Dell Technologies has moved its artificial intelligence strategy from experimental pilots to full-scale production, according to remarks made at the Six Five Summit: AI Unleashed 2026 on Tuesday.
The company, which serves as its own test case for AI deployment, has reduced a portfolio of 900 potential AI projects to 13 targeted initiatives. Dell analyzed 6,800 roles to identify areas where autonomous agents could enhance productivity, with early deployments showing gains between 20% and 40%.
John Roese, Dell’s global chief technology and AI officer, described the company as "Customer Zero" for internal testing of autonomous agents. He emphasized that AI agents are designed to augment rather than replace roles, stating that "agents do not take your job, they change your job."
The shift reflects a broader transition in enterprise AI from chatbot-driven data access to agentic systems capable of digitizing workflows. Roese noted that the first phase of generative AI in business focused on unlocking proprietary data, while the current phase involves automating entire work processes.
Dell’s financial performance over the past 12 months underscores the scale of its AI investments. Revenue reached $134 billion, a 38.6% increase, while diluted earnings per share climbed to $12.51. The company’s stock rose 3.76% to $449.47, trading near its 52-week range of $110.22 to $514.00. Wall Street price targets for Dell range from $360 to $700.
Technical deployment details highlight the evolving infrastructure demands of agentic AI. Dell reported a shift in CPU-to-GPU ratios from traditional AI workloads of 8-to-1 or 16-to-1 to ratios closer to 1-to-1 or 2-to-1, as CPUs take on more orchestration and coordination tasks. The company sources tokens from four to five different channels, including on-premises open models, frontier models in controlled virtual private clouds, APIs, and device-based models.
Headless agents, which operate without direct human interaction, are expected to account for roughly 70% of future deployments. Use cases span from high-cost CEO decision-support agents priced at $100,000 per month to low-cost CRM cleanup tools under $0.50 per record.
