Hyperscale colocation leasing starts at 4 MW per deployment, typically carries a 10 to 15 year commitment, and represents the lowest-priced segment per kilowatt because buyers consolidate large blocks of capacity [S2].
The wider 2026 market can be bounded, not reduced to one leasing figure: global data center capacity is projected to grow 97 GW from 2025 through 2030, with almost 100 GW of new supply and a 14% compound annual growth rate across the period [S1][S3].
2026 Leasing Volume: What the Available Data Establishes
No verified public source in the research material gives a single 2026 hyperscale colocation leasing volume in megawatts. JLL instead forecasts that nearly 100 GW of new data center capacity will be added between 2026 and 2030 and that global capacity will double over that period [S1][S3].
JLL also states that hyperscalers will drive expansion through a dual strategy of leasing capacity and building their own facilities [S1]. That dual structure is essential: the 97 GW capacity increase covers owned, leased and other capacity additions, so it cannot be presented as hyperscale colocation leasing volume. A defensible 2026 answer therefore gives the contract threshold and market envelope, not an invented MW total.
The 2026 demand context is measurable. Global capacity could reach 200 GW by 2030, while the Americas account for about 50% of global capacity and the United States represents about 90% of the regional total [S1]. APAC is forecast to expand from 32 GW to 57 GW, a 12% CAGR through 2030 [S1].
Hyperscale, Wholesale and Retail Capacity Bands
Hyperscale colocation is defined at 4 MW or more, wholesale spans 250 kW to 4 MW, and retail covers installations up to 250 kW [S2]. These bands determine the unit economics and contract structure more clearly than a headline market forecast.
Retail deployments generally use 1 to 3 year leases, wholesale commitments usually run 3 to 7 years, and hyperscale requirements often extend 10 to 15 years with pre-negotiated expansion rights [S2]. Hyperscale buyers therefore trade flexibility for scale, lower unit pricing and greater contractual control over future phases.
Pricing is normally quoted in dollars per kilowatt per month under a modified gross lease. The base rate covers power distribution, cooling, physical security, redundancy and building management, while electricity is commonly passed through separately [S2]. Annual escalators typically range from 2.5% to 5% [S2].
Decision Criteria for Large Capacity Leasing
Power availability is the first gate for a large colocation requirement because the average grid-connection wait in primary data center markets exceeds four years [S1]. JLL identifies power, rather than location or cost, as the primary site-selection criterion and notes that global occupancy reached 97% [S3].
Contract duration and pricing are the second decision set. Hyperscale leases offer the lowest per-unit rate, but their 10 to 15 year term and potential annual escalator of 2.5% to 5% reduce near-term flexibility [S2]. Operators with greater negotiating leverage can also use expansion rights to stage deployments, while tenants retain the option to lock capacity before it is needed [S3].
Delivery and technical design form the third set. More than half of projects faced delays in 2025, so contract dates require schedule buffers [S3]. AI-oriented facilities may also require rack densities approaching 100 kW and liquid-cooling capability [S3].
Electricity pricing must be evaluated separately from the lease rate. A small movement in utility rates applied across several hundred megawatts can change annual cost by millions of dollars [S2]. Capacity planning should therefore include both the base infrastructure charge and the power pass-through exposure.
Criteria-Based Comparison of Colocation Options
Retail colocation fits deployments up to 250 kW with the highest per-unit pricing and the shortest typical commitment, while wholesale capacity of 250 kW to 4 MW occupies the middle of both pricing and term structure [S2]. Hyperscale starts at 4 MW and offers the lowest per-unit price through scale, but the 10 to 15 year term creates the greatest contractual lock-in [S2].
The relevant comparison is therefore not price alone. For a 4 MW threshold, the hyperscale segment is the only listed band that meets the requirement without splitting the deployment across retail or wholesale contracts. The trade-off is price per kilowatt against duration, expansion flexibility and power-contract exposure.
Occupancy reinforces the distinction. At 97% global occupancy, landlords retain negotiating leverage, and significant rent growth is anticipated through 2030 [S3]. North American average vacancy is below 2%, with many tier-one markets below 1%; European vacancy is just under 4% [S2].
Who Uses Hyperscale Colocation and When
Hyperscale colocation is suited to cloud operators, AI infrastructure operators and large enterprises that need at least 4 MW but prefer operating space from a colocation provider rather than building a complete facility. These buyers can use long leases, phased capacity and pre-negotiated expansion rights to match deployment plans [S1][S2].
The model is not a good fit for small deployments. A customer below 250 kW falls into retail colocation, where pricing is higher but the typical 1 to 3 year term provides materially greater flexibility [S2]. It is also a weak substitute for tenants that require a dedicated facility, unusual power architecture or a short deployment window that cannot tolerate construction delays.
Hybrid portfolios are becoming the default corporate model. On-premises footprints are expected to shrink, but sensitive workloads remain on-site as enterprises combine owned infrastructure, colocation, hyperscale capacity and edge sites [S3].
Power, Delivery and Contract Constraints
Grid access is the binding physical constraint in the 2026 leasing market. Connection waits exceeding four years, 97% occupancy and project delays affecting more than half of 2025 projects make delivery timing a contractual issue rather than an operational detail [S1][S3]. Sites without secured power can become stranded despite zoning approval and demonstrated tenant demand [S3].
Equipment and construction lead times also affect developability. Transformers, generators and UPS systems face extended delivery periods, while labor scarcity and financing costs feed into the lease rate [S2]. Tenants should include schedule buffers and should not assume that a contractual delivery date is guaranteed without qualified suppliers and execution capacity [S3].
Technical constraints are converging with AI deployment. Rack densities approaching 100 kW require liquid-cooling readiness, and even non-AI tenants are advised to preserve pathways for higher densities and cooling upgrades [S3]. A low base rate does not remove the need to price power, cooling, expansion and long-term escalation together.
Sourcing and Market Tracking
The most defensible public figures for tracking 2026 hyperscale colocation leasing are the 4 MW hyperscale threshold, 10 to 15 year contract range, 2.5% to 5% annual escalators, 97% global occupancy and more than four years for average grid-connection waits [S1][S2][S3].
Any claimed 2026 leasing total should be expressed with a defined perimeter: contracted MW only, signed versus operational capacity, and the period covered. The research does not supply those values, so the 97 GW 2025 to 2030 capacity increase must not be presented as a hyperscale leasing subtotal [S1][S3].
The next trackable signals are regional additions against the forecast 13 GW of new EMEA supply, APAC movement from 32 GW to 57 GW by 2030, and whether announced hyperscale leases progress from pre-commitment to energization [S1]. The data center busway market analysis is relevant to the electrical infrastructure needed to deliver those capacity phases.