Newspapers and journals have recently been flooded with reports about Artificial Intelligence and that the ‘AI bubble’ is ready to pop. The idea is certainly a concerning one.
Often compared to the late 1990s dot-com bubble, economists are worried that the trillions being invested in AI has vastly outpaced the sector’s profitability. J.P. Morgan Chase anticipates that $5 trillion will be invested in AI infrastructure between now and 2030, which, when compared with the relatively modest revenues of OpenAI and Anthropic of around $44 billion, is setting more than a few nerves on edge. Investor concerns are further amplified by the fact that nearly 80% of stock gains in 2025 were driven by the ‘Magnificent Seven’ (Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla). Overly-optimistic valuations concentrated in such a small number of ticker symbols means that investors are increasingly vulnerable to a crash in the stock price of just a few companies. The Bank of England has warned that the US-Iran war could tip the bubble over the edge – things appear to be fragile.
However, not all is doom and gloom. The dot-com bubble, when it crashed, ultimately pushed the tech sector to innovate [pdf] and was followed by investment in internet infrastructure. Nothing is guaranteed, but whether the bubble pops or not, one thing is for sure: AI’s profit potential has not yet been fully realised.
It’s not just OpenAI and Anthropic who aren’t generating the desired ROI. An MIT study from 2025 found that 95% of businesses that are attempting to implement generative AI are failing to achieve the desired results. Investment is being wasted, and so the problem is one that consultants to tackle head-on.
How should businesses be maximising AI’s profitability?
Megan Butler, AI Workforce lead at KPMG believes the potential is there. There are three ways that businesses can more successfully make it happen.
1. Strategy: more nuanced integration
Businesses need to go beyond investment to proper integration. AI needs to become central to how businesses operate, beyond just the abstracts of ‘saving time’ or ‘cutting costs’. So, how can businesses actually implement this?
Well, for starters, a nuanced analysis of exactly where AI can help is necessary. The human aspects of business, such as client and staff relationships, cannot be replaced by AI – so don’t waste time or resources on this. Identify specific areas of business that AI can do, as well as evaluating how well it can do them. For example, research tasks often require human checking to avoid AI hallucinations or missed information. On the other hand, when it comes to analysing data or summarising information, AI is invaluable.
Recognising these nuances and understanding where AI is most useful will allow businesses to automate the right processes and make AI the core of that function, rather than an optional add-on tool.
2. Banish fear: changing social behaviour
Peter Drucker famously stated that ‘culture eats strategy for breakfast’. In other words, having a goal is not enough. The most important thing is to get people on board and to banish fear by normalising the use of AI in the culture of the company.
There is a fear surrounding the non-human nature of AI, which makes some businesses hesitant to fully deploy the technology. The anxious vision of a dystopian, robot-ruled world pervades various aspects of society. From the tearing down of subway advertisements in Paris for Friend.com, an AI ‘companion’ business to the preference for human-generated content on Reddit, there is a clear anxiety about the encroachment of the machines and a desire for authentic human connection.
The key to overcoming the AI taboo is twofold:
- Engage people in conversation about AI. Find out what their concerns are and address them.
- Communicate the areas of value that human workers provide. Reassure staff that AI is a tool to empower them, not something that will replace them.
3. Contingency: prepare for the bubble to pop
Given the aforementioned concerns about the popping of the AI bubble, businesses should prepare. Given the current geopolitical climate, contingency planning for a potential AI crash is not only wise, but essential.
Key takeaway
Whether the AI bubble pops or not is not in the hands of individual business leaders or consultants – all that can be done is to prepare for this eventuality.
Where consultants can help is by helping businesses maximise the profitability from their AI investments.
To do this, nuanced integration needs to replace abstract and blanket approaches, and the social norms surrounding AI adoption need to move towards a more AI-positive atmosphere.
India Jordan Jones is a final-year undergraduate student at the University of Oxford, reading English Language and Literature. She is interested in a career in consulting or commercial law and passionate about sustainability and energy matters in business.
Image: DALL-E
🔴 Found these ideas useful?
Sharpen your edge