Is the AI bubble about to burst?  Not if the leaders of the biggest tech firms are to be believed.

Gathered around a table in the White House, several of the world’s most powerful tech CEOs heaped praise on President Donald Trump for his leadership, saying his approach to AI will position the US as the dominant player in the emerging sector.

Between them, Mark Zuckerberg, Tim Cook, Sergey Brin, Bill Gates, Sam Altman and others manage the companies driving the current AI revolution – and if they are to be believed, the US AI sector is on the cusp of greatness.

Google co-founder Brin said: “It’s a real incredible inflection point right now in AI and the fact that your Administration is supporting our companies instead of fighting with them — it’s hugely important. It’s a global race and I think we’re at the cusp where these AI models are about to become profoundly useful… so we’re very grateful for your Administration’s support.”

Oracle boss Safra Catz agreed, adding that “AI is going to change everything” while Alphabet CEO Sundar Pichai added: “The AI moment is one of the most transformative moments any of us have ever seen or will see in our lifetimes.”

But elsewhere this weekend, evidence suggests the AI hype cycle could be losing steam.

Slowing rates

A study from Apollo Academy released this weekend showed that AI adoption rates at larger companies are starting to slow.

The study is based on US Census Bureau data, which is released biweekly and surveys 1.2 million firms. One question added to the survey since 2023 is whether a business has used AI tools, such as machine learning, natural language processing, virtual agents, or voice recognition, to help produce goods or services in the last two weeks.

From November 2023, the figure increased in US businesses of all sizes, but perhaps unsurprisingly, the fastest rise was on firms with more than 250 employees. But according to the latest figures, this number fell in June and July for larger businesses and also flattened out at smaller firms.

The data measures enterprise integrations, not usage of AI tools across the board, but it suggests that adoption of major top down implementations are slowing.

This is backed up by spending data from Ramp, released earlier this year. Ramp’s AI Index suggests that the meteoric rise in business AI spending might be showing signs of deceleration.

The figures show that AI penetration among US businesses had reached 41.7% as of April, but the growth trajectory flattened since the end of last year. It is worth noting that Ramp’s index data is taken from among its customers so may not be reflected across the entire US, or global markets.

Torsten Sløk, chief economist at Apollo Academy, highlights that the slowdown signals caution among large enterprises, while Arpit Gupta, associate professor of finance at NYU Stern, suggests that “trillions in AI capex should probably be reconsidered.”

For Andy Ward, SVP International at Absolute Security, the reports highlight a growing concern that businesses are beginning to question the return on investment from AI projects.

He says: “AI can transform detection and response, but if it’s deployed without robust resilience strategies, real-time visibility, and clear governance, it risks adding more vulnerabilities than it solves.

“Our research shows that over a third (34%) of CISOs have already banned certain AI tools like DeepSeek entirely, driven by fears of privacy breaches and loss of control. As attackers leverage AI to reduce the gap between vulnerability and exploitation, our defences must evolve with equal urgency. Now is the time for security leaders to ensure their people, processes, and technologies are aligned, or risk being left dangerously exposed.”

ROI problems

 

David Tyler, founder of Outlier Technology, believes the data from MIT – which showed 95% of Generative AI projects failed to show any value – suggests the hype cycle is on the verge of bursting.

He claims that investors have been “blinded by the hype machine’s promises” and have “continually increased the amount of money being thrown into AI”. Major players are reporting losing $2.40 for every $1 they make; and a recent economic modelling of data centres showed they’d need 3.5 billion users, each paying $35 a month, just to break even, Tyler explains. “That amounts to 64%  of all of the people with access to the internet in the entire world.

“So, the economics don’t stack up, the majority of projects aren’t providing any value, and ChatGPT 5 has been critiqued as being not quite the ‘game changer’ promised ahead of its release. All of which has contributed to the increased mainstream attention on the AI bubble.”

However, he says that AI can still play an important role going forward if enterprises shift expectations.

