Google, Blackstone Launch $25 Billion AI Cloud Joint Venture
Google and Blackstone (NYSE: BX) announced Monday May 18 a US-based joint venture to create a new artificial intelligence cloud company offering Google’s Tensor Processing Units (TPUs) as compute-as-a-service. Blackstone is making an initial equity commitment of $5 billion. Total investment, including leverage, could reach $25 billion, according to Bloomberg reporting. The venture targets 500 megawatts of computing capacity online by 2027.
Under the partnership terms, Blackstone will be the majority owner of the venture, which will offer data-center capacity, operations, networking, and TPU access. Google will supply the TPU hardware, software, and services. The structure allows Google to extend the commercial reach of its custom AI chips beyond Google Cloud, while Blackstone packages capital, property, and operating expertise into a platform for enterprises that want high-performance AI infrastructure without building it themselves.
Benjamin Treynor Sloss, a longtime Google executive, will serve as chief executive of the new venture. “This joint venture with Blackstone helps meet growing demand for TPUs,” Google Cloud CEO Thomas Kurian said. Jon Gray, Blackstone’s President and COO, described the move as “a generational opportunity to invest in AI infrastructure.” Blackstone manages $1.3 trillion in total assets and runs more data center capacity than any other private investor globally.
The deal arrives amid an unprecedented AI infrastructure capital cycle. As the Rio Times reported on the NextEra-Dominion $67 billion merger announced the same day, US utility consolidation is being driven by exactly this AI power demand. Combined Big Tech AI infrastructure spending is projected to top $700 billion in 2026 — up from earlier $600 billion forecasts — as Alphabet, Amazon, Microsoft, and Meta all signal acceleration rather than deceleration in capex.
Key Points
The Deal
| Element | Detail |
|---|---|
| Initial Equity Commitment | $5 billion (Blackstone) |
| Total Investment (with leverage) | Up to $25 billion |
| Capacity Target (2027) | 500 megawatts |
| Structure | Blackstone majority owner; Google supplies TPUs |
| CEO | Benjamin Treynor Sloss (ex-Google) |
| Blackstone AUM | $1.3 trillion |
| Alphabet Q1 Cloud Revenue | $20 billion (+63%) |
| 2026 Big Tech AI Capex | $700+ billion projected |
Why It Matters
The structural innovation is splitting Google’s TPU commercial channel from Google Cloud itself. By spinning out a separate TPU-cloud company, Google can pursue large infrastructure contracts without warping Google Cloud’s reported margins. This mirrors CoreWeave (which IPO’d in 2025 in the biggest US tech IPO since 2021) and the OpenAI-backed Stargate venture — Google is borrowing the model but substituting its own custom silicon for Nvidia GPUs.
For Blackstone, the deal extends its dominance in digital infrastructure. The Blackstone Digital Infrastructure Trust raised $1.75 billion in its US IPO on May 13, 2026 (87.5 million shares at $20), and the firm recently created a dedicated AI infrastructure arm called BXN1. Blackstone now controls more data center capacity than any other private investor globally — a position that became strategic as AI workloads moved capacity demand from commercial real estate to power-intensive computing.
Alphabet’s Q1 2026 results released last month support the deal’s strategic logic. Q1 consolidated revenue reached $109.9 billion (+22 percent year-on-year); Google Cloud revenue specifically grew 63 percent to approximately $20 billion. TPU-based workloads have grown to power “many of the world’s top AI labs, capital market firms, and companies running the most complex high-performance computing applications,” per Blackstone’s announcement. Spinning out additional TPU supply removes a constraint on Google’s ability to satisfy that demand.
For Brazilian investors, the deal provides a benchmark. As the Rio Times reported on CPFL Energia’s Q1 2026 print, Brazilian utilities are positioning for the same AI-driven power demand cycle. The Brazilian data-center market is projected to triple by 2030, with major projects emerging across São Paulo, Ceará, and Pernambuco states. The Google-Blackstone playbook — combining hyperscaler chip access with private-equity infrastructure capital — could be replicated in Brazil.
Combined with the simultaneous NextEra-Dominion $67 billion utility merger announcement, May 18, 2026 stands as a turning point in AI infrastructure capital structure. Big Tech is no longer building everything in-house — power generation, data-center construction, and chip distribution are increasingly being syndicated to specialised partners. The model accelerates capacity expansion while preserving Big Tech balance-sheet flexibility. Expect more announcements of this structure through 2026-2027.
AI capex $700B+ projected. Demand exceeds even hyperscalers’ build capacity.
TPU as alternative to Nvidia. Custom-silicon thesis spreads beyond Google Cloud.
Blackstone scale unmatched. Most data center capacity of any private investor.
$700B capex bubble risk. AI demand could decelerate; ROI uncertain.
500 MW = small initial scale. Execution complexity at hyperscale.
CoreWeave competition. Crowded AI cloud market by 2027.
Frequently Asked Questions
What is the Google-Blackstone deal?
Google and Blackstone announced Monday May 18, 2026 a US-based joint venture to create a new artificial intelligence cloud company. Blackstone is making an initial $5 billion equity commitment, with total investment potentially reaching $25 billion including leverage per Bloomberg reporting. The company will offer Google’s Tensor Processing Units (TPUs) as compute-as-a-service. Target is 500 megawatts of computing capacity online by 2027. Benjamin Treynor Sloss, a longtime Google executive, will serve as CEO.
What are TPUs?
TPUs (Tensor Processing Units) are Google’s custom-designed AI chips, purpose-built for training and inference of advanced AI models. They have been developed and deployed in Google production for more than a decade and currently power workloads for many of the world’s top AI labs, capital market firms, and high-performance computing applications. TPUs compete with Nvidia GPUs in the AI accelerator market. The joint venture will offer TPUs as an alternative cloud-computing option beyond traditional Google Cloud access.
Why is AI infrastructure capex so large?
Combined Big Tech AI infrastructure spending is projected to top $700 billion in 2026 — up from earlier $600 billion forecasts. Alphabet, Amazon, Microsoft, and Meta all signaled during recent earnings calls that AI spending would not slow.
Alphabet’s Q1 2026 results showed cloud revenue growing 63% to approximately $20 billion. The capex cycle reflects sustained corporate demand for AI training and inference capacity from hyperscalers, enterprises, and AI startups — a structural shift that has pushed even the largest companies to syndicate infrastructure development with private-equity partners like Blackstone
Updated: 2026-05-18T19:00:00-03:00 by Rio Times Editorial Desk
Google Blackstone $25B AI cloud | TPU joint venture | Benjamin Treynor Sloss | Jon Gray | AI infrastructure 2026 | The Rio Times
This article was produced by The Rio Times’ automated newsroom system. How we use AI · Report an error
Read More from The Rio Times