2025 was a stock market year of contrasts. A year that began with a political shock, continued with a strong recovery – and ultimately left more questions than certainties. The so-called “Trump Tariff” crash in the spring briefly sent tech stocks into a steep decline, but the shares of the “Magnificent Seven” subsequently rebounded sustainably to their old highs. However, for investors in the Big Tech sector, surprisingly little remained in net terms. With two notable exceptions: Alphabet and Nvidia.
Continue after ad
While the broader market – measured by the benchmark S&P 500 – gained around 16 percent in 2025, only two of the “Mag 7” managed to beat this benchmark in the end: Alphabet with a gain of around 63 percent and Nvidia with a good 30 percent. The rest, however, lagged the US benchmark: Microsoft and Tesla recorded value increases of 13 percent, Apple of 11 percent, Meta achieved around 10 percent, while Amazon barely exceeded 2 percent. For Big Tech standards, it was a rather surprising year of stagnation. The view into the portfolio was even more disappointing for German investors. Due to the strong dollar devaluation, US stocks in euros suffered a hefty discount of 13 percent. In net terms, German owners of Big Tech stocks only made money with Nvidia and Alphabet.
Alphabet: From a written-off laggard to the success story of the year
That Alphabet, of all companies, would become the stock market star of 2025 was considered quite unlikely at the turn of the year: The Google parent company was then branded as a loser in the AI race. Gemini was considered immature, the search market threatened, and ChatGPT a potential gravedigger of a business model that had dominated for decades. Wall Street was merciless – and overlooked how robust the fundamental reality actually was.
What followed in the second half of the year was one of the most impressive re-ratings in recent stock market history. Alphabet caught up in the AI race, technologically matched ChatGPT, and established Gemini as at least a serious platform. At the same time, tailwinds from unexpected quarters provided additional momentum: The major antitrust lawsuit ended without a breakup. And as a cherry on top, it was announced in the fall that Berkshire Hathaway, led by star investor Warren Buffett, had built a four-billion-dollar position, an enormous psychological effect for the tech giant. For years, it had been perceived more as a mature cash flow value than a growth story.
As a result, Alphabet doubled its stock market value within six months. At times, the internet pioneer from Mountain View even broke the four-trillion-dollar mark in market capitalization. Suddenly, Alphabet was on par with Apple, close behind Nvidia, and left Microsoft behind. A scenario that few would have thought possible at the beginning of 2025.
Fundamentally, this revaluation was fueled by several sources: Google Cloud gained market share, YouTube is increasingly developing into a global entertainment and commerce machine, and the seemingly aging advertising business picked up again with double-digit growth. Above all, however, perception changed: Alphabet was no longer considered an AI laggard.
Continue after ad
Nvidia: The backbone of the AI trade – and first signs of fatigue
Meanwhile, Nvidia remained the industrial backbone of the AI revolution in 2025 and even became the most valuable company in the world in the summer. Revenues continued to grow at a breathtaking pace, margins remained historically high, and demand for Hopper and Blackwell chips exceeded supply. For hyperscalers, cloud providers, and AI startups, there was no way around Nvidia.
And yet, 2025 was a year in which cracks in the narrative became visible for the first time. Not fundamentally, as growth in profit and revenue continues to be extremely dynamic at over 60 percent. But the expectation remains equally ambitious: In the second half of the year, the tone on Wall Street noticeably shifted – from boundless euphoria to greater skepticism about how long CEO Jensen Huang can maintain the growth pace. It was increasingly questioned how much of the AI growth fantasy is now priced into the Nvidia stock.
Within the tech industry itself, concerns are dismissed. Tech star analyst Dan Ives, for example, stated two weeks ago that the major AI bull market cycle could last for another two years. Lisa Su, CEO of chip giant AMD, echoed this sentiment weeks ago: The speed at which AI is improving is higher than any technology she has ever experienced. A compliment to the tech industry – but also an indication of how short-lived innovation cycles have become. This is precisely where the skeptics’ criticism comes in.
OpenAI, the mood swing – and the big bubble debate
As is well known, the AI boom of recent years originated with OpenAI. The launch of ChatGPT three years ago triggered a chain reaction: billions in investments, rising capex budgets at hyperscalers, massive orders for Nvidia, strategic alliances with Microsoft, and infrastructure deals up to Oracle. However, in the fall of 2025, the mood began to shift.
Suddenly, an old question dominated the debate in the capital markets again: Is this really a sustainable industrial revolution – or the overshooting phase of a classic tech bubble? Memories of the dot-com bubble at the turn of the millennium, of Cisco, of the year 2000 kept returning.
One of the loudest voices of caution was Michael Burry. The hedge fund manager, known from the Hollywood blockbuster “The Big Short,” drew a provocative parallel: Nvidia is not Enron – but Cisco. Around the turn of the millennium, Cisco was the indispensable foundation of the internet age – technically dominant, highly profitable, seemingly irreplaceable. And yet, after the peak in early 2000, a price drop of almost 90 percent followed, and a decade-long sideways trend, which only ended recently with record highs after a quarter of a century.
Burry’s argument is less a crash scenario than a valuation warning. Even structural winners can disappoint investors if expectations are astronomical. He is particularly critical of the depreciation practices for AI hardware: If servers become obsolete faster than the capitalized years suggest, profits could come under pressure in the future – an issue that has so far hardly appeared in the consensus of analyst models.
2026: Mega IPOs from OpenAI, Anthropic and SpaceX expected
Accordingly, 2025 is likely to be remembered as the first year of differentiation in the AI hype. Whether this ushered in the beginning of the end of the great AI supercycle or merely a phase of consolidation remains to be seen. So far, the AI boom has mainly played out indirectly: Investors have bet on infrastructure beneficiaries like Nvidia, cloud platforms like Microsoft or Alphabet, and capex winners along the value chain. However, the next 12 months are likely to be a litmus test for tech and AI investors.
With the potential IPOs of OpenAI and Anthropic, the AI hype is entering a new phase: it is becoming much more investable. In addition, there is another elephant in the room: SpaceX. A possible IPO – whether as a whole company or through individual divisions like Starlink – would not only be one of the largest IPO events in stock market history but also a stress test for the current market environment. If the mega-IPOs, with valuations expected to exceed $300 billion (Anthropic), over $500 billion (OpenAI), and even over $800 billion (SpaceX), are successful, it would be an enormously strong signal.
Risk appetite is alive, the tech bull market continues. However, if one or more of these IPOs fail – due to investor reluctance, valuation discounts, or general weakness in the US economy – the AI euphoria is likely to be significantly curbed. 2026 will thus become the year of truth for Big Tech and the AI ecosystem.
(wpl)