Is the AI bubble about to burst? It's not whether the leaders of the biggest tech companies should be trusted.
Some of the world's most powerful tech CEOs gathered around the White House tables praised President Donald Trump for his leadership, saying his approach to AI positions the United States as the dominant player in the emerging sector.
Among them, Mark Zuckerberg, Tim Cook, Sergey Brin, Bill Gates, and Sam Altman manage the companies driving the current AI revolution.
Bryn, co-founder of Google, said: “It's a really incredible inflection point in AI right now, and the fact that your administration is supporting businesses rather than fighting them. That's very important. I think it's a global race and I think you're in CUSP where these AI models are becoming very useful.
Oracle boss Safra Cats added, “AI will change everything,” while Alphabet CEO Sundar Pichai added:
But elsewhere this weekend, it suggests that the AI hype cycle may be losing steam.
Slow speed
Apollo Academy survey released this weekend showed that AI adoption rates in large companies are beginning to slow down.
The survey is based on data from the US Census Bureau, which is released every other week and surveys 1.2 million companies. One question added to the survey since 2023 is whether the business has created products or services in the past two weeks using AI tools such as machine learning, natural language processing, virtual agents, and voice recognition.
Since November 2023, it has increased in all sizes of US companies, and perhaps not surprising, the fastest rise was in companies with over 250 employees. However, the number fell for large businesses in June and July, and even small businesses flattened.
The data measures corporate integration rather than using AI tools entirely, but suggests that adoption of key top-down implementations is slower.
This is backed up with consumption data from RAMP released earlier this year. RAMP's AI index suggests that an increase in meteors in business spending may indicate signs of slowing down.
The figures show that AI penetration among US businesses reached 41.7% as of April, but shows that the growth trajectory has been flattened since the end of last year. It is worth noting that RAMP index data is retrieved among customers and may not be reflected across the US or global market as a whole.
Torsten Sløk, chief economist at Apollo Academy, stresses the slowdown signal warnings among large companies, but Arpit Gupta, an associate professor of finance at NYU Stern, suggests that “the AI Capex signs should be rethinked.”
For Andy Ward, Absolute Security's SVP International, the report highlights growing concerns that companies are beginning to question the return on investment from AI projects.
He states: “AI can translate detection and responses, but when deployed without robust resilience strategies, real-time visibility and clear governance, there is a risk of increasing vulnerability than resolved.
“According to our research, a third (34%) of CISOs have already banned certain AI tools like Deepseek, driven by the fear of privacy invasion and the fear of loss of control. As attackers leverage AI to reduce the gap between vulnerability and exploitation, our defenses need to promote urgency equal to equal urgency.”
ROI Issues
David Tyler, founder of Outlier Technology, suggests that the hype cycle is on the verge of rupture, with MIT data showing that 95% of the generated AI projects showed no value.
He claimed that investors were “blinded by the promises of the hype machine,” and “continuously increased the amount that was thrown into AI.” Key players report losing $2.40 per dollar. It also explains that recent data center economic modeling requires 3.5 billion users, each pays $35 a month, and is evenly uniform just to break. “This represents 64% of all people with internet access worldwide.
“So economics isn't stacked up, most of the projects offer no value, and ChatGpt 5 has been criticized for not having a “game changer” promised prior to its release. All of these have contributed to the increase in mainstream attention in the AI bubble.”
But he says that AI can continue to play an important role in the future if companies change their expectations.
“We need to recognize that AI has specific uses, but there is no single model that will revolutionize the whole world,” adds Tyler. “We must always be part of the process somewhere in the line to humans, and instead of plowing trillions of dollars into the small benefits we can make and the unattainable magic “all” solutions that are all for all men, we should focus on specific processes that can be improved with AI. ”
A recent study from Qlik highlights the issue of ROI. Several important takeaways are shown from the report.
- Only 11% say most of the AI initiatives have brought tangible benefits
- Just 51% evaluate AI using KPIs directly linked to Business Performance
- Almost a quarter (23%) say that the majority of AI use cases are still stuck in the experimental stage
- HR (37%) and finance (30%) are considered the sectors where AI has the least specific profit
James Fisher, Chief Strategy Officer at Qlik, says the data doesn't mean that AI is all about hype.
“That's because many companies don't have the data to make the most of these projects. As we all know, the quality of the data used in AI projects determines whether it sinks or swims,” he explains.
“Instead of writing down AI as a hype that doesn't bring about business benefits, we need to take a step back, get the data and analytics foundations in turn, set clear use cases, and expand the adoption of AI so that the project shows value.”
Adoption issues
There is also sector-specific evidence suggesting a type of slowdown. A recent survey by Telecoms Consultants STL Partners shows that a total of 62% of new Telco Genai projects aims to monetize AI through business or consumer solutions.
However, in the telco market, STL has found that new announcements are slowing down. Between June and August 2025, the number of genai projects announced by carriers rose by just 12%. This is far less than the 41% increase in the previous three months.
According to Ugo Orsi, Digitate's chief customer officer, Genai is currently in the process of demonstrating its capabilities with businesses. This is why there was so much hype.
“What we have is adoption issues, not ability issues. That means companies don't know how to unlock ways to reach their full potential in Gen AI. Others don't have the courage to start a journey,” explains Orsi.
“The bottom line is that “first perceived hype” was supported by the corporate 'curiosity' to understand the art of possibilities. Now we are in a different time frame: recruitment is progressing at a much slower pace. Gen AI/AI is still an incredible technology, and companies need time to adopt it. ”
Dorian Selz, founder and CEO of Squirro, agrees that many companies are moving from a new phase that has launched pilots, tested Genai's capabilities and focused on value-driven adoption.
“The idea that the AI hype cycle is about to burst will miss the bigger picture,” Selz says. “Yes, the AI pilot epidemic will be reduced sharply. Many such pilots have been launched without a clear business case.
“Companies are doing past experiments and focus on ROI, integration and risk management. All of this requires time and discipline. It's more difficult than most people think.”
Many large organizations are preventing AI projects from scaling due to concerns about responsibility, compliance, data privacy and ethical use, so the impact of regulations — or the lack of a clear framework may also be delaying adoption.
Selz adds: “AI cycles that are far from bursting will enter a mature stage that prioritize sustainable and responsible growth over short-term excitement, particularly when governance and transparency are built from the start.”
This view is supported by Solutions Engineering Emea Gavin Guinane's Glean Head, and many organizations have admitted to AI pilots over the past 18 months.
“It's not a bubble burst. It's a necessary pause as it focuses on leaders who are focused on building a durable foundation for long-term success,” says Ginane.
“Companies are discovering that AI transformation is not about flashy demonstrations. It's about trust, security and adaptability. Systems need to protect sensitive information and work seamlessly across the tools employees already use. Without these basics, recruitment will stall.
“The real story isn't retreating, it's maturity. Companies are investing in AI that can move beyond the hype cycle, expand responsibly, expand, deeply integrated, and have measurable impact. Those focused on these fundamentals are what transform today's pilots into lasting transformation.”
Part 2 examines why some industry figures believe the hype cycle hasn't bursted.
