AI shocks markets and sectors worldwide

Recent AI developments are rippling across global markets, with investors revaluing software, infrastructure and service businesses. In February, multiple large-language-model launches and AI automation tools coincided with sharp share price moves in listed software firms and data providers. Coverage described an “AI scare trade” in which investors sold companies seen as vulnerable to AI automating white‑collar and advisory work, even when near‑term earnings remained solid.

ServiceNow reported fourth‑quarter 2025 revenue of $3.6 billion, up 20.66% year over year, with 98% renewal rates and 35% free‑cash‑flow margins. Despite these metrics and a new $5 billion buyback, the stock sold off as traders focused on the risk that AI agents could disrupt seat‑based licensing. Similarly, real estate brokers such as CBRE, JLL and Cushman & Wakefield fell as markets weighed whether AI could erode fee-heavy advisory models.

India is becoming one of the largest AI user bases for global platforms. OpenAI CEO Sam Altman wrote that India now has 100 million weekly active ChatGPT users, making it the company’s second‑largest market after the United States. ChatGPT’s global weekly active users were around 800 million in October 2025 and reported to be approaching 900 million.

Altman highlighted that India has the largest number of student users of ChatGPT worldwide. He noted that India’s focus on access, AI literacy and supporting infrastructure positions it to broaden who benefits from AI and to shape how democratic AI is adopted at scale. Government initiatives such as the IndiaAI Mission aim to expand computing capacity and support startups, while the upcoming India AI Impact Summit in New Delhi is expected to showcase international collaboration and commercial applications.

Several listed companies reported that AI-related products are driving revenue momentum. DigitalOcean, which focuses on small and mid‑sized businesses, said AI revenue had doubled year over year in each of the last five reported quarters. In the first three quarters of 2025, total revenue reached $659 million, up 14.5% year over year, an acceleration from 2024. Management guided that AI revenue likely doubled again in the fourth quarter and projected further growth when it reports results for the period ended Dec. 31, 2025, on Feb. 24.

Fortinet reported full‑year 2025 results that beat expectations and raised 2026 revenue guidance, while upgrading its FortiSIEM product. The company highlighted rapid annual recurring revenue growth in Secure Access Service Edge, security operations and cloud‑based services. However, its shares trade on a price‑to‑earnings multiple above the firm’s own fair‑value estimate, underscoring how investors are balancing AI‑driven growth against valuation and hardware‑cycle risk.

Rising AI workloads are also driving physical infrastructure expansion. In India, consultants expect a consolidation wave among data‑centre operators as demand from hyperscalers, AI workloads and enterprise cloud migration lifts utilisation in major hubs such as Mumbai and Chennai to 65–80%. Vacancy in key markets fell to 4.3% in the first half of 2025, described as a “sell‑out environment” for multi‑megawatt blocks.

Operators with scalable power access and AI‑ready facilities are seen as best placed, while sub‑scale and single‑market players may become acquisition targets. Analysts warned that rapid build‑outs can temporarily create overcapacity and low utilisation, leaving smaller firms exposed until returns improve. Larger telecom and technology companies are expected to use acquisitions to expand geographically and upgrade regional data‑centre assets from CPU‑ to GPU‑based services.

AI’s energy needs are prompting new power‑supply arrangements. In West Virginia, Mon Power and Potomac Edison proposed a 1,200‑MW natural‑gas plant adjacent to an existing coal station, alongside three new solar farms totalling 70 MW. The utilities told regulators that the projects are intended to address energy deficits and reduce potential capacity shortfalls while supporting reliable service.

Separately, Oklo announced an agreement with Meta Platforms to develop a 1.2‑GW nuclear power campus in Ohio, with Meta prepaying for power under a long‑term contract. Oklo is also advancing its Aurora small reactor project and a U.S. Department of Energy‑backed nuclear fuel recycling programme, and has partnered with Siemens Energy on power conversion systems. The Meta deal is viewed as a step toward supplying dedicated nuclear power for data‑centre and AI workloads, though Oklo remains pre‑revenue and reported a US$96 million net loss over the last year.