Humanoid robot kick boxers might disagree but Bill Gates, the Bank of England and the IMF all say that we are in the middle of an AI bubble. To add to the impending sense of doom a Massachusetts Institute of Technology (MIT) report estimates 95% of corporate generative AI investments are failures.
For Bill Gates it is not as bad a bubble as the 17th century Dutch Tulip Mania which saw investors in – you’ve guessed it – tulips, wiped out by the first recorded speculative bubble in modern financial history, but it could be as bad as the late dotcom crash of the 2000’s, he believes.
The Bank of England Financial Policy Committee politely warns: “On a number of measures, equity market valuations appear stretched, particularly for technology companies focused on Artificial Intelligence. This, when combined with increasing concentration within market indices, leaves equity markets particularly exposed should expectations around the impact of AI become less optimistic.”
IMF chief economist Pierre-Olivier Gourinchas warns markets are overheated with shareholders at risk but, reassuringly, believes the bursting of the bubble is less likely to be a systemic event. The MIT report “The GenAI Divide: State of AI in Business 2025” examined 300 AI deployments, spoke to companies and employees and concluded most AI pilots fail in the early stages.
The humanoid robot kick boxers meanwhile are warming up for their next televised event in Shenzen, China, in December. The first televised show in May demonstrated that humanoid kickboxing apparently still has some way to go as a TV sport. But the event organised by robotics company EngineAI is more about China showcasing the latest advances in AI and robotics technology. Worries about AI hype appear less important.
Pumping Up The AI Bubble
Nvidia ceo Jensen Huang does not believe in the bubble and as boss of the world’s first $5 trillion market cap company which produces computer chips for AI processing, he would understandably make this point. Nvidea recently announced a $500 billion AI chip order and plans to build seven supercomputers for the U.S. Government.
Huang told Bloomberg TV, “All of these different AI models we’re using – we’re using plenty of services and paying happily to do it.”
Huang’s generous use of the pronoun “we” is insightful. It is estimated that over one billion people interact with AI every day, 5o-70 million are Americans or approximately 15-20% of the U.S. population, and many of these services are free to the end user. ChatGPT was the fastest growing app in history reaching 100 million uses in two months in 2023.
Recent quarterly earnings from companies Alphabet, Meta and Microsoft outlined plans for huge capital expenditure on AI infrastructure. Alphabet’s Google boosted its spending plans for this year from $8 billion to $93 billion while Meta signaled spending could hit $100 billion next year. Microsoft forecast spending of around $140 billion next year.
There is a lot more spending to come. McKinsey estimates that data centres for AI will need $5.2 trillion in capex by 2030 to keep up with AI “compute” demand. It will not all come from the Mag7 but will need external financing, including debt.
Morgan Stanley Research estimate that AI Infrastructure capex between 2025 and 2028 will be $2.9 trillion, with $1.5 trillion expected to be met by external capital, including $800 billion from private credit.
AI has a wider impact than capex and Amazon’s recent plans to cut 30,000 corporate employees illustrates that and how AI is not just about a bubble. The job cuts are explicitly aimed at making the company leaner and Amazon’s ceo Andy Jassy has said that as AI automates more repetitive and routine work more jobs will go.
Beyond The AI Bubble
Big numbers and forecasts of what the future will bring risk ignoring the impact on businesses now and how AI is being applied. AI has had a strong start in financial services across blockchain data networks and in the highly regulated investment management industry.
Investment management technology provider Clearwater Analytics, for instance, has recently launched CWAN GenAI, an embedded generative AI platform that’s already deployed across more than $10 trillion in institutional assets.
Unlike other systems which graft AI tools on to legacy systems, it is fully integrated into production operations, delivering results including a 90% reduction in manual reconciliation, 80% faster regulatory reporting and 50% faster financial close cycles. The platform has around 800 AI agents created by clients and internal teams.
Sandeep Sahai, Clearwater ceo says, “This is about fundamentally re-imagining how institutional investment operations function in real-time.”
Robocap, a U.K. fund manager, commissioned independent research company Pureprofile to interview 100 senior professional investors at pension funds, insurance asset managers, family offices and wealth managers, who collectively manage $1.183 trillion in assets.
All survey participants agreed that robotics and AI offer exciting investment opportunities, with third (32%) strongly agreeing with this view. These are the types institutional investors that provide the finance for AI infrastructure capex as well as investing in the sector through any bubbles and beyond.
Jonathan Cohen, founder and cio of Robocap says, “We have already seen huge growth in the AI and robotics sectors and given the major advances in computing power, big data and AI models, investors there will be an explosion in investment in the next decade.
“AI and robotics attract sensational headlines, but they’re here to stay and already fulfil an important role in human society by doing difficult tasks, quickly and efficiently. The potential for AI and robotics to continue to transform the world, creating even more investment opportunities is also here to stay.”
Robocap is a leading investor in robotics, automation and AI listed stocks, with pension funds, insurance asset managers, family offices and wealth managers who collectively manage $1.183 trillion in assets.
Believe Some Of The AI Hype
The research found all professional investors questioned had increased allocations to robotics and physical AI investments and they expect this trend to continue. Over the next three years, 95% of professional investors said allocations will increase including 15% predicting a dramatic increase.
They are all coming under pressure currently from clients to invest more into AI and robotics companies and 90% expect that pressure to continue to grow over the next three years.
They are not immune from the worries about an AI bubble – nearly two out of five (37%) of investors questioned were very concerned about the issue of making false claims about the impact of AI on operations.
AI washing can also involve companies claiming to use AI when in fact they are using less-sophisticated algorithms or overstating the efficacy of their AI over existing techniques, or even falsely claiming that their AI solutions are fully operational when they are not.
Institutional investors are hard-headed about a range of other AI issues – 80% are concerned about privacy and data security, while nearly three-quarters (71%) are worried about technological vulnerabilities and the potential for AI to be hacked or manipulated.
Two-thirds of respondents said they are most concerned about autonomous AI systems making decisions without human intervention, while 61% are focused on the potential for job displacement and a negative impact on employment, as in the case of Amazon.
Concern about those issues is however superseded by worries about what over-regulation will do to the sector. Around 35% of those questioned strongly agree and 65% slightly agree that regulation in the U.K. and E.U. around AI and robotics is too stringent and that has curtailed creativity and innovation.
In other markets such as the U.S. and China, with its kick boxing robots there is a more relaxed approach which benefits companies there.
Cohen adds guidance, “At a time when countries are competing to attract the biggest and most successful businesses, it makes sense to have progressive legal frameworks in place and it is essential that policymakers support these businesses by creating long-term plans that help them achieve their full potential.
“A robust AI and robotics strategy is seen as material to the success, growth and profitability of today’s leading companies.”
This guidance also applies to humanoid kick boxers and AI bubbles.