On July 30, 2026, Beijing time, Nscale and Anyscale announced that they had signed a definitive acquisition agreement. TechCrunch's public feed source published the same event at Coordinated Universal Time 15:19:15, describing the former as a UK artificial intelligence cloud infrastructure company and the latter as a software enterprise helping businesses scale AI workloads across data centers and servers. The agreement has not yet closed; both parties expect to complete it in the second half of 2026 after meeting closing conditions and obtaining regulatory approval; financial terms have not been disclosed.
Transaction Parties and Proposed Platform
Nscale's official announcement was released from London and San Francisco, positioning itself as an integrated software, computing power, data center, and electricity full-stack artificial intelligence cloud platform. Anyscale, founded by the creators of distributed computing framework Ray, provides a computing platform for AI workloads. The announcement confirms Josh Payne as CEO and founder of Nscale, Keerti Melkote as CEO of Anyscale; existing materials do not list post-closing management reporting relationships. Anyscale's official article is co-authored by five authors but does not label their post-closing management positions or role changes.
Based on the companies' descriptions of their current businesses, Nscale provides power, data centers, GPU computing power, and software for running AI workloads; Anyscale’s software layer targets machine learning engineers and AI platform teams. The proposed combination covers data processing, model training and fine-tuning, inference, reinforcement learning, and agent deployment. These projects are the companies' statements on product scope and reasons for the transaction; public materials do not provide third-party performance tests, cost comparisons, or completed integration results.
Brand Continuity and Team Transition
Nscale's announcement states that post-closing, Anyscale will continue to operate under its existing brand and serve current customers; customers can still choose the infrastructure for their AI workloads. As both parties integrate further, customers will also have the option to run Anyscale’s software layer on Nscale’s full-stack platform. Anyscale also notes that its platform post-closing will cover major cloud service providers, with portability continuing as part of Ray and Anyscale's product roadmap; these arrangements are commitments post-closing. Neither company has announced when new options will be available, their pricing, or customer migration plans.
Regarding personnel, Nscale states that the entire team distributed across the US, Europe, and India will join Nscale, totaling approximately 200 people. The announcement does not specify when these employees will complete the transition nor disclose job adjustments, management structures, office locations, compensation arrangements, retention conditions, or layoff plans. The figure of about 200 people comes from the acquirer's announcement and is not an employee count from regulatory filings; Anyscale’s official article also does not provide a more detailed regional distribution. Existing materials do not clarify whether employee transitions are subject to any conditions beyond transaction completion.
Ray Governance and Technical Rationale
The governance of the Ray project is separately retained in both companies' narratives about the transaction. Nscale's announcement states that Ray was donated to the PyTorch Foundation in 2025 and continues with open-source and community governance; Anyscale’s article also notes that the project, along with PyTorch and vLLM, falls under the foundation's governance. Nscale says it will join this foundation, while Anyscale writes its plan as joining at a platinum member level. Neither material states that membership procedures have been completed, nor does the transaction change Ray's current open governance status.
Anyscale attributes the rationale for the deal to increasing scale and complexity in AI systems, stating that software must consider rack and cluster topology, different hardware, computing power splits, and failures at each layer. The company believes joint optimization can address issues beyond single-layer optimization and cites Nscale’s power, data centers, and accelerated computing as reasons for choosing them. These are Anyscale's explanations of the technical path and motivation for cooperation, not measured conclusions on performance, reliability, or cost levels post-merger.
Growth Claims and Disclosure Boundaries
Anyscale also reports that its most recent quarter was its best yet, with revenue growing by over 70% year-over-year; the article does not disclose revenue amounts, comparison bases, accounting standards, or independent audit materials. The list of transaction advisors has been disclosed: Goldman Sachs International and Morgan Stanley provided financial advisory services to Nscale, Latham & Watkins served as its legal advisor; Qatalyst Partners and Fenwick & West provided financial and legal advisory services to Anyscale respectively.
This transaction is still between signing and closing, with public materials not disclosing the acquisition price, payment method, valuation, ownership structure, integration budget, board changes, customer migration arrangements, or employee retention terms. The content about full-stack platforms, more computing power, joint optimization, scale, and lower costs are company statements that cannot be taken as realized synergies. Verifiable points in the future include regulatory approval, actual closing time, foundation membership status, and post-closing product and organizational arrangements; however, current sources do not provide definite results.