At its annual Dell Technologies World conference, Dell is unveiling new laptops, servers and storage to help partners meet a market in which 75 percent of organizations say AI is key to their strategy and 65 percent are successfully moving AI concepts into production, according to one executive.
Dell Technologies said AI projects that have been in the works for years are now moving out of proof of concept and into production, setting the company and its partners up to have a “breakthrough year,” said Sam Grocott, senior vice president of product.
Dell Technologies is unveiling new laptops, servers and storage at its Dell Technologies World conference this week to help partners meet a market in which 75 percent of organizations say AI is key to their strategy and 65 percent are successfully moving AI concepts into production, Grocott said during a media pre-briefing ahead of the show.
“We are super excited about the maturing market and the way this is going and where I think this is setting us up to have a really breakthrough year,” Grocott said. “Last year we introduced the Dell AI Factory and we’ve been busy over the last year with over 200 new releases and innovations across the Dell AI Factory. We’ve also deployed more than 3,000 of these across many enterprises.”
[RELATED: ‘Fast and Furious’ Dell Brings AI To Speed Partner Sales: Channel Chief]
Grocott said Dell’s AI Factory approach, which drives users to place AI workloads on prem, can be up to 62 percent more cost-effective for inferencing large language models (LLMs) than the public cloud and helps organizations securely and easily deploy enterprise.
Grocott said the market is noticing and Dell is picking up customers among traditional technology clients such as GPU as a service, hyperscale and cloud services providers, but also seeing AI deployments in retail, finance, health care, and public sector.
To meet the demand for new devices and capabilities, Dell has new AI PCs, leading with its Dell Pro Max lineup of desktops and notebooks that incorporate Qualcomm chips to power the CPU as well as Nvidia GB300s to tackle the AI or graphics workloads. Additionally, Dell’s newest top-end PowerEdge servers, XE9780 and XE9780L, come in both air-cooled and liquid-cooled varieties, and can incorporate either AMD or Nvidia GPU architectures.
Michael Dell introduced the audience to the devices on stage at Dell Technologies World 2025, calling the products the “infrastructure backbone” of enterprises as they create intelligence with their data.
“Dell is the infrastructure, the backbone enabling enterprises to think faster, to act smarter and to dream bigger,” he told the crowd. “Today we have now more than 3,000 customers running Dell AI Factories with a lot of success. The Dell AI Factory is up to 60 percent more cost-effective than the public cloud. Now with agents, and test time compute and deep reasoning, the models are helping us think and they’re thinking and acting on their own, with autonomy.”
Dell Pro Max Plus laptop with Qualcomm
During a pre-briefing, the company said the Dell Pro Max Plus laptop comes with a Qualcomm AI-100 NPU that provides 32 AI cores. It is the industry’s first-ever enterprise-grade discrete NPU in a mobile form factor.
Dell said the device is designed to enable AI inferencing at the edge with an eye towards AI engineers and data scientists who can now run large models at scale at the edge with the ability to power a 109 billion parameter model on the device, such as Meta Llama 4.
Dell PowerCool Enclosed Rear Door Heat Exchanger (eRDHx)
Dell said this is another industry first, with self-contained airflow systems that captures 100 percent of IT heat. It can do this with warm water systems. The devices can save customers 60 percent on their cooling costs compared to the standard rear door heat exchanger units available today, Dell said.
The upside for compute is that organizations can deploy 16 percent more dense AI and HPC racks without increasing their power consumption, Dell said. It can cool up to 80-kilowatt racks and includes features such as hot swapable fans.
This pairs with the new Dell Integrated Rack Controller, which can be placed inside OpenManage Enterprise software. That combination offers IT administrators leak detection and thermal management, through rack scale visibility.
Dell AI Data Platform
The platform is built to ingest, discover and process data from the edge to the data center or in the cloud, said Varun Chhabra, senior vice president of infrastructure and telecom marketing at Dell.
One of the enhancements to the platform being unveiled is Project Lightning, which when it is released later this year will be the fastest parallel file system in the world, double the speed of traditional file systems. The AI data platform also integrates vector search for direct data set creation and querying within the Dell Data Lake House, which helps eliminate external systems for those capabilities.
Dell has also built query functions around SQL-accessible LLM functions, text summarization, sentiment analysis, with all of it baked into a customer’s instance of Dell’s AI Data Platform.
PowerEdge XE9785 and XE9785L servers with AMD Instinct MI350 Series
These are the newest versions of Dell’s flagship servers that come equipped with 288 gigabytes of HBM3E memory per GPU and deliver up to 35 times greater inferencing performance and are designed to house the AMD MI350 series chips.
The new server comes in both air-cooled and liquid-cooled versions.
Eventually they will be used to power Dell’s AI Platform With AMD, which currently runs on XE9680s and has also been upgraded in time for Dell Tech World.
Chhabra said while the platform was just introduced last year, this is the fourth update that combines Dell’s PowerEdge servers, PowerScale storage, and PowerSwitch networking with AMD’s ROCm AI accelerator framework, and MI300X GPUs for an 86 percent increase in customer’s time to value, Chhabra said.
Dell PowerEdge servers with Nvidia HGX B300
The newest versions of Dell’s PowerEdge servers, the XE9780 and XE9785, run Nvidia’s most advanced GB300 GPUs inside both liquid-cooled and air-cooled systems.
The air-cooled versions are 10U systems that can support Intel Xeon 6 as well as AMD Epyc 5 processors, and can fit inside standard 19-inch racks. It provides 1.8 gigabytes per second of GPU to GPU throughput for faster model training as well as faster inferencing.
It also features eight ConnectX 800-gigabit network interfaces per server, which lets customers cluster GPU servers together to create large scale systems for training and inferencing.
The newest version of PowerEdge also has liquid-cooled versions that come in 3U form factors and can create the highest density GPU systems that Dell can deploy. It comes standard with the ability to install 192 GPUs per rack, and it is customizable up to 256 HGX B300 GPUs in a single rack.
The machines are built to handle highly demanding AI workloads as well as HPC clusters.
