Anthropic is actively promoting Claude as a foundation for developer-built businesses, yet its aggressive move into enterprise applications is triggering alarm among founders concerned that their model supplier could transform into a direct rival.
A nervous warning to wrappers
Gerald Zingraf, president and COO of GlacierGrid, voiced his unease on LinkedIn after Anthropic announced its Tuesday expansion of Claude's enterprise offerings, likening the company's trajectory to Amazon's approach to private-label products.
Makes me nervous for wrappers and pure SaaS companies
Gerald Zingraf
Zingraf draws on firsthand experience: he previously sold products through Amazon's platform, then watched the retailer launch competing offerings. His worry centers on Anthropic's visibility into which third-party integrations its customers rely on most, and whether the company might leverage that intelligence when deciding what to build internally.
He contends that Anthropic maintains the same vantage point, though he provides no documentation that the company actually mines integration data to spot competitors or shape its product direction. For some builders, however, this risk has already materialized.
When partners become rivals
Scott Stevenson, CEO of legal AI startup Spellbook, recounted his experience when Anthropic entered his market in a November 2 interview with Forbes.
Spellbook harnesses Claude models to handle contract reviews, propose edits, and streamline legal operations. Stevenson recalls Anthropic staff pledging the company would not pursue direct competition in legal software—a promise that preceded Anthropic's launch of Claude for Legal, which incorporates contract-review functionality that directly overlaps with Spellbook's offerings.
Anthropic has been a great partner, but it made us realize that they are a massive company and two parts of the org can have conflicting mandates
Scott Stevenson
Stevenson added that such competition is ultimately unavoidable. Despite the tension, Spellbook continues using Claude. Anthropic still showcases the startup as a customer, citing approximately 530,000 monthly contract reviews performed by its Claude-driven agents, and Stevenson participated in Anthropic's October 28 announcement of legal artifacts.
The situation reveals the precarious position of developers whose applications depend on cutting-edge language models. Spellbook still gains from Anthropic's technology even as Anthropic develops competing applications for overlapping customer segments. Stevenson's account contains no indication that Anthropic appropriated Spellbook's proprietary data to construct its legal product.
When the platform ships your product
Todd Saunders, founder and CEO of Dalton Mills, articulated comparable concerns in a March 10 post on X following Anthropic's rollout of Claude Code Review.
Anthropic is the new Amazon
Todd Saunders
Saunders cautioned developers building against Claude's SDK that Anthropic could monitor how applications are used and eventually fold competing capabilities directly into Claude.
On November 10, Anthropic unveiled Claude Sheets, Claude Whiteboards, and Claude Data, while showing Claude Meetings in preview. The company also expanded interactive connectors, permitting partners to embed their applications directly within Claude, including offerings from Adobe and Atlassian.
What the policies leave unsaid
Anthropic possesses information about customer interactions with third-party apps and services—a fact the company discloses in its privacy policy dated September 10, 2026. What remains opaque is whether Anthropic examines which integrations gain the most traction when determining its next product priorities.
Enterprise customers operate under different rules. Anthropic's commercial terms grant customers ownership of their inputs and outputs, with Anthropic pledging not to use customer content for model training. The company's commercial privacy documentation also specifies that feedback data excludes content routed through connectors, including local and remote MCP servers, unless users directly paste it into a chat.
Neither policy clarifies how Anthropic handles aggregated integration data.
The tradeoff developers accept
Anthropic continues inviting developers to construct on Claude. The exchange, however, involves a cost: access to its expanding base of enterprise customers comes bundled with Anthropic's control over the interface and its ability to introduce competing features without consulting the developers who depend on it.
At present, Anthropic is asking developers to place faith in it as both a technology vendor and a business ally, even as it penetrates sectors where those same developers are attempting to establish enterprises. The durability of this arrangement may hinge less on Anthropic's current assurances than on what the company chooses to release next.
The New Stack reached out to Anthropic for a statement and has not yet received one.