Artificial intelligence news.
Straight from the original announcements.

CoreWeave Opens LGA1, a New GPU Data Center on the Hudson River in NYC

October 2, 2026, 16:12Source: CoreWeave
CoreWeave Opens LGA1, a New GPU Data Center on the Hudson River in NYC
Image: CoreWeave

CoreWeave has introduced LGA1, a new data center on the Hudson River at the mouth of the Lincoln Tunnel in New York City. According to the company, the site offers sub 1ms latency to Manhattan and is aimed at businesses across the East Coast.

Key points
  • CoreWeave introduced LGA1, a data center on the Hudson River at the mouth of the Lincoln Tunnel in NYC.
  • The company claims sub 1ms latency to Manhattan and low latency for the 20 million+ people in the metro area.
  • LGA1 offers NVIDIA GPU instances including Quadro RTX, RTX Ampere workstation and A40 GPUs, plus CPU-only instances and Block and Shared File System storage.
  • CoreWeave now has 3 data centers, linked by 400Gbps+ of dark fiber.
  • Customers can launch a Virtual Server in LGA1 from the CoreWeave Cloud UI.

What LGA1 offers

CoreWeave says LGA1 serves the more than 20 million inhabitants of the New York metropolitan area with ultra-low latency, including sub 1ms latency to Manhattan. The company presents the facility as a way for East Coast businesses to use its ultra-low latency virtual workstations, its broad fleet of GPUs and what it calls industry-leading pricing.

The campus is ISO 27001 certified, SSAE 18 SOC 2 compliant and Energy Star Certified, according to CoreWeave. It is built around a Tier 1 internet backbone and is designed for compute-intensive use cases that demand performance, reliability and security.

Hardware and storage

LGA1 is stocked with NVIDIA GPU accelerated Cloud Instances. CoreWeave lists the Quadro RTX series, the newest RTX Ampere workstation GPUs and A40 data center GPUs. The company also describes "nearly unlimited GPU compute and availability" at the site.

Beyond GPUs, the data center offers:

  • CPU-only instances
  • High performance Block storage
  • High performance Shared File System storage

Network and how to get started

With LGA1, CoreWeave now has 3 data centers that, in its words, provide low latency access to accelerated compute across all of North America. The company says each of its data centers, including its campus in Las Vegas, has redundant 200Gbps+ public internet connectivity from Tier 1 global carriers. The sites are linked to each other by 400Gbps+ of dark fiber transport, which CoreWeave says allows easy, free transfers of data within CoreWeave Cloud.

Customers can try LGA1 by launching a Virtual Server from the CoreWeave Cloud UI. According to CoreWeave, workloads can be scheduled in any of its US regions, whether they are containerized micro-services or Virtual Servers, using its region selector. The company also invites customers with questions to contact its team.

What it means for customers

For organizations on the East Coast, the main change is proximity: GPU and CPU resources are now available from a facility inside the New York area, with the latency figures CoreWeave cites for Manhattan. Free data transfers between its data centers, as described by the company, may also matter to customers who spread workloads across regions. Pricing details were not included in the announcement.

Related news

NVIDIA brings Korea’s leading open models to production with Dynamo
For developersNovember 9, 2026, 06:00

NVIDIA brings Korea’s leading open models to production with Dynamo

NVIDIA is releasing Dynamo-optimized deployment recipes for four of Korea’s leading open models, including models from SK Telecom, LG AI Research, Upstage and Motif…

NVIDIA outlines five ways to make every AI memory byte work harder
Chips and infrastructureNovember 5, 2026, 17:00

NVIDIA outlines five ways to make every AI memory byte work harder

NVIDIA has outlined five co-designed technologies aimed at increasing the amount of useful AI work produced by each byte of memory. The company says the approach enables…

NVIDIA Details Full-Stack Approach to AI’s Growing Memory Pressure
Chips and infrastructureNovember 5, 2026, 17:00

NVIDIA Details Full-Stack Approach to AI’s Growing Memory Pressure

NVIDIA is detailing a full-stack codesign strategy aimed at easing the memory pressure created by larger AI models, longer contexts and agentic applications. The company…