AI is at an extreme: Market cap, returns, capex surpass tech-bubble highs

At the moment, global markets are dancing entirely to the tune of artificial intelligence (AI). It is everywhere, and either you believe and participate in the bull market, or you sit on the sidelines, doomed to mediocre relative performance while waiting for sanity to return. As of today, the bulls have the upper hand, as markets are convinced that AI is transformational and here to stay. It will determine the future winners and losers, both at the company and sovereign level. At the beginning of an industry platform shift, one should not worry about valuations, just focus on buying the winners, so goes the current thinking. 

The believers are convinced that AI will solve the demographic and fiscal challenges of the West. Through a surge in productivity, growth will accelerate, solving the debt trap problem that most Western democracies are on the precipice of. We don’t have to inflate our way out of the fiscal challenges, AI led growth will solve the budget arithmetic. 

With a need for fewer workers, especially at entry level, the demographic challenges of the West will be less corrosive. AI will also widen the span of corporate outcomes, with the winners seeing a surge in growth and margins and the losers going out of business. America is seen as a winner in the AI race, as is China. The AI trade is in full force, whether you look at the internals of the US market or consider the outperformance of China/Korea and Taiwan. India is seen as an AI loser or laggard. 

While one can debate the ultimate outcome, timing, and true benefits of AI, only the passage of time will reveal its real impact. The trade and thematic are currently all-pervasive. AI carries extreme weight in the markets, and measures of concentration across market capitalisation, returns, earnings, and capital expenditure are at all-time highs. 

Today in the US, 65 per cent of all venture capital investment is going into AI- or machine-learning–based startups. OpenAI is poised to raise money at a $500 billion valuation, despite its own internal projections showing it will lose over $100 billion over the next five years. Every single large language model (LLM) company is raising billions in new capital, and Oracle has seen a $240 billion single-day surge in market capitalisation based on a $300 billion five-year Cloud deal signed with OpenAI (which has estimated current revenues of $12 billion and no visibility on free cash flows). Oracle’s pop in market capitalisation needs the markets to stay buoyant so that OpenAI can raise the cash required to fund the contract. Circular logic at its best!  

Today, the Magnificent Seven (Mag-7) holds a 32 per cent weight in the S&P 500. In January 2023, just after ChatGPT was launched, this number was only 18 per cent. Nvidia, with an 8 per cent weight in the S&P 500, now has the largest single-stock weight in the history of the index. Its current market capitalisation is equivalent to 15 per cent of US gross domestic product! 

The S&P TMT (tech/media/telecom) weight at 45 per cent, is exactly the same as it was in the tech bubble of 2000. The difference of course is that this is all AI now, with telecom and media almost non-existent. If we look at the top 10 companies in the S&P 500 (basically the Mag-7, Broadcom, Berkshire and JPMorgan), they account for a record 40 per cent share of the index and 25 per cent share of corporate earnings. We have never seen such concentration of company size and earnings. 

Even the returns of the index are extremely concentrated. Since January 2021, 55 per cent of the entire gain in the S&P 500 was accounted for by the top 10 stocks. If you were not adequately invested in these 10 giants, you had no chance to keep up with the broader markets. This concentrated returns profile continues to date, with the Mag-7 up about 50 per cent since Liberation Day (April 2, 2025), while the remaining 493 S&P 500 stocks have gained only 20 per cent. 

Even the earnings profile of the market is heavily skewed towards the Mag-7 and the AI theme. In 2023 and 2024, the Mag-7  saw earnings growth of about 35 per cent within the S&P 500, while earnings for the remaining 493 stocks grew only 3 per cent. Consequently, the relative performance of the technology sector versus the broader S&P 500 index is even stronger today than what we saw at the peak of the 2000 bubble. 

Even when we look at corporate capital expenditure, the concentration in technology and AI is remarkable. The Mag-7 and Oracle account for over 35 per cent of total S&P 500 capex. US hyperscalers (the major tech companies) have doubled their share of private domestic investment since 2023. For these hyperscalers, capex has now crossed 20 per cent of sales, compared with under 10 per cent previously. Even on operating cash flow, they are using over 65 per cent to fund data centre buildouts. To put this in perspective, their capex-to-sales ratio is 20 per cent, and research & development-to-sales is 15 per cent, meaning 35 per cent of sales is being reinvested into growth. Truly unprecedented numbers. 

Telecom companies, with their massive fibre-optic buildouts, were seen as the poster child of overinvestment during the 2000 technology bubble peak. At their peak in 2000, telecom companies’ capital expenditure accounted for 0.8 per cent of US gross domestic product. Today, hyperscalers’ capex is already at 1.2 per cent of US gross domestic product (GDP), with the current projection being that this number will cross 1.4 per cent by 2028, a full 75 per cent higher than the telecom capex peak during the dot-com bubble1. ( source: Apollo Global Chartbook). 

While all these numbers can keep rising, we seem to be in uncharted territory as far as the measures of concentration for market capitalisation, returns, and capex go — all driven by AI and the belief that it will change our world. 

This hype and hysteria will have to come back to earth at some stage. AI may be truly revolutionary technology, but can still be a bad investment from these elevated levels. When the hype cycle deflates, the Mag-7, China/Korea/Taiwan and numerous other trades will unwind. India has not participated in the AI trade and is seen as an AI loser. We may need to wait for the AI hype cycle to turn for India to start attracting foreign capital again. In the meantime, we should focus on kickstarting our domestic economy and implement a growth agenda. India may never fully participate in the AI hype cycle, but we can position ourselves to benefit from its inevitable disenchantment and the cycle of disillusionment. 

 

The author is with Amansa Capital  1. All data from Apollo Global Chartbook