Blackstone, Google Launch $5B AI Cloud Venture

Bisnow reports that Blackstone and Google are teaming up to launch a new AI cloud provider aimed at the booming market for high-performance computing infrastructure. The still-unnamed venture will provide customers with on-demand access to Google’s Tensor Processing Units (TPUs), the company’s proprietary AI chips, according to The Wall Street Journal.

The partnership launches with $5B in equity capital from Blackstone, which will hold a majority ownership stake. The firms expect that initial investment to support roughly $25B in computing infrastructure deployment as demand for AI workloads continues to accelerate.

A Push Into the Neocloud Market

The venture marks a major expansion into the so-called “neocloud” sector, a fast-growing corner of the cloud industry focused on AI-native computing services. Unlike traditional cloud providers, neocloud operators specialize in delivering access to high-performance chips and AI infrastructure with faster deployment timelines and more flexible pricing.

That segment has largely been led by startups including CoreWeave, Crusoe, Lambda, Nebius, Nscale, and Core Scientific. According to JLL, more than 190 neocloud operators are now competing globally as enterprises and AI developers scramble for computing power.

Google and Blackstone are entering the market with considerable scale advantages. Longtime Google executive Benjamin Treynor Sloss will serve as CEO of the new company, which plans to deploy 500 MW of AI capacity by the end of 2027. Blackstone said identified data center sites are already in place for the initial buildout.

The Details Behind the AI Infrastructure Buildout

The new platform will rely on Google’s TPUs rather than Nvidia GPUs, the dominant processors currently powering most AI workloads. Google historically used TPUs internally across its own data centers but has increasingly commercialized the chips over the past year.

In late 2025, Google expanded TPU access through major AI computing agreements with Anthropic and Meta. Company executives also signaled during the firm’s latest earnings call that TPU commercialization would become a larger strategic priority, particularly for AI inference workloads, which require massive amounts of computing capacity at lower operating costs. The expansion also aligns with Google’s broader push to redirect capital toward AI infrastructure after cutting back on office space commitments across several major markets.

For Blackstone, the partnership further expands one of the largest digital infrastructure portfolios in commercial real estate. The investment giant built much of its data center footprint through its acquisitions of QTS in 2021 and AirTrunk in 2024, giving it substantial hyperscale and colocation capacity across North America and Asia-Pacific markets.

Institutional Capital Floods Into AI Data Centers

The deal underscores how aggressively institutional investors are scaling exposure to AI infrastructure. Blackstone recently consolidated many of its AI and digital infrastructure investments under a new platform called Blackstone N1, launched in April 2026.

Earlier this month, Blackstone N1 also formed a $1.5B joint venture with Anthropic, adding to a growing list of institutional partnerships tied to AI computing demand. Other major investors including Brookfield, KKR, DigitalBridge, and Blue Owl have also expanded allocations toward powered land, data centers, and AI cloud infrastructure over the past year.

The broader market opportunity remains enormous. Global data center demand tied to AI workloads is expected to more than triple by 2030, according to McKinsey’s 2025 AI infrastructure outlook, with power availability becoming one of the industry’s biggest bottlenecks.

Why It Matters

The partnership highlights the convergence of cloud computing, AI chip development, and commercial real estate infrastructure. Rather than relying solely on third-party cloud operators, Google is pairing proprietary silicon with Blackstone’s real estate and capital platform to create vertically integrated AI infrastructure at scale.

That strategy could reshape competitive dynamics in the cloud market, particularly as enterprises seek alternatives to Nvidia-heavy ecosystems. It also reinforces how data center development has become one of the most capital-intensive and strategically important sectors in CRE.

What’s Next

The venture’s initial 500 MW deployment target is likely just the beginning. Blackstone said computing capacity will continue scaling after 2027 as additional sites come online and AI demand grows.

Market watchers will also be tracking whether other hyperscalers pursue similar partnerships pairing proprietary chips with institutional infrastructure capital. As AI workloads expand, access to power, land, and deployable data center capacity may become as critical as the chips themselves.