By Ji Zhang, portfolio manager for global real estate portfolios at Cohen & Steers
Artificial Intelligence is a transformative technology that we believe will be a net positive for real estate, though it has the potential to create both winners and losers across the investing landscape.
Over the past few decades, we have witnessed several transformative technology cycles: the proliferation of personal computing, the growth of the internet, the rise of social media, and the mobile revolution. Artificial intelligence (AI) appears to be the next technology with this transformative potential.
While AI has been progressing for decades, OpenAI’s release of ChatGPT in late 2022 brought language-capable artificial intelligence into the mainstream, catalysing a rapid acceleration in investment, with the potential for mass adoption of technology. We believe generative AI is likely to have wide-ranging applications across the economy, and this technology has the potential to transform sectors across markets, including real estate.
Though still early, we believe AI has strong potential to be a net positive for more than half the investable universe for REITs, improving operational efficiency for sectors such as healthcare and hotels, while leading to increased demand for data centres and towers. On the other hand, traditional office, which represents 3.1% of the investable universe for REITs, could see lower demand due to job displacement from AI.
AI revenue is expected to grow significantly over the next several years. Leading research firms estimate annual AI growth between 18% and nearly 30%. Bank of America estimates the growth will culminate in a total addressable AI market of $900 billion by 2026.
Cloud service providers at the leading edge of artificial intelligence have pointed to a growing revenue opportunity and corresponding acceleration in capital expenditures to capitalize on AI.
A large supplier of AI hardware and software reported it has seen a surge in AI-related demand, resulting in second quarter 2023 data centre revenues up 141% from the first quarter. Most of that growth is being led by accelerated computing requirements from cloud service providers, social media platforms and corporate enterprises across the automotive, financial services, health care and telecom industries.
One of the top-10 cloud providers has identified next-generation AI as a potential $10 billion business line and has signalled higher data centre investments to support AI adoption and continued cloud migration. To meet increasing AI revenue opportunities, Morgan Stanley expects capital expenditure growth among the largest cloud providers to accelerate from 7% in 2023 to 15% in 2024.
The business cases driving data centre demand are a combination of existing cloud workloads and novel AI applications.
While AI use cases driving data centre demand are intriguing, further understanding is needed for how consumers and enterprises are going to leverage AI. However, it is already clear that AI is changing how many companies operate, which in turn is increasing their need for computing power, cloud-based data and ultimately, data centre capacity.
A few notable AI examples:
- More than 27,000 organizations—including Airbnb, Dell and Scandinavian Airlines—are incorporating Microsoft Copilot, an AI platform, to increase the productivity of their software developers.
- Telecom providers are utilizing AI to improve fleet dispatch for field technicians.
- Pharmaceutical firms such as Amgen are using AI to help with drug discovery and protein engineering.
- Financial services firms, like J.P. Morgan, are adding AI to their IT workloads. Jamie Dimon, J.P. Morgan’s CEO, noted the firm already has 300 AI use cases and has already spent $2 billion on cloud migration but is still early in its digital transformation.
As AI’s use cases proliferate, many workloads will begin in the cloud. Annual cloud revenue contribution from AI and machine learning is expected to climb to $111 billion (representing 36% annualized growth) by 2030, according to estimates from Bank of America.
We believe there is a clear bull case for data centres as AI could drive incremental new demand, while existing data centres could benefit from non- AI applications. AI workloads require significantly more computational power than non-AI applications (as much as 5x by some estimates), which could absorb a significant amount of available power across major data centre markets. As a result, end users with traditional workloads might gravitate toward existing facilities, as enterprises will likely not require the same level of power density that new AI workloads will likely require. One risk for legacy data centres, however, is that older facilities could either face greater capital requirements to be upgraded or risk becoming obsolete.
Outsized cloud revenue growth and digital transformation have benefited data centre REITs so far, and we expect AI to be an additional secular tailwind.
AI’s impact beyond data centres
AI has the potential to disrupt numerous real estate sectors beyond just data centres, though we view much of the current investor discussion as speculative. Ultimately, we believe that discerning winners and losers will require careful analysis, and that capitalizing on opportunities will require active management.
We are currently considering both positive and negative scenarios across most real estate sectors. AI and automation are likely to improve margins for operationally intensive businesses such as senior housing and hotels. AI is also likely to improve the success rate of new drug discovery, leading to increased research funding and demand for life science property.
In addition, further investment in wireless networks will be needed as mass adoption broadens to mobile applications, which should be a positive for cell towers. Within office, job displacement as a result of AI could lead to lower demand for space, putting further pressure on occupancy.
