Upscale AI Achieves $2 Billion Valuation Amidst Critical AI Infrastructure Demand
June 24, 2026, 9:37 pm
Upscale AI secured $190 million in Series A-1 funding. This elevated its valuation to $2 billion. Total funding now stands at $500 million. The company pioneers AI networking infrastructure. It develops a full-stack, open-standard AI fabric. This system tackles critical network bottlenecks in large-scale AI training and inference. Upscale AI's technology includes advanced scale-up switches with deterministic latency via SkyHammer chips. It also features scale-out switches integrating NVIDIA’s Spectrum-X and AI-optimized SONiC. Leading hyperscalers and neocloud providers are evaluating the solutions. The capital injection will accelerate commercialization and global delivery of its AI-native networking technology. This marks a significant step in optimizing modern AI infrastructure.
Upscale AI just closed a significant funding round. The AI networking firm announced $190 million in Series A-1 financing. This capital boost elevates the company’s valuation to $2 billion. It brings total funding to a substantial $500 million. This marks a major milestone for the Santa Clara-based company.
Premji Invest led the latest round. Other key investors joined. NVIDIA, Salesforce Ventures, Seligman Ventures, and Temasek participated as new backers. Existing investors also reaffirmed their commitment. These included Maverick Silicon, Mayfield, Prosperity7 Ventures, StepStone Group, and Tiger Global. The strong investor confidence underscores Upscale AI's market position.
The newly secured funds serve a clear purpose. Upscale AI will scale its operations. It will accelerate delivery of its advanced AI-native networking technology. Demand for robust AI infrastructure is surging. This investment empowers the company to meet that demand.
Upscale AI operates as a pure-play AI networking infrastructure company. It builds full-stack technology. This includes silicon, systems, and software solutions. Its core mission is to create an open-standard AI fabric. This fabric connects accelerators, memory, and storage. The goal is a high-performance AI engine.
Modern AI workloads present unique challenges. Training and inference for frontier models require immense computational power. They also demand seamless data flow. Network bottlenecks frequently impede performance. Upscale AI's platform aims to reduce these crucial bottlenecks. It targets large-scale, synchronized AI computations.
Data centers rely on specialized switches. These manage data traffic. Scale-up switches move data within a single rack. Scale-out devices handle traffic between racks. Upscale AI develops both. Its switches are specifically optimized for artificial intelligence workloads.
AI model calculations are sequential. One delay can cascade. Graphics processing units (GPUs) exchanging data slowly cause inefficiencies. This leads to unexpected latency spikes. Such delays postpone subsequent computations. Upscale AI directly addresses this issue.
Its scale-up switch series offers deterministic latency. Data movement speeds become predictable. This eliminates unexpected delays. The feature relies on a custom chip. Upscale AI calls this chip SkyHammer.
SkyHammer supports multiple open-source network protocols. These protocols are optimized for scale-up traffic. UALink is one such technology. It enables GPUs to access data in each other's memory. This makes remote memory appear local. SkyHammer also works with ESUN. ESUN is an Ethernet protocol variant. It is specifically optimized for AI tasks.
Upscale AI also offers scale-out switches. These link together racks of graphics cards. The product line leverages NVIDIA’s Spectrum-X chip series. These processors power NVIDIA's Ethernet switch lineup. NVIDIA sells these alongside its GPUs.
Traditional traffic between GPUs often routes through a central processing unit (CPU). This adds latency. NVIDIA's Spectrum-X switches support RoCE technology. RoCE allows data packets to bypass the CPU. This significantly reduces delay. The devices also gather telemetry data. This helps administrators pinpoint technical issues.
Upscale AI's scale-out switch series combines Spectrum-X silicon with open-source software. This software is called SONiC. Microsoft originally developed SONiC for its public cloud. It streamlines many network management tasks. However, using it can be complex. Upscale AI developed an AI-optimized version of SONiC. This version simplifies the user experience.
The company's hardware is not merely theoretical. It is undergoing active evaluation. Multiple hyperscalers and neocloud operators are testing the technology. These customer engagements span both scale-out and scale-up networking environments. The deployments are already underway.
The vision is clear: AI infrastructure needs redefining. It must operate at a cluster scale. Networking often forms the most critical bottleneck. Upscale AI builds a high-performance, open-standard AI fabric. It is purpose-built for large-scale, synchronized workloads.
This investment strengthens Upscale AI's strategic vision. It enables crucial business scaling. The company aims to meet the escalating demand for open, interoperable AI infrastructure. Leadership points to the team’s ability to tackle complex, system-level challenges. They execute with remarkable clarity.
The future of AI performance hinges on networking. It ties compute, memory, and storage together. Upscale AI reimagines AI networking. It offers an open-standard, full-stack approach. This ensures accelerators, memory, and storage operate in perfect synchrony.
Building advanced AI necessitates a rethinking of data center infrastructure. Organizations demand scalable, interoperable systems. They want to avoid vendor lock-in. Upscale AI addresses this challenge. It combines a commitment to open networking standards. It also features a zero-packet-loss architecture. This maximizes hardware utilization.
Upscale AI's work is vital. It unlocks the full potential of high-density AI workloads. The team is positioned to drive innovation. Their long-term view anticipates AI infrastructure evolution. The company plans to build lasting leadership in this expanding market.
