Data Centre Stocks Slip as AI Infrastructure Realities Bite
Investor confidence in AI data centres wavers as projects face community pushback, soaring costs, and supply chain hurdles.
For twenty years, data facilities functioned as secure, highly profitable real estate investments. Functioning as the fundamental digital plumbing of the internet, they regularly attracted conservative capital. However, the heavy computational demands of artificial intelligence have triggered an intensive race to construct fresh capacity. Recently, financial markets have begun questioning the long-term sustainability of these operations.
Market sentiment shifted sharply when NextDC, a major Australian data facility developer, saw its stock price drop by 40% following revelations that delays at its flagship Melbourne location would cause revenue targets to be missed. The setbacks occurred after local resistance prompted utility distributor Jemena to halt the installation of twenty-two-metre concrete power poles along residential streets in Footscray. Because artificial intelligence hardware requires massive electrical supply to remain functional, power access is paramount.
Financial struggles extend beyond Australia. Firmus abandoned a multibillion-dollar public listing on the local exchange due to weak market demand, while London-based Nscale has faced difficulties finding buyers for a Wall Street debut. Additionally, CoreWeave experienced a 15% single-day share price decline amid questions over its financial footing. These fluctuations highlight mounting scrutiny regarding underlying artificial intelligence business models.
Escalating Costs and Short Hardware Lifespans
Global spending on model training and facility construction has reached staggering figures, with total investments projected to hit US$10 trillion by the year 2032. While early capital came from internal corporate reserves or direct investor funding, developers increasingly rely on corporate bond markets. Meanwhile, persistently high inflation and elevated interest rates have driven up borrowing and construction expenses.
Analysts note that leading developers like OpenAI and Anthropic face expectations of 84% annual revenue growth over six years to satisfy investors, even as consumer and corporate spending growth on paid tools shows signs of deceleration. Furthermore, unlike traditional facilities that lasted for decades with minimal maintenance, modern AI chips feature extremely short lifespans. Each hardware generation demands specialized cooling and power upgrades, locking operators into perpetual capital expenditure cycles.
Supply anxieties have grown so intense that Amazon recently admitted to hoarding billions of AI chips despite lacking the physical capacity to install them. This dynamic has largely benefited hardware manufacturers like Nvidia, though circular financing practices—where Nvidia invests in its own chip-buying customers—have drawn additional oversight.
Community Friction and Resource Strain
Beyond corporate balance sheets, social licence is eroding due to local pushback against noise, visual disruption, thermal output, and environmental footprints. Public resistance threatens to intensify if facility operations drive up consumer utility costs. Projections indicate that facilities in New South Wales might consume up to 25% of Sydney's daily water supply within a decade, while Victorian transmission requests exceed the state's entire electrical supply capacity.
What it means for developers
Developers navigating this shifting landscape must contend with tighter capital markets, expensive infrastructure upgrades, and strict utility constraints. Efficiency has become paramount, requiring a shift toward leaner architectures that minimize resource footprints and energy consumption. As the industry matures, developers can try top AI models cheaply through one API at https://apixoai.online to build and test applications more cost-effectively without committing massive upfront resources to individual proprietary systems.
Source: Shares in this data centre company are down 40% as investors lose faith in the AI story — The Conversation AI. Written by the Apixo team from that report.
One key for Claude, GPT, GLM, DeepSeek and more. Pay per token with crypto.
Get your API key
