Last price
$10.41
The RPA incumbent now repositioning as the enterprise agent runtime — thesis-native if the pivot holds, disintermediated if it doesn't.
Live quote sourced from Yahoo Finance. Prices cited in narrative below reflect the original memo date and may be stale.
UiPath is the most exposed public incumbent to both sides of the Sequoia thesis: if agents are the new software, UiPath's runtime position is gold; if LLM-native startups (Adept, Claude Computer Use, native agent frameworks) disintermediate the RPA logic layer, UiPath's legacy is a liability. The pivot is explicit: Agent Builder turns UiPath into the orchestration fabric for LLM-based agents; robots become one tool in the agent toolbox; Autopilot products extend copilot surface across developers and business users. Outcome pricing exists in RFPs but isn't yet the headline motion.
The framing is binary and loud: does UiPath own the enterprise agent runtime, or does it get eaten by LLM-native agents running on Claude/Gemini/GPT directly against enterprise systems? The bull view is that governance, compliance, audit trails, human-in-the-loop workflows, and reliability requirements force enterprises to route their agents through a governed fabric — and UiPath has been building that for a decade. The bear view is that AI-native agent companies are free of RPA's rigid worldview and will simply out-execute on the greenfield. There is very little middle ground here.
Enterprises are moving agents from demos to production and hitting a wall: you cannot point an LLM at a live ERP and hope. Regulated customers need audit logs, role-based access, credential vaulting, human-in-the-loop gates, dry-run testing, and failure isolation. UiPath has those muscles. Agent Builder re-uses the governance layer for LLM-based agents, and Agent Orchestration mixes agents, robots, and humans in the same workflow with the same audit trail. The incumbent account base (banks, telcos, healthcare, public sector) is the most governance-constrained buyer and the slowest to adopt greenfield agent startups — giving UiPath a window.
LLM-native agent platforms (Adept, Anthropic's Computer Use, Cognition, a dozen AI-first startups, plus every hyperscaler's native agent framework) can in principle act on the same systems UiPath sits on, without the rigid rule-authoring of classic RPA. If they catch up on governance, UiPath's moat narrows fast. Meanwhile, Microsoft Power Automate and Salesforce Agentforce bundle agentic automation into the seat an enterprise already pays for. The pivot story is credible but not yet evidence-backed at scale.
| Product | Role | Thesis fit | Status |
|---|---|---|---|
| Agent Builder + Agent Orchestration | LLM-based agent runtime + orchestration | Core | Live; early-stage deployments |
| Unattended / attended robots | Classic RPA execution | Core | Cash cow + expansion base |
| Autopilot for Studio / Everyone | Developer + business-user copilots | Core | Live; net-expansion driver |
| Process Mining + Task Mining | Discovery + prioritisation | Supporting | Mature |
| Document Understanding + Communications Mining | Intelligence layer | Core | Mature + extending via LLM |
| AI Trust Layer + governance | Compliance + audit | Core moat | Shipped |
Enterprises cannot run unsupervised LLM agents on their core systems. UiPath's decade of compliance, audit, human-in-loop, credential, and failure-containment tooling is exactly what regulated buyers need. Competitors must replicate that muscle before they can win large-logo deals.
Agent Builder is architected for LLMs (Claude, Gemini, GPT) with retrieval, memory, human-in-the-loop gates, and robot-tool invocation. It is not wrapping RPA in LLM costume; it is a greenfield product that reuses the governance stack. That distinction matters for enterprise architects.
11K customers and 2K $100K+ accounts translate to a large net-expansion surface for Autopilot products and Agent Builder. Net new logo growth has re-accelerated; expansion within the base is the bigger revenue lever.
Net cash and consistent FCF give UiPath multi-year runway to execute the agent pivot without capital-market dependency. That's a rare advantage among thesis-native incumbents.
The incumbent threat isn't a startup — it's Power Automate inside M365 and Agentforce inside Salesforce. Those platforms reach every enterprise employee at zero additional sticker price and are improving fast. UiPath has to justify a separate line item.
Startups building agents from scratch — without classic RPA's rigidity — will typically win greenfield use cases, and increasingly the more interesting ones. UiPath's land motion may slow as the agent conversation shifts away from process automation.
R&D on Agent Builder + Autopilot depresses operating margin while classic RPA growth decelerates. The numbers can look worse before they look better, which is what activists usually attack.
If the enterprise ultimately consolidates on (say) Microsoft + Salesforce + Azure OpenAI + Anthropic, UiPath must live as an integrator across their stacks. That's a defensible but thinner business than being the primary runtime.
UiPath is the single best 'evolution vs. obsolescence' binary in the public set. If the thesis of 'enterprise agent runtime' becomes real, UiPath has the best incumbent position to win it — governance, execution rails, customer relationships, cash. If instead enterprises consolidate on hyperscaler-bundled agents or AI-native startups, UiPath's existing surface is stranded cost. The thesis is more asymmetric here than almost anywhere in the index; the next 18 months of Agent Builder traction are the tell.
Best-positioned incumbent in the agent runtime fight, with real governance moat — but the pivot is still probabilistic. Size to the binary.