By Dell TechnologiesLearning Objectives
- Understand what Dell AI Factory is and why on-premises AI infrastructure matters
- Identify the key components of Dell's AI server lineup and NVIDIA partnership
- Evaluate Dell's competitive position in the AI infrastructure market
What Is Dell AI Factory?
Dell AI Factory is Dell Technologies' comprehensive AI infrastructure platform, launched in March 2024 in partnership with NVIDIA. It provides everything an organization needs to build, train, and run AI systems on-premises: servers with the latest GPUs, high-speed networking, AI-optimized storage, and as-a-service consumption options.
While much of the AI conversation focuses on cloud services, many organizations — especially governments, financial institutions, and healthcare systems — need to keep their AI infrastructure on-premises for data sovereignty, security, and cost reasons. Dell AI Factory addresses this market as the world's number one server vendor.
💡Key Concept
On-Premises AI Infrastructure: Running AI workloads on servers physically located in your own data center rather than renting from cloud providers. This gives organizations full control over their data, avoids ongoing cloud costs, and meets regulatory requirements around data residency. The trade-off is higher upfront capital investment and the need for internal hardware expertise.
Current Server Lineup
Dell's PowerEdge AI servers span from single-GPU rack servers to massive liquid-cooled GPU clusters:
| Server | GPUs | Cooling | Use Case |
|---|---|---|---|
| PowerEdge XE9780/XE9785 | 8x NVIDIA Blackwell B300 | Air or liquid | Flagship training and inference |
| PowerEdge XE8712 | Up to 144 NVIDIA Blackwell GPUs per rack | Liquid | Large-scale training clusters |
| PowerEdge XE9785L | 8x AMD Instinct MI355X | Liquid | AMD GPU alternative to NVIDIA |
| PowerEdge R770/R7715/R7725 | NVIDIA RTX PRO 4500 Blackwell | Air | Edge AI and smaller workloads |
| Dell Pro Max with GB300 | Desktop AI supercomputer | Air | 20 petaFLOPS; 748 GB coherent memory |
Coming in late 2026: PowerEdge XE9812 with NVIDIA Vera Rubin NVL72 platform, and the XE9880L/XE9882L/XE9885L series with Rubin NVL8.
Beyond Servers
Dell AI Factory is more than just servers:
Storage
Dell PowerScale and a new AI data platform provide high-throughput storage for training data pipelines. Project Lightning, announced in May 2025, is a parallel file system Dell claims is the world's fastest for AI training workloads.
Networking
PowerSwitch SN5610 and S2201 switches running Dell SONiC with NVIDIA Spectrum-X technology — purpose-built for the high-bandwidth, low-latency networking that GPU clusters demand.
Dell APEX (As-a-Service)
A pay-per-use consumption model for on-premises AI infrastructure. Instead of a large upfront capital purchase, organizations pay based on usage — bringing cloud-like economics to on-premises hardware. Hybrid cloud options available with Microsoft Azure.
Software and Services
Knowledge assistant (generally available), agentic AI platform with integrations from Cohere North, DataRobot, and ClearML. Dell's professional services team helps organizations design, deploy, and manage AI infrastructure.
AI Revenue and Orders
Dell's AI business has grown explosively:
| Metric | FY2025 | FY2026 | Growth |
|---|---|---|---|
| AI Server Revenue | $9.73 billion | $24.56 billion | 2.5x |
| Total AI Orders (fiscal year) | Not disclosed | $64 billion+ | Record |
| AI Backlog (entering next FY) | Not disclosed | $43 billion | Record |
| Total Company Revenue | $95.6 billion | $113.54 billion | +18.8% |
Dell's $43 billion AI server backlog entering fiscal year 2027 is so large that it has reportedly triggered DRAM shortages and downstream hardware repricing across the industry.
Dell AI Factory vs. Competitors
| Vendor | Strength | Best For |
|---|---|---|
| Dell AI Factory | Number 1 server market share; largest AI backlog ($43 billion); full-stack (servers + storage + networking + APEX) | Enterprises and governments needing complete on-premises AI infrastructure |
| HPE | GreenLake hybrid cloud/AI-as-a-service; Cray XD670 with HGX B300; strong in HPC | High-performance computing and hybrid cloud AI |
| Supermicro | 10-15% lower hardware costs; 2-4 week faster delivery; ~70% liquid cooling market share | Cost-sensitive deployments needing maximum rack density |
| Lenovo | TruScale pay-as-you-go; gigawatt-scale AI factory program with NVIDIA | Large-scale AI factories with flexible payment models |
| NVIDIA DGX | Reference architecture; tightest GPU integration; DGX SuperPOD | Organizations wanting NVIDIA's own optimized systems |
Company Details
| Detail | Info |
|---|---|
| Company | Dell Technologies (NYSE: DELL) |
| Founded | 1984 |
| CEO | Michael Dell (founder) |
| Headquarters | Round Rock, Texas |
| Employees | ~108,000 |
| Revenue (FY2026) | $113.54 billion (+18.8% year-over-year) |
| AI Server Revenue (FY2026) | $24.56 billion (2.5x year-over-year) |
| AI Backlog | $43 billion entering FY2027 |
| Market Cap | ~$117 billion |
| AI Factory Customers | 4,000+ (enterprises, governments, neoclouds) |
| Server Market Share | Number 1 globally (IDC, calendar year 2025) |
| Website | dell.com/ai |
Strengths
- Number one server vendor globally — largest market share by revenue (IDC 2025) with the broadest PowerEdge lineup
- $43 billion AI backlog — provides exceptional revenue visibility and validates demand
- Full-stack offering — servers, storage, networking, software, and services in one vendor; reduces integration complexity
- GPU flexibility — supports both NVIDIA Blackwell and AMD Instinct GPUs, avoiding single-vendor lock-in
- APEX as-a-service — brings cloud-like consumption pricing to on-premises AI infrastructure
- Sovereign AI trust — governments across Europe, Middle East, and Asia choose Dell for domestic AI infrastructure
Limitations and Considerations
- Enterprise-only pricing — no published pricing; all configurations are custom-quoted through sales teams
- Capital-intensive — on-premises AI infrastructure requires significant upfront investment (unless using APEX)
- Hardware complexity — deploying and managing GPU clusters requires specialized expertise that many organizations lack
- No cloud AI services — Dell sells infrastructure, not AI APIs; you still need software platforms (Databricks, NVIDIA NIM, etc.) on top
- Lead times — the massive backlog means delivery times can stretch to months for popular configurations
Key Takeaways
- Dell AI Factory is the leading enterprise AI infrastructure platform — combining the world's number one server lineup with storage, networking, and as-a-service options
- AI server revenue reached $24.56 billion in FY2026 (2.5 times year-over-year growth) with a $43 billion backlog entering FY2027
- Designed for organizations that need on-premises AI infrastructure for data sovereignty, security, or cost reasons — supported by 4,000+ customers including sovereign governments
- Supports both NVIDIA and AMD GPUs with configurations ranging from single-GPU rack servers to 144-GPU liquid-cooled clusters
