Blackstone Buys 75% of AWS Data Center JVs for $5B in AI REIT Deal
Blackstone Digital Infrastructure Trust has signed a definitive agreement to acquire a 75% stake in two AWS data center joint ventures for approximately $5 billion at a 7.0% cap rate, marking a landmark securitization of AI infrastructure into a public REIT. The portfolio includes 13 newly built powered shells with 647MW of IT capacity, with AWS retaining a 25% stake and leasing back the entire portfolio under a 12-year initial triple-net lease extendable to 32 years. The transaction comes as SemiAnalysis slashed its 2028 and 2030 AI accelerator capex forecasts, highlighting shifting dynamics in AI infrastructure investment.
Table of contents
Blackstone’s $5B AWS Data Center Deal: What the Announcement Says
Blackstone Digital Infrastructure Trust (NYSE: BXDC) signed a definitive agreement on November 16, 2026, to acquire a 75% stake in two Amazon Web Services data center joint ventures for approximately $5 billion at a 7.0% cap rate, marking one of the most high-profile securitizations of AI-focused data center infrastructure to date. The portfolio includes 13 newly built powered shells with roughly 647 megawatts of IT capacity, positioned to support the fast-growing demand for AI compute workloads.
Under the terms of the deal, AWS will retain a 25% ownership stake in the joint ventures and lease back the entire portfolio under a triple-net lease structure. The initial lease term runs for 12 years, with options to extend up to a fully extended 32-year period. The transaction is expected to close in the fourth quarter of 2026, pending customary regulatory approvals and closing conditions.
The announcement comes against a turbulent backdrop for AI infrastructure investment, as research firm SemiAnalysis recently slashed its 2028 AI accelerator capex forecast by 28% and cut its 2030 projection from $1.531 trillion to $923 billion. The contrast between the large-scale REIT acquisition and the reduced capex outlook highlights a shifting market dynamic, as institutional capital moves to consolidate existing AI infrastructure while near-term spending growth cools.

Why This Deal Is a Marker for AI Infrastructure as an Asset Class
The Blackstone-AWS transaction is being framed as a marquee example of AI infrastructure being securitized into a public REIT, a shift that could reshape how institutional and retail investors gain exposure to the AI boom. Unlike investments in chipmakers or hyperscaler stocks, which are tied to product cycles and revenue growth, REIT-held data centers offer steady, long-term income streams backed by contractual leases.
The 7.0% cap rate on the deal reflects the market’s pricing of AI data center assets as a relatively stable income play, supported by AWS’s strong credit rating and the long duration of the lease. Triple-net lease terms mean AWS is responsible for most operating costs, including taxes, insurance, and maintenance, reducing the risk for the REIT and its investors.
For the broader AI ecosystem, this kind of securitization can unlock new pools of capital for data center buildouts, as REITs can raise funds from public markets to finance new construction. It also signals that AI infrastructure is maturing from a speculative growth sector to a more established, income-generating asset class.

Context: Slashed Capex Forecasts and a Cooling AI Investment Climate
The deal lands at a time when analysts are rapidly revising down their expectations for near-term AI infrastructure spending. SemiAnalysis’s latest cuts reduce the 2028 AI accelerator capex forecast by 28%, and slash the 2030 forecast by more than $600 billion, from $1.531 trillion to $923 billion. The revisions reflect growing expectations that hyperscalers will slow their pace of new data center buildouts as they assess actual demand for AI workloads, navigate macroeconomic headwinds, and adjust to tighter financing conditions.
At first glance, the $5 billion Blackstone-AWS deal might seem to contradict the bearish capex outlook, but the two trends are actually complementary. The deal involves existing, fully built data center assets with a long-term lease from a top-tier tenant, not speculative new construction. This suggests that while investors are growing more cautious about funding new, unleased data center capacity, they are still willing to pay a premium for high-quality, contracted AI infrastructure assets with steady cash flow.
The market split — strong demand for existing assets, weaker appetite for new buildouts — points to a maturing AI infrastructure market, where growth is shifting from explosive capacity expansion to more measured, demand-driven investment.

What This Means for AI Content Creators
For creators who rely on an ai video generator or ai image generator for daily content production, shifts in AI infrastructure investment hit closer to home than many realize. Every cloud-based AI tool runs on data center capacity like the portfolio Blackstone is acquiring, so consolidation, pricing pressure, and capacity shifts can translate to changes in tool availability, speed, or cost over time.
On the positive side, the securitization of AI infrastructure into long-term REIT holdings can bring more stability to the tooling ecosystem. When data center assets are backed by long-term leases and predictable capital, tool providers have more visibility into their infrastructure costs and capacity, which can lead to more consistent service and pricing for creators.
On the other hand, slower capex growth could mean slower rollouts of next-generation models, as providers face tighter constraints on new capacity to support more powerful, compute-heavy AI tools. Creators may also see pricing shifts as providers pass through changes in infrastructure costs.
That’s why many creators are building more resilient workflows by using multi-model platforms that give them access to a range of AI tools across different providers. By spreading their work across multiple tools and infrastructure stacks, creators reduce their risk of disruption from any single provider’s outages, price hikes, or capacity limits. MagicBoat brings together leading AI creative tools in one unified platform, so creators can switch between models, experiment with new tools, and build flexible workflows that adapt to changing market conditions. If you want to build resilient creative workflows across multiple leading AI tools, try these models for free.
Key Takeaways
- Blackstone Digital Infrastructure Trust is acquiring 75% of two AWS data center joint ventures for ~$5 billion at a 7.0% cap rate, with AWS retaining 25% and leasing back the full portfolio.
- The portfolio includes 13 newly built powered shells with ~647MW of IT capacity, under a 12-year initial triple-net lease extendable to 32 years, with closing expected in Q4 2026.
- The deal is a landmark securitization of AI infrastructure into a public REIT, marking the maturing of AI data centers as a stable income-generating asset class.
- The transaction comes as SemiAnalysis slashed its 2028 AI accelerator capex forecast by 28% and its 2030 forecast from $1.531T to $923B, reflecting cooling near-term spending growth.
- For creators, infrastructure shifts impact tool reliability and cost, making multi-model platforms a valuable strategy for building resilient creative workflows.
Keywords
blackstone aws data center deal, ai infrastructure reit, ai capex forecast, data center investment, ai compute infrastructure, aws data center joint venture, bxdc data center acquisition