What is an AI Bubble & its Risks to Enterprise AI Strategy?
The AI sector is experiencing what many analysts describe as ‘bubble’ conditions.
As unprecedented investment levels drive valuations that approach peaks last seen during the dotcom era of the late 1990s, UBS, the investment bank and financial services firm, gathers research.
The research shows the risks and broader implications of the ‘AI bubble’ as it appears in more and more business discussions across the world.
What is an AI bubble?
An AI bubble refers to a situation where companies developing AI technologies are valued far beyond their current revenue streams and proven business models.
This occurs when investor enthusiasm for AI’s potential drives stock prices and investment commitments based on projected rather than demonstrated returns.
The scale of current AI investment provides context for these concerns.
Sam Altman, CEO of OpenAI sums it up when speaking to Business Insider: “When bubbles happen, smart people get overexcited about a kernel of truth. Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes.
“Is AI the most important thing to happen in a very long time? My opinion is also yes.”
The bigger picture
UBS finds that technology companies have committed capital expenditures and research budgets that rival entire industries, with 2024 spending alone surpassing the total research and development expenditure of all listed European companies combined.
- The AI bubble is when AI companies are valued much higher than their current profits justify, driven by hype and big investments hoping for future success. This risks big losses if AI growth or profits don’t meet expectations, similar to the dotcom bubble, with challenges like high costs, energy use and global competition adding uncertainty.
This investment surge reflects the strategic importance and pressure that major players place on securing dominant positions in the emerging AI market.
Furthermore, leading technology companies generate a substantial share of overall economic profit in the US, providing the financial foundation necessary to fund large-scale AI development programmes.
These companies’ strong balance sheets have enabled continued investment even as questions mount about returns.
The speculative foundations that create valuation risks
UBS reports that the fundamental problem lies in the speculative nature of many AI use cases currently driving investment decisions.
Revenue potential from AI applications remains largely projected into the future rather than realised through current operations, creating a disconnect between spending and proven returns.
Joe Tsai, Alibaba’s Cofounder says to Business Insider: “I start to see the beginning of some kind of bubble... I start to get worried when people are building data centers on spec.”
Machine learning (ML) systems promise to transform industries from healthcare to financial services.
However, many of these applications remain in experimental phases, with commercial viability dependent on technological breakthroughs and market adoption that may take years to materialise.
Current technology sector valuations incorporate expectations for AI-driven growth that leave little margin for error should cash flows fall short of projections.
This mirrors conditions during the dotcom bubble, when internet companies commanded premium prices based on anticipated rather than actual profitability.
When reality failed to meet expectations, significant price corrections followed.
The uncertainties threatening returns
Several factors could affect the ability of AI investments to deliver expected returns over the medium term.
“The bubble talk is completely wrong. AI will fundamentally change everything over the next five years.”
UBS finds that returns on capital investments in AI infrastructure remain unclear, as companies build data centres and purchase computing equipment without proven revenue models to justify the expenditure.
Energy supply constraints present another challenge to AI expansion plans.
The computational requirements for training and operating AI systems demand enormous electricity consumption, creating infrastructure bottlenecks that could limit deployment speeds and increase operational costs beyond current projections.
Rising global competition adds further uncertainty to return calculations.
How the change in AI competition could impact the AI bubble crisis
Regions with rapidly increasing research and development activity are developing their own AI capabilities, potentially reducing the competitive advantages that current technology leaders expect to maintain.
This competitive pressure could erode profit margins and market share assumptions built into current valuations.
The timeline for widespread AI adoption across traditional industries remains uncertain.
Many potential applications face regulatory hurdles, technical challenges and institutional resistance that could delay commercialisation beyond investor expectations.
These uncertainties create conditions where substantial portions of current AI investment may not generate anticipated returns.
Companies are essentially placing bets on future technological capabilities and market adoption rates that remain largely unproven, while stock prices reflect assumptions about success that may prove overly optimistic.
The risk is that when reality meets expectations, the resulting corrections could reshape technology sector valuations and broader economic projections that assume AI will drive significant productivity improvements across multiple industries.
Lisa Su, AMD CEO says to Business Insider: “The bubble talk is completely wrong. AI will fundamentally change everything over the next five years.”
Whereas, Thomas Siebel, CEO C3.ai says: “There is absolutely an AI bubble and it’s huge. The market is way overvaluing some startups.”