“We need to realise that AI does have specific uses, but there is not one single model which is going to revolutionise the entire world,” adds Tyler. “We will always need humans to be part of the process somewhere down the line, and we’re better off focusing on the small gains we can make and the specific processes we can improve with AI, rather than ploughing trillions of dollars into an unattainable magic ‘fix all’ solution which will be all things to all men.”

A recent survey from Qlik highlights the problem with ROI. Some key takeaways from the report showed:

  • Only 11% say most of their AI initiatives have delivered tangible gains
  • Just 51% evaluate AI using KPIs tied directly to business performance
  • Nearly a quarter (23%) say the majority of their AI use cases are still stuck in the experimental phase
  • HR (37%) and finance (30%) are seen as the departments where AI is having the least tangible benefit

James Fisher, chief strategy officer at Qlik, says this data does not mean that AI is all hype, however.

“It’s because many businesses don’t have the data ready to power these projects to their full potential. As we all know, the quality of the data being used for AI projects determines whether it sinks or swims,” he explains.

“Rather than write AI off as hype that doesn’t bring business benefit, we need to take a step back and get our data and analytics foundations in order, set clear use cases, and scale AI adoption as projects show their value.”

Adoption issues

 

There is also sector-specific evidence suggesting a slowdown of sorts. According to a recent study by telecoms consultants STL Partners, a total of 62% of new telco GenAI projects aim to monetise AI through enterprise or consumer solutions.

However, the telco market has seen a slowdown in the new announcements, STL found. Between June and August 2025, the number of GenAI projects announced by telcos grew by just 12% – much less than the 41% increase in the preceding three-month period.

According to Digitate’s chief customer officer Ugo Orsi, GenAI is currently in the stage where it is demonstrating its capabilities to enterprises, and this is why there has been so much hype, which might feel like it is losing momentum.

“What we have is an adoption issue, not a capability issue: i.e., Enterprises do not know how to unlock the full potential of Gen AI or worse, others do not have the courage to start the journey,” explains Orsi.

“The bottom line is that the “hype’ perceived at the beginning was fuelled by the ‘curiosity’ of enterprises to understand the art of the possible. Now we are in a different time frame: adoption which is proceeding at a much slower pace.  Gen AI / AI is still an incredible technology, and enterprises need time to adopt to it.”

Squirro founder and CEO Dorian Selz agrees that we are moving from a stage where many businesses were launching pilots, to test the capabilities of GenAI, into a new phase, focussed around value-driven adoption.

“The idea that the AI hype cycle is about to burst misses the bigger picture,” says Selz. “Yes, the craze for AI pilots will be cut sharply. Many such pilots have been started without a clear-cut business case.

“Enterprises are moving past experimentation and focusing more on ROI, integration, and risk management, all of which takes time and discipline. And is harder than many think.”

The impact of regulation – or the lack of clear frameworks – could also be slowing adoption, add Selz, as many large organisations hold back from scaling AI projects because of concerns around liability, compliance, data privacy, and ethical use.

Selz adds: “We continue to see strong demand for AI solutions that deliver measurable outcomes, particularly where governance and transparency are built in from the start. Far from bursting, the AI cycle is entering a more mature phase that prioritises more sustainable and responsible growth over short-term excitement.”

This view was backed by Glean head of solutions engineering EMEA Gavin Guinane, acknowledges that many organisations had rushed into AI pilots over the past 18 months – with only a fraction taking those to scale.

“That isn’t a bubble bursting – it’s a necessary pause as leaders focus on building durable foundations for long-term success,” claims Guinane.

“Enterprises are discovering that AI transformation isn’t about flashy demos. It’s about trust, security, and adaptability. Systems must deliver reliable results, safeguard sensitive information, and work seamlessly across the tools employees already use. Without those fundamentals, adoption will stall.

“The real story is not retreat, but maturity. Companies are moving beyond hype cycles and investing in AI that can scale responsibly, integrate deeply, and deliver measurable impact. Those that focus on these basics will be the ones to turn today’s pilots into lasting transformation.”

 

In part two we will look at why some industry figures believe the hype cycle isn’t bursting and AI is set for even further growth in the coming years

Personalized Feed
Personalized Feed