Anthropic's Signup Bug, Cognizant Deal, Open-Weights Stance

Three distinct Anthropic stories landed over the past 48 hours: a technical incident that temporarily shut the door on new Claude signups, an enterprise partnership with Cognizant, and a clarification from CEO Dario Amodei on where Anthropic stands on open-weights AI.

Key points

• A 200-line userscript reportedly caused Claude’s payment form to render as a SEPA form, prompting Anthropic to halt new user registrations while the issue propagated for six business days before resolution.
• Cognizant has expanded its partnership with Anthropic to embed Claude directly into Cognizant’s industry-specific platforms.
• Anthropic CEO Dario Amodei publicly clarified the company’s position on open-weights AI following Nvidia CEO Jensen Huang’s push for more open AI development, per Digit.in.
• All three stories touch on different parts of Anthropic’s business: product reliability, enterprise distribution, and competitive positioning.

What happened with the signup bug?

The technical write-up published on Medium describes a scenario where a userscript of roughly 200 lines interfered with Claude’s card payment form, causing it to display as a SEPA (Single Euro Payments Area) form instead. The downstream effect was significant enough that Anthropic closed new signups while the problem worked its way through the system over six business days.

The incident is a useful reminder of how surface-level UI bugs can have outsized consequences when payment flows are involved. A prospective subscriber hitting a SEPA form instead of a standard card entry would almost certainly abandon the process, and depending on how many users encountered the broken state, the revenue and growth impact could be meaningful. Anthropic has not publicly commented on the scale of the disruption based on available sources.

For investors tracking user growth as a proxy for Claude’s commercial momentum, any period of closed signups is worth flagging, even if the root cause turns out to be narrow in scope.

Does the Cognizant deal matter for enterprise distribution?

Cognizant Technology (CTSH) is one of the larger IT services firms globally, with deep relationships across financial services, healthcare, and manufacturing. Embedding Claude into Cognizant’s industry platforms means Anthropic gains a distribution layer it does not have to build or sell itself. Cognizant handles the client relationships; Claude handles the inference.

The partnership is described as an expansion of an existing arrangement, which suggests Anthropic and Cognizant have already been working together and are now deepening the integration. Details on revenue sharing, exclusivity, or specific platform names were not available in the sources at time of publication.

This follows a pattern Anthropic has been building across consulting and IT services partners. Each deal of this type adds indirect enterprise reach without requiring Anthropic to staff a direct salesforce at the same scale. From a pre-IPO perspective, diversified distribution through established enterprise intermediaries reduces customer concentration risk and broadens the addressable pipeline.

Where does Anthropic stand on open-weights AI?

Following Jensen Huang’s public advocacy for more open AI development, Anthropic CEO Dario Amodei offered a clarification of the company’s position. The detail available from Digit.in is limited, but the framing suggests Amodei felt it necessary to distinguish Anthropic’s approach from the direction Huang was promoting.

This site covered Anthropic’s open-weights stance on July 27, when the company stayed out of an open-weights advocacy letter that had gathered 50 signatories. Amodei’s latest comments appear to be a continuation of that positioning rather than a reversal, though the sources do not provide the full text of his remarks. Readers should treat this as a signal of ongoing public debate rather than a new policy announcement.

Anthropic’s caution on open-weights models is grounded in its stated safety-first mission. Whether that stance affects enterprise sales or talent recruitment in a market where Meta’s Llama models are widely adopted remains an open question that will likely surface in any eventual IPO roadshow.

Putting it together

None of Monday’s stories individually moves the needle on Anthropic’s trajectory, but together they sketch a company managing product reliability problems, expanding enterprise distribution, and defending a differentiated safety posture in public. The signup bug is the most operationally unusual of the three; the Cognizant deal is the most directly revenue-relevant; and the open-weights comments are the most strategically telling for long-term positioning.