AI Infrastructure

Los Angeles – September 25, 2026 -- MeetKai is rolling out a sovereign AI stack built on NVIDIA accelerated computing across six countries -- Brazil, Ukraine, Pakistan, Kazakhstan, Uzbekistan and Bangladesh -- home to more than 700 million people, following a public-tender win in Brazil and plans for the first national-scale AI factory in Ukraine.

SERPRO awards MeetKai Brasil the top score in Brazil's national sovereign AI tender

MeetKai Brasil won the public tender run by SERPRO, Brazil's federal government IT company, to deploy a sovereign AI platform and locally trained models on NVIDIA accelerated computing inside the country. The deployment includes a sovereign coding agent for secure workflows and targets regulated sectors -- banks, insurers, energy companies and state-owned enterprises -- whose regulators require workloads to remain under Brazilian control.

Singapore – September 25, 2026 -- Brightray Veridian Pte. Ltd., a wholly owned subsidiary of Brightray Science Inc., has signed a Prefabricated Module Sales Framework Agreement with a Macao-based developer for a planned 40MW artificial intelligence data center in Coloane, Macao, with delivery targeted for the third quarter of 2027.

Framework agreement covers 40MW AI data center at Coloane site

New York – September 25, 2026 -- Sixty-nine percent of enterprise technology decision-makers now weigh energy more heavily in AI and data infrastructure decisions than a year ago, according to new research from EnterpriseDB (EDB) presented at the United Nations Global Compact Leaders Summit this week. The survey of 300 respondents, conducted in September 2026, found reliability of supply cited by 49% of respondents as a leading energy factor, with cost and energy efficiency each cited by 48%.

London – – September 25, 2026 -- Nscale Limited has closed a $3.36 billion convertible loan note financing led by Third Point, with NVIDIA committing an additional $1 billion tranche expected to fund in mid-November 2026.

Nscale secures $2.36 billion at initial closing, with NVIDIA's $1 billion tranche to follow

The financing structure splits into two tranches: a $2.36 billion initial close and a subsequent $1 billion commitment from NVIDIA. Goldman Sachs & Co. LLC acted as placement agent for the round.

Las Vegas – September 25, 2026 -- Ault & Company, Inc. has invested $55 million directly into Hyperscale Data, Inc. through convertible preferred stock purchases, building a beneficial ownership stake of approximately 62% in the NYSE American-listed company.

Ault & Company files 13D amendment disclosing 62% stake in Hyperscale Data

The $55 million figure excludes additional funds spent by Ault & Company and its affiliates on Hyperscale Data common stock through open market purchases. Details of the ownership position appear in Schedule 13D/A Amendment No. 19, filed with the U.S. Securities and Exchange Commission on September 15, 2026.

Austin, Texas – September 24, 2026 -- Forge AI Deployment has committed $30 million to webAI to build and deploy private, specialized AI systems for enterprise customers, the companies announced.

Forge pairs $30 million commitment with webAI's collaborative intelligence platform

The deal combines webAI's technology with Forge's enterprise deployment and services operation. Forge will work directly with customers to build AI systems around proprietary data, workflows and operations, running on infrastructure those customers control, from individual specialized AI Personas to company-wide deployments.

Los Angeles – September 24, 2026 -- Stand8, a global technology consulting firm, has launched the Stand8 Model Gateway, a new enterprise AI service designed to reduce model spending and compress the path from proof of concept to production AI deployment into a 90-day sprint.

Gateway routes each AI task to the most cost-effective model without touching client data

The gateway is deployed inside the client's own cloud account on AWS, Microsoft Azure, or Google Cloud, sitting between applications and the models they call. Stand8 designs and tunes the routing logic but does not sit in the data path itself. The system evaluates each task and routes it to the most appropriate option among leading closed models, open-source and open-weight models, or smaller fine-tuned models, while layering in caching, identity, policy, and audit controls.

Wilmington, Del. – September 24, 2026 -- Synteq has launched its largest single-site expansion to date, deploying a full-scale cloud and GPU compute facility in Chicago that extends the company's infrastructure into the Northern US Central region.

Chicago site runs Synteq's newest core platform end to end

The Chicago deployment combines high-density GPU compute for AI and machine learning workloads with virtualized cloud, bare metal servers, and block storage, all operated on Synteq's own network backbone, AS400304, and managed entirely by the Synteq team.

Menlo Park, Calif. – September 24, 2026 -- PhoenixAI has become a foundational database partner for Cloudera, making its cloud database engine available on the newly launched Cloudera Anywhere Cloud Marketplace to deliver sub-second real-time analytics for agentic AI workloads.

PhoenixAI database engine enters Cloudera's private and sovereign cloud ecosystem

Enterprise customers can now deploy PhoenixAI's database engine directly within Cloudera's secure private and public cloud environments. The integration targets organizations scaling autonomous AI agents and real-time decision systems that require sub-second access to fresh enterprise data while meeting governance and data sovereignty mandates.

Hangzhou – September 24, 2026 -- Amap Platform has launched Qianyu, a spatial-intelligence open platform that lets enterprises deploy location-based AI agents through natural-language prompts with no coding required. The announcement was made at the 2026 Apsara Conference in Hangzhou.

Qianyu packs 80 million points of interest into machine-readable agents

As of August 2026, the platform integrates more than 80 million POIs, over 5 million kilometers of roads and 7,500 square kilometers of three-dimensional models under a 'Point-Line-Plane-Dynamic' framework. Coverage extends to addresses, travel times, building outlines and area-level heatmaps.

Cui Yong, General Manager of Amap Platform, said the company has converted spatial intelligence into ready-to-use agents for developers and enterprises, enabling AI to support real-world business decisions such as store site selection and fleet routing rather than only answering digital queries.