As investors increasingly focus on whether the technology sector is entering an AI bubble, Goldman Sachs Group Inc (NYSE:GS, ETR:GOS) analysts say current market conditions do not yet point to a dramatic re-rating of technology stocks.
Following the Goldman Sachs Communacopia + Technology Conference in September and recent company announcements on long-term AI capacity, the firm’s Technology, Media, and Telecom research team examined adoption trends, capital deployment, energy requirements, and broader market dynamics.
In a note to clients, Goldman highlighted that “identifying and calling out a bubble has been a historically difficult dynamic for public market analysts and investors.”
Consumer adoption of AI has been rapid, primarily through day-to-day queries and computing activities.
While early monetization has focused on professional subscription models, Goldman said 2026 could be “a key year for monetization initiatives (advertising and commerce) to scale.”
The firm added that the increase in human-computer interaction may act as a long-term secular growth driver, similar to the shift from desktop to mobile computing.
Enterprise adoption has continued to scale internally, with companies deploying AI to reduce operational friction, enhance productivity, and optimize costs.
Although early results over the past 12 to 18 months have been mixed, companies at the conference consistently cited AI as “a source of margin maintenance/upside and a resulting ability to optimize technology industry headcount.”
Revenue-generating enterprise adoption has been less scalable, despite significant stock re-ratings across enterprise and application software in anticipation of AI-driven disruption.
Goldman Sachs highlighted a “wide gap between the current demand for AI services against the capacity that is currently available,” particularly at cloud providers.
Investments aimed at medium- and long-term capacity raise questions about the visibility of returns, with AI-related spending projected to reach $3 trillion to $4 trillion by decade-end.
This level of investment “would require additional examples of profit pool disruption to emerge” across enterprise productivity and consumer spending efficiency.
Datacenter power requirements are also rising, with a “more than 165% rise in power requirements by 2030” expected relative to 2023, according to the analysts.
Goldman projects that 60% of future demand will require new generation, and grid investment expectations have been raised to $780 billion through 2030.
On the question of a bubble, the analysts noted some parallels to the late 1990s, including rising private valuations, intra-industry capital deployment, and stock performance anchored on a single theme.
However, Goldman wrote that factors including lower levels of public market IPOs, scrutiny of AI returns, valuations below late 1990s peaks, and significant profit and capital deployment argue against a current bubble scenario.
The firm concluded, “We continue to monitor levels of technology adoption, technology spend, and their collective impact on the industry/market framing, but don’t see the signs today for any such dramatic re-rating” like those experienced in the early 2000s or during the 2007–2008 market downturns.