Amazon, Meta, Alphabet shares plunge as record AI capex worries investors

Artificial Intelligence has once again sent shockwaves in stock markets by expanding its capacity and reach, where it is now planning to replace human in white collar jobs. This threat was not new after knowing the capabilities of Artificial Intelligence. However, it could reach at this stage so early, is what most of us are stunned about. Many industry leaders are dismissing the claims, while some are warning to look for alternatives. Irrespective of the when the actual replacement happens, stock markets are already reacting to it in that way.

However, we are witnessing a paradoxical reaction in the stock market currently, in relation to the AI. Shares of software services companies are falling, because new AI models are threatening the traditional SaaS model. At the same time, shares of companies that are investing heavily in AI are also plunging 10-12%. Let us understand why there is such a paradoxical situation in the market.

Why are software stocks plummeting?

Software companies which provide enterprise level software solutions to corporations are now under the threat of going obsolete as Anthropic AI’s new model is touted to replace humans with coding skills. The really worry lies not in loosing business contracts, but the repricing of the business model as traditional model involves higher marketing and sales spends, which help gaining contracts.

With AI doing the same work at fraction of a cost, traditional software companies loose the edge and consequently loose the competitive advantage. Acknowledging these threats and risks, the investors are now repricing the valuations and exiting the most prone stocks leading to massive selloff in software stocks across the globe.

Why megacap AI-enablers falling?

Megacap companies like Alphabet, Meta, Amazon, Apple, Microsoft and Tesla, which are know becoming AI enablers are also witnessing sharp plunge in their stock price. Majority of these companies have now announced record level spending in AI infrastructure and data centre capacities.

All these major companies shares fell 10-20% post quarterly earnings, despite most of the comapanies reporting steady growth in earnings. The primary reason for soured investor reaction was due to doubling down on the expenditure in AI infrastructure.The total expenditure by these companies is expected to cross $630 billion, which could eclipse few of small countries budgets. The record level spending received less excitement and optimism and more of caution from investors.

Does heavy capex in AI lead to heavy returns?

The answer to this question is not straightforward yes or no. Since the launch of AI in November 2022, there has been several launches of new models with more advanced capabilities, but very few of them are witnessing positive yields as of date. This worries investors more than anything else as they sound cautious about the aggressive bets without no empirical evidence of it yeilding better returns. Recent study by Deloitte showed that a meaningful ROI from AI can only be expected after 2-4 years as compared to 12 months for traditional software services applications.

Additionally, the AI companies like Open AI, Anthropic, Perplexity and Gemini too are yet to derive any meaningful bottomline results. Recently,the Open AI CEO said annualised revenue of crossed $20 billion in 2025, but could could not provide any further details on operational profitability. Open AI aims to raise $100 million at the valuation of $830 billion.

Anthropic AI, which has rattled the software company stocks, has reportedly generated $9 billion in revenue in 2025, but the company aims to spend $12 billion in training models and another $7 billion running them. It also aims to raise $10 billion at $350 billion valuation. While the Perplexity AI has annualised run rate of $148 million for 2025 and holds valuation of $20 billion. The financial metrics for Gemini are not available, but the google’s cloud segment revenue stood at $17.5 billion with revenue from Generative AI model products jumping 400%.

Is 2026 a Deja vu of 1999 dot com bubble?

They say "history never repeats, but often rhymes". Something similar is visible in the current AI scenario, where the parallels are drawn to 1999 dot com bubble. Similar to the current record spending on AI infrastructure, the internet companies spent heavily on optic fibre cables with expectation of internet traffic would double every 100 days. However, 95% of the capacity lied unused nearly for a decade after the crash.

Looking at the initial reaction of investors on record AI-capex, investors feel, the history could repeat in AI space as well. Investors feel the inflated and aggressive bets in AI infrastructure and data centers hold the similar risks of 1999 bubble burst. Will it happen sooner or later, only time will tell.