In an analysis published in May 2026, New York City Comptroller Mark Levine highlighted the metropolis's dual position, noting that no other city is as exposed to the opportunities and risks of artificial intelligence as New York City. Because municipal tax revenue is closely tied to an increasingly AI-driven financial industry, Levine's office outlined five economic future scenarios. Accordingly, the probability of a productivity surge with limited economic disruption stands at 35 per cent, followed by a 25 per cent probability of an investment slowdown and declining markets. The report put the likelihood of widespread job losses at 20 per cent, whilst an unreservedly positive growth scenario was estimated at just 15 per cent.
These model projections coincide with a volatile global environment. Goldman Sachs forecast global AI investment of one trillion dollars for 2026, including 581 billion dollars in the United States, but also pointed to a net monthly loss of 16,000 jobs over the past year. The International Monetary Fund recently warned of sudden market disruptions resulting from potential AI corrections and rising interest costs.
Significant increase in Manhattan leasing activity
Regardless of macroeconomic warnings, New York's office property sector is recording strong demand from technology firms. According to surveys by property services firm Savills, AI companies leased around 1.7 million square feet of office space in Manhattan during the first half of 2026. This represents an 8.2 per cent share of the borough's total new leasing volume, a significant jump compared to 0.7 per cent across the full year 2020. The first half of 2026 thus already exceeded the leasing volume of 2024 and 2025 combined.
At the same time, a qualitative maturation of demand is becoming apparent. The share of subleases in the transaction volume of AI firms fell from 42.3 per cent in the period from 2020 to 2024 to 17.5 per cent between 2025 and mid-2026, pointing to a growing preference for direct leases. Furthermore, the average lease term lengthened from 49 months in 2020 to 90.9 months in the first half of 2026. Gabe Marans, Vice Chairman at Savills, highlighted that New York, alongside the Bay Area, has emerged as the leading hub for sector-specific applied AI solutions.
Low market share limits systemic risks
Although the growth rates appear substantial, a look at the overall volume puts the risks for the wider market into perspective. According to Savills, despite the increases, AI companies currently occupy only around one per cent of total office stock in Manhattan, which corresponds to roughly 4.2 million square feet in absolute terms. Marans explained that even an abrupt bursting of the AI bubble would hardly derail the New York office market, given that the sector's weighting within the overall market remains modest.
Nevertheless, a potential cooling poses noticeable risks for individual market participants. Jeff Gural, Chairman and Principal of GFP Real Estate, emphasised that the greatest risk for owners of Class B buildings currently lies in new lettings to AI tenants. While owners were still able to deliberately avoid such commitments during the dot-com boom, there is now little alternative in the current market environment to leasing space to this growing sector.
