Will AI really replace software?

Starting from January 2, 2026, the soaring prices of AI storage (led by companies such as Western Digital(NASDAQ:WDC), Seagate Technology(NASDAQ:STX), Micron Technology Inc(NASDAQ:MU), and SanDisk(NASDAQ:SNDK)) , semiconductor equipment (such as ASML Holding ADR(NASDAQ:ASML), Lam Research Corp(NASDAQ:LRCX), and Applied Materials(NASDAQ:AMAT)), chip manufacturing (such as Taiwan Semiconductor(NYSE:TSM) and Intel Corporation(NASDAQ:INTC)), as well as gold (ETF: SPDR Gold Shares(NYSE:GLD)) and silver (ETF: iShares Silver Trust(NYSE:SLV)) have had a severe draining effect on niche sectors, with the software sector being the most severely impacted. From January 8, 2026, major software stocks generally began to decline. The situation persisted until early February 2026, when the fragile software sector encountered panic selling. The S&P North American Software Index plummeted by 15% in a single month, marking its largest decline since the 2008 financial crisis. This sharp drop not only inflicted heavy losses on investors but also triggered intense market discussions about whether "AI will replace software."

On February 3, the U.S. AI sector experienced panic selling due to the simultaneous emergence of multiple negative factors. The automated tools launched by AI startup Anthropic sparked market concerns about AI replacing human labor, causing the stock prices of legal software companies such as Thomson Reuters(NASDAQ:TRI) to plunge by over 15% in a single day. According to Goldman Sachs data, the S&P North American Software Index has declined for three consecutive weeks, with a 15% plunge in January, marking its largest single-month decline since October 2008. Jeffrey Favuzza from the equity trading department at Jefferies described the current situation as the "SaaS apocalypse," noting that the current trading style is entirely characterized by panic selling with the mentality of "get me out at any cost," and no signs of a bottom or stabilization have been observed yet.

The U.S. software sector includes, but is not limited to, the following companies: Microsoft Corporation(NASDAQ:MSFT), Adobe Systems Incorporated(NASDAQ:ADBE), Salesforce Inc(NYSE:CRM), Intuit(NASDAQ:INTU), ServiceNow Inc(NYSE:NOW), SAP SE ADR(NYSE:SAP)... 

The market's fear of AI replacing software is not unfounded. As AI technology continues to advance, its capabilities in automation, data processing, and decision support are growing stronger, making the moats of traditional software companies increasingly shallow. The new automated tools launched by Anthropic can automate tasks such as contract review and legal brief preparation, directly impacting the core businesses of legal software companies. Similarly, Alphabet Inc C(NASDAQ:GOOG)'s Project Genie tool, which can create immersive virtual worlds through text or image prompts, poses a threat to video game stocks (such as Unity Software(NYSE:U)).

Despite the widespread market panic, numerous industry leaders and experts believe that AI will not completely replace software but will instead drive the transformation and upgrading of the software industry. Nvidia CEO Jensen Huang, while attending an AI conference hosted by Cisco in San Francisco, explicitly refuted the view that "AI will replace software," calling it "illogical." He emphasized that the development of AI cannot be separated from the existing software ecosystem, and that the design of current software tools features clear and well-defined characteristics, providing solid support for AI applications.

Huang's viewpoint has been widely recognized within the industry. In fact, AI is driving the software industry to transition from "software-defined" to "physical intelligence." Physical AI, as the next climax of AI development, focuses on enabling intelligent agents to perceive, reason, and execute actions in the physical world. This means that AI is no longer confined to on-screen dialog boxes but is integrated into physical hardware such as humanoid robots, autonomous vehicles, and drones, becoming "embodied intelligence."

During this transformation process, the role of software has not been weakened but has instead become even more crucial. Physical AI requires an extremely complex software stack for support, including world models, physics engines, digital twin platforms, and more. Traditional software code is being transformed into "action strategies" for AI, while software engineers need to acquire new skills in building high-quality physical data and simulation environments.

Therefore, AI should not be seen as replacing software but rather as an auxiliary medium that promotes the transformation and upgrading of software, making it more complete, intelligent, and user-friendly.