Salesforce doubles down on agents at Dreamforce; Nscale files for NYSE listing

Salesforce Inc. has made a strategic bet on artificial intelligence agents, and early indicators suggest the wager is working. During Dreamforce 2026, Marc Benioff, Salesforce's chief executive officer, shared the stage with Jensen Huang, Nvidia Corp.'s chief executive officer, to introduce AIforce, a new AI interface layer. The conference also showcased Headless 360, a capability that lets customers interact with Salesforce applications independently of the traditional user interface.

Salesforce's 'system of intelligence' emerges

At Dreamforce, Salesforce unveiled its vision for what it calls an Enterprise AI Harness—a unified architecture that brings together data, business logic, processes and governance to support composable agents. This direction represents movement toward a "system of intelligence" that serves as a framework for agents to interpret and navigate enterprise operations.

This whole headless thing points to the fact that AI agents are not another feature cycle, meaning agents aren't a feature of the enterprise. They're forcing a complete reconstruction of the enterprise stack … from the applications, the end interfaces, down through security, infrastructure and capital allocation.

John Furrier, executive analyst at theCUBE Research

Dave Vellante, chief analyst at theCUBE Research, elaborated on the architecture's role: "Headless becomes that system of intelligence. That's where the data is harmonized, the process knowledge gets injected. You see harnesses within the LLMs, but they're not deterministic. AIforce … allows any client surface to basically create the interface. The reality is this stuff is not mature — Agentforce is still in its early days — but there's clear business value."

Salesforce has made agents central to its strategy, partly to sidestep concerns about a potential "SaaSpocalypse" that has unsettled the investment community. The company's emerging product portfolio relies heavily on composable agents, which theoretically offer customers enhanced flexibility and control.

Furrier noted that Salesforce's underlying strengths lie in its accumulated data systems, workflow infrastructure, identity management and business context. He suggested that Slack could function as an entry point for agentic capabilities, since unstructured data aligns well with graph structures and the computational patterns that AI systems favor.

On-premises deployment has gained momentum as a revenue driver for AI workloads. Salesforce continues to refine its pricing model, according to Vellante. Research from Qualitate indicates that 75% of respondents who have modeled or experienced the financial consequences of headless access anticipate their Salesforce expenditures will grow rather than shrink.

You're seeing Claude create the interface. You're seeing Slack create the interface. Virtually, any harness is going to be able to create the interface. That's going to drive token demand. Where's that token demand going to come from? Core infrastructure in the data center. A lot of that is going to be on-prem. There's no question about it.

Dave Vellante

Nscale moves toward public markets

During the episode, news broke that Nscale Global Holdings Ltd., a United Kingdom-based developer of AI-optimized data centers, submitted paperwork for an initial public offering on the New York Stock Exchange. The Nvidia-backed company, which has not yet achieved profitability, is positioning itself to capitalize on expanding demand for data center resources.

Furrier characterized Nscale as following a playbook similar to CoreWeave's approach—a fully integrated operation spanning the entire stack. CoreWeave acquired Anyscale, creating a substantial cloud infrastructure platform aimed at serving enterprise AI needs. Nscale now competes alongside Nebius and CoreWeave in this space. Furrier predicted the offering would generate significant market interest.

The conversation also turned to how agents are evolving within enterprise environments. As AI systems begin making increasingly autonomous choices grounded in operational data, Salesforce's headless offering could fundamentally alter enterprise workflows.

https://www.youtube.com/embed/O71RdBGRmmw?feature=oembed

Furrier raised a broader question about agent development: "What happens when agents start to learn about how enterprises actually work over time? You have a new training inference dataset. Spirit Airlines bankruptcy — there's a bid for the data, 20 years of data. The buyers of that data were for training purposes. That brings up the commoditization of the system of record. At what point does that become disposable?"