On March 6, Anthropic launched Claude Marketplace, a platform that lets enterprise customers purchase third-party software built on Claude using their existing Anthropic spend commitments. No new vendor contracts. No separate security reviews. No procurement cycles.

The launch came one day after the Pentagon formally designated Anthropic as a supply-chain risk, making it the first American company ever to receive a label previously reserved for foreign adversaries like Huawei. That designation, the culmination of a weeks-long standoff over whether Claude could be used for autonomous weapons and domestic mass surveillance, threatened to freeze enterprise adoption.

Anthropic's response was not to retreat. It was to go deeper into enterprise.

Six partners went live on day one: Snowflake, GitLab, Harvey, Replit, Rogo, and Lovable. The roster is deliberately broad, spanning data infrastructure, DevOps, legal AI, software development, financial research, and no-code app building.

To understand the Marketplace launch, you need to understand the week that preceded it.

In July 2025, Anthropic signed a $200 million contract with the Pentagon that made Claude the first frontier AI model approved for use on classified networks. As part of that agreement, the Pentagon agreed to abide by Anthropic's acceptable use policy.

Then the Pentagon sought to renegotiate. It demanded that Anthropic allow the military to use Claude "for all lawful purposes" without limitation. Anthropic had two redlines: no mass surveillance of American citizens, and no fully autonomous weapons systems with no human in the targeting loop. The Pentagon refused those restrictions. When a February 27 deadline passed without agreement, President Trump ordered federal agencies to cease using Anthropic's technology, and Defence Secretary Pete Hegseth designated the company a supply-chain risk.

The move drew bipartisan criticism. Senator Kirsten Gillibrand called it "shortsighted, self-destructive, and a gift to our adversaries." Thirty former military and intelligence officials wrote to Congress urging an investigation into the precedent being set.

Hours after the designation, OpenAI announced it had secured a deal to replace Anthropic in classified military environments.

Here is the part that did not make most headlines. According to NPR, the dispute triggered a consumer surge: more than a million people signed up for Claude each day in the days following the standoff, lifting it past ChatGPT and Gemini as the top AI app in over 20 countries on Apple's App Store. The public, it turns out, rewarded the company that said no.

But consumer downloads do not pay for frontier model research. Enterprise contracts do. Which makes the Marketplace launch a pointed strategic move: Anthropic is doubling down on commercial enterprise adoption at precisely the moment its government revenue is under threat.

Enterprise AI adoption has a well-documented bottleneck, and it is not the technology. It is procurement.

Every new AI tool a team wants to deploy requires vendor evaluation, security review, legal sign-off, and budget approval. For large organisations, that process can take months. By the time a tool is approved, the landscape has shifted, the team has moved on, or a competitor has already shipped.

Claude Marketplace collapses that entire cycle into a single step: allocate a portion of your existing Anthropic commitment. Anthropic manages all invoicing for partner spend. The enterprise has already done the security review on Claude. The budget is already approved.

As VentureBeat reported, observers noted that Claude Marketplace effectively lets enterprises "pre-approve" apps, bypassing the cautious approval process that kills adoption. That single detail solves a problem that has frustrated enterprise AI deployment for years.

The timing carries an additional subtext. For any enterprise general counsel worried about Pentagon exposure, Anthropic is saying: the supply-chain designation is narrowly scoped to direct military contracts. Microsoft, Google, and Amazon all confirmed they can continue working with Anthropic on non-defence projects. The Marketplace makes it even easier to keep buying.

The six launch partners are not random. They map to the core workflows that define enterprise operations.

Harvey handles legal AI for contracts, due diligence, and litigation support. