C-suite survey adds fuel to the 'AI bubble' fire

This is the thrust of a survey carried out by RWS, a company that provides translation and intellectual property (IP) services to enterprises.

It found that 87% of respondents are under pressure to rapidly implement GenAI solutions at speed and scale, while 76% said there was "overwhelming excitement" at their organisation about its potential benefits.

However, 68% admitted that with so much noise around GenAI at the moment, they find it difficult to identify genuine innovators. 36% raised concerns that due to their company's enthusiasm for all things GenAI, it is diverting resources that could be better deployed elsewhere.

"Recent advances in GenAI have triggered an innovation race," said Mark Lawyer, general manager of linguistic AI at RWS. "In a bid to be first past the post there's a real risk of failing to see any value from AI investments. It's critical that business leaders are not reactive or piecemeal, but rigorous and purposeful in their approach."

The RWS survey is the latest in an expanding catalogue of material that is somewhat tempering the excitement surrounding the technology.

Thanks largely to Nvidia's recent ascension to the world's most valuable company – albeit temporarily – the conversation about GenAI is shifting away from topics like governance and use cases, to concerns that asset prices have completely decoupled from the underlying market's true value, and therefore the whole thing is a massive bubble that could pop at any moment.

James Ferguson, founding partner of macroeconomic research firm MacroStrategy Partnership, drew parallels between today's GenAI market and the dotcom bubble that burst so spectacularly in 2000.

Speaking on a recent Bloomberg podcast, he warned that there is still a trust issue when it comes to the output from GenAI, necessitating human verification.

"If you have to check it, then AI is effectively, in my mind, useless," he said.

Ferguson also warned that companies like Nvidia are continuously churning out new and better AI hardware in the form of graphics processing units (GPUs), resulting in the rapid depreciation of their older products.

On top of that, Ferguson also highlighted that AI is very energy intensive, making it expensive to run, "and it's yet to prove that it can pay for itself."

Ferguson isn't the only one concerned about GenAI's revenue potential, or lack thereof.

David Cahn, head of growth at investment firm Sequoia, last month warned that the yawning chasm between the implied revenue expectations of deploying and operating AI infrastructure and the actual revenue being generated by the AI ecosystem is getting even wider.

In September 2023, he suggested the industry would need to turn over $200 billion to cover its costs. As of June, Cahn puts this number at $600 billion. This is a problem because today the AI market is generating around $100 billion of revenue, leaving it $500 billion in the red.

Cahn is bullish about the AI market's long-term potential, but says the road to riches won't necessarily be smooth.

"We need to make sure not to believe in the delusion that has now spread from Silicon Valley to the rest of the country, and indeed the world. That delusion says that we're all going to get rich quick, because AGI (artificial general intelligence) is coming tomorrow, and we all need to stockpile the only valuable resource, which is GPUs."

Unfortunately, no one can predict with absolute certainty when a bubble will burst, but the growing chorus of concerned voices suggests that all is not well.