From AI to Absorption: Office Demand, AI Talent Concentrations, and What it Means for Data Centers
John McWilliams • 1/18/2024
FULL REPORT NOW AVAILABLE
Artificial intelligence (AI) is an integral part of modern society and business, having re-entered the spotlight in late 2022 with the launch of OpenAI’s ChatGPT, as hype surrounding the technology boomed. As a result, AI occupiers are poised to become a significant tenant class in commercial real estate as AI technology continues to attract venture capital, reshape industries and further embed itself in the fabric of the business world.
Read Full Report
Upticks in office space demand for AI tenants have already been recorded and the San Francisco Bay Area, with more than 2.5 million square feet (msf) of tenants in the market, is an unsurprising hot spot for AI occupiers. Austin, New York City and Greater Seattle are also well-established markets for AI tenants looking to tap into highly skilled tech-talent pools that can support AI business expansion.
Demand for AI talent has been growing across the country and has not been exclusive to established AI markets. Many companies have increasingly sought to bolster their internal technology teams in the wake of the AI revolution which has led to a nationwide search for uniquely skilled talent and the growing presence of emerging AI markets. In fact, unique AI job postings peaked in June 2022 which saw 8,700 postings over the course of the month. Emerging AI markets are overwhelmingly located in Sunbelt states, have a high number of AI-related jobs that continues to grow, and have seen a significant increase in the number of AI-related degree completions.
The United States is a global leader in AI and currently hosts 65 of the top 100 globally ranked AI company headquarters, according to Pitchbook. Growth in the number of AI companies, the scale at which they operate and the increased prevalence of generative AI has pressured the data center industry to meet the new demand. Increased rack densities, continuously rising compute requirements and more efficient power usage effectiveness (PUE) ratios, are prominent trends for data center operators in the era of generative AI. The semiconductor industry, which supplies data centers with microchips, is facing material shortages and production slowdowns due to labor shortages in a time of heightened demand and increased buy-side competition. However, AI occupiers and data center operators can both benefit from higher latency locations for large language model (LLM) training, spurring interest in low-cost markets.
Going forward, data center users and operators are expected to continue seeking out opportunities in states that can meet their sustainability goals, the most prominent of which is renewable power. The ideal future location for a data center operation would have plenty of available renewable power, a low risk of natural disaster, ample water availability, robust waste management infrastructure and a business-friendly disposition.
Read the full report here.
Demand for AI talent has been growing across the country and has not been exclusive to established AI markets. Many companies have increasingly sought to bolster their internal technology teams in the wake of the AI revolution which has led to a nationwide search for uniquely skilled talent and the growing presence of emerging AI markets. In fact, unique AI job postings peaked in June 2022 which saw 8,700 postings over the course of the month. Emerging AI markets are overwhelmingly located in Sunbelt states, have a high number of AI-related jobs that continues to grow, and have seen a significant increase in the number of AI-related degree completions.
The United States is a global leader in AI and currently hosts 65 of the top 100 globally ranked AI company headquarters, according to Pitchbook. Growth in the number of AI companies, the scale at which they operate and the increased prevalence of generative AI has pressured the data center industry to meet the new demand. Increased rack densities, continuously rising compute requirements and more efficient power usage effectiveness (PUE) ratios, are prominent trends for data center operators in the era of generative AI. The semiconductor industry, which supplies data centers with microchips, is facing material shortages and production slowdowns due to labor shortages in a time of heightened demand and increased buy-side competition. However, AI occupiers and data center operators can both benefit from higher latency locations for large language model (LLM) training, spurring interest in low-cost markets.
Going forward, data center users and operators are expected to continue seeking out opportunities in states that can meet their sustainability goals, the most prominent of which is renewable power. The ideal future location for a data center operation would have plenty of available renewable power, a low risk of natural disaster, ample water availability, robust waste management infrastructure and a business-friendly disposition.
Read the full report here.
Want more information about data centers?
Subscribe to stay updated about what's next for the data center sector of commercial real estate. We also deliver full services from site selection, appraisals, and facilities management to sales and leasing, using deep data center expertise.
Subscribe
Learn more Related Insights
Podcast • Economy
Cushman & Wakefield: Behind the Numbers
Our experts cut through the noise to provide you concise commentary on only the most important economic and sector-specific commercial real estate data points.
11/6/2024
Research
2024 Global Data Center Market Comparison
Explore a unique way to assess and score 92 global primary and emerging data center markets utilizing 14 criteria.
Jacob Albers • 3/26/2024
Research • Economy
10 Critical Questions for 2024
It was an uphill battle for CRE in 2023. Higher interest rates, elevated inflation and the threat of a recession all contributed to a year of turbulence and uncertainty. Despite the challenges, the industry demonstrated bright spots of resilience and adaptability. What does that mean for 2024?
Kevin Thorpe • 1/2/2024