The AI networking space is booming. Upscale AI is a fast-growing player. This investment will fund immediate priorities. It also advances long-term ambitions. The company makes rapid progress building its product portfolio. This capital infusion ensures continued execution.
The era of frontier AI demands new solutions. Upscale AI aims to power this new era. It builds the next-generation networking company. Its technology supports ultra-low-latency networking. This enables performance and scalability across AI training, inference, and generative AI deployments. This funding confirms its leadership position. It paves the way for wider adoption of its innovative solutions.
Upscale AI just closed a significant funding round. The AI networking firm announced $190 million in Series A-1 financing. This capital boost elevates the company’s valuation to $2 billion. It brings total funding to a substantial $500 million. This marks a major milestone for the Santa Clara-based company.
Premji Invest led the latest round. Other key investors joined. NVIDIA, Salesforce Ventures, Seligman Ventures, and Temasek participated as new backers. Existing investors also reaffirmed their commitment. These included Maverick Silicon, Mayfield, Prosperity7 Ventures, StepStone Group, and Tiger Global. The strong investor confidence underscores Upscale AI's market position.
The newly secured funds serve a clear purpose. Upscale AI will scale its operations. It will accelerate delivery of its advanced AI-native networking technology. Demand for robust AI infrastructure is surging. This investment empowers the company to meet that demand.
Upscale AI operates as a pure-play AI networking infrastructure company. It builds full-stack technology. This includes silicon, systems, and software solutions. Its core mission is to create an open-standard AI fabric. This fabric connects accelerators, memory, and storage. The goal is a high-performance AI engine.
Modern AI workloads present unique challenges. Training and inference for frontier models require immense computational power. They also demand seamless data flow. Network bottlenecks frequently impede performance. Upscale AI's platform aims to reduce these crucial bottlenecks. It targets large-scale, synchronized AI computations.
Data centers rely on specialized switches. These manage data traffic. Scale-up switches move data within a single rack. Scale-out devices handle traffic between racks. Upscale AI develops both. Its switches are specifically optimized for artificial intelligence workloads.
AI model calculations are sequential. One delay can cascade. Graphics processing units (GPUs) exchanging data slowly cause inefficiencies. This leads to unexpected latency spikes. Such delays postpone subsequent computations. Upscale AI directly addresses this issue.
Its scale-up switch series offers deterministic latency. Data movement speeds become predictable. This eliminates unexpected delays. The feature relies on a custom chip. Upscale AI calls this chip SkyHammer.
SkyHammer supports multiple open-source network protocols. These protocols are optimized for scale-up traffic. UALink is one such technology. It enables GPUs to access data in each other's memory. This makes remote memory appear local. SkyHammer also works with ESUN. ESUN is an Ethernet protocol variant. It is specifically optimized for AI tasks.
Upscale AI also offers scale-out switches. These link together racks of graphics cards. The product line leverages NVIDIA’s Spectrum-X chip series. These processors power NVIDIA's Ethernet switch lineup. NVIDIA sells these alongside its GPUs.
Traditional traffic between GPUs often routes through a central processing unit (CPU). This adds latency. NVIDIA's Spectrum-X switches support RoCE technology. RoCE allows data packets to bypass the CPU. This significantly reduces delay. The devices also gather telemetry data. This helps administrators pinpoint technical issues.
Upscale AI's scale-out switch series combines Spectrum-X silicon with open-source software. This software is called SONiC. Microsoft originally developed SONiC for its public cloud. It streamlines many network management tasks. However, using it can be complex. Upscale AI developed an AI-optimized version of SONiC. This version simplifies the user experience.
The company's hardware is not merely theoretical. It is undergoing active evaluation. Multiple hyperscalers and neocloud operators are testing the technology. These customer engagements span both scale-out and scale-up networking environments. The deployments are already underway.
The vision is clear: AI infrastructure needs redefining. It must operate at a cluster scale. Networking often forms the most critical bottleneck. Upscale AI builds a high-performance, open-standard AI fabric. It is purpose-built for large-scale, synchronized workloads.
This investment strengthens Upscale AI's strategic vision. It enables crucial business scaling. The company aims to meet the escalating demand for open, interoperable AI infrastructure. Leadership points to the team’s ability to tackle complex, system-level challenges. They execute with remarkable clarity.
The future of AI performance hinges on networking. It ties compute, memory, and storage together. Upscale AI reimagines AI networking. It offers an open-standard, full-stack approach. This ensures accelerators, memory, and storage operate in perfect synchrony.
Building advanced AI necessitates a rethinking of data center infrastructure. Organizations demand scalable, interoperable systems. They want to avoid vendor lock-in. Upscale AI addresses this challenge. It combines a commitment to open networking standards. It also features a zero-packet-loss architecture. This maximizes hardware utilization.
Upscale AI's work is vital. It unlocks the full potential of high-density AI workloads. The team is positioned to drive innovation. Their long-term view anticipates AI infrastructure evolution. The company plans to build lasting leadership in this expanding market.
The AI networking space is booming. Upscale AI is a fast-growing player. This investment will fund immediate priorities. It also advances long-term ambitions. The company makes rapid progress building its product portfolio. This capital infusion ensures continued execution.
The era of frontier AI demands new solutions. Upscale AI aims to power this new era. It builds the next-generation networking company. Its technology supports ultra-low-latency networking. This enables performance and scalability across AI training, inference, and generative AI deployments. This funding confirms its leadership position. It paves the way for wider adoption of its innovative solutions.