned to prepare an exit. They’re already building the next hype vehicle to jump into once AI starts sinking. Unfortunately, that next bandwagon—quantum computing—may be an even bigger dead end.

By now, it’s widely understood that AI’s explosive growth is unsustainable. Concepts like the efficient compute frontier and the Floridi conjecture suggest that modern AI models are already close to their maximum potential. Even massive increases in scale deliver only marginal gains. ChatGPT-5 illustrates this perfectly: vastly more data, compute, and capital than GPT-4, yet only incremental improvements in performance.

This presents a serious problem. As they currently exist, generative AI systems are neither particularly useful nor meaningfully profitable.

An MIT study found that 95% of corporate AI pilots failed to improve productivity or profitability. In the remaining 5%, AI was confined to tightly controlled back-office tasks—and even then, gains were minimal. A METR report showed that AI coding tools often slow developers down, introducing strange, hard-to-trace bugs that take longer to fix than writing the code manually. Other research indicates that for 77% of workers, AI has increased workload rather than reduced it.

In short, today’s AI is too unreliable to deliver the productivity revolution that justifies the scale of investment behind it.

For that speculation to pay off, AI would need to become dramatically better—which would require exponentially more spending. That’s a problem when even the industry leader is bleeding money. OpenAI reportedly loses money on every $200-per-month subscription and would need pricing closer to $2,000 just to break even.

Meanwhile, Big Tech—propped up by venture capital and investment banks—has been pouring hundreds of billions of dollars annually into AI. The technology is hitting fundamental limits, profitability remains distant, and expectations are wildly detached from reality. With GPT-5 underwhelming, Meta scaling back its AI ambitions, and interest rates threatening to rise, even the institutions that inflated this bubble are warning of an impending crash. Goldman Sachs has gone so far as to suggest that when the AI bubble bursts, it will also take the data-centre boom down with it, damaging not just OpenAI, Meta, Google, Anthropic, and xAI, but also infrastructure providers like Amazon, Oracle, and Nvidia.

In other words, when this collapses, it will hurt just about every tech bro and finance guy you know.

Luckily for them, there’s an escape plan: quantum computing.

Quantum computers, in theory, are extraordinary. Instead of bits that are either 0 or 1, they use qubits that can exist in superposition. This allows, at least theoretically, exponential computational power. Quantum machines have already solved niche problems in minutes that would take classical supercomputers longer than the age of the universe.

Naturally, this has led to grand promises: breakthroughs in chemistry, revolutionary AI models, even speculation that quantum systems could finally unlock human-level intelligence. Some have even claimed that the human brain itself operates as a quantum computer.

You can see why investors are salivating.

Google, Microsoft, and Amazon are all building quantum hardware. Nvidia is developing quantum platforms. OpenAI has hired leading quantum physicists. Musk has begun floating quantum ideas for his AI ambitions. Smaller quantum startups are exploding in valuation—Quantinuum recently raised $600 million, doubling its valuation to $10 billion.

Quantum computing is being positioned as the lifeboat for the AI bubble.

But here’s the problem: quantum computing is not what it’s being sold as.

A fully functional, general-purpose quantum computer is still likely a decade or two away. These machines are extraordinarily difficult to build and operate. But hardware isn’t the biggest obstacle—software is.

Quantum computers are not universally faster than classical ones. They only outperform them on very specific problems. Worse, they cannot run standard algorithms. Reading a qubit collapses its quantum state, effectively turning it into a normal bit. To extract value, quantum algorithms must carefully manipulate wave interference before measurement—and designing such algorithms is fiendishly difficult.

So far, we have only a handful of useful quantum algorithms, mostly limited to narrow mathematical or physics problems. We have none that meaningfully apply to chemistry at scale or to training neural networks. Many researchers suspect such algorithms may not even exist, as AI training relies on unstructured data and mathematical operations poorly suited to quantum systems.

In other words, even if usable quantum hardware arrives sooner than expected, current evidence suggests it won’t solve AI’s fundamental problems.

The idea that the brain is a quantum computer—the spark behind “quantum AI”—has also been largely debunked by recent research. But that hardly matters. The myth is already out there, circulating freely, ready to be monetised.

If the quantum hype grows fast enough, it may temporarily delay the AI reckoning. But eventually, reality will catch up. Promised breakthroughs won’t arrive. The narrative will collapse. Hundreds of billions of dollars—money that could have gone toward wages, infrastructure, or real innovation—will evaporate.

And when it all goes wrong, the billionaires will already be gone, pockets lined, leaving the rest of us to deal with the fallout.

                                                                                                                                                          .   .   .

Thanks for reading! Don’t forget to follow me on linkedin, Twitter, and Instagram.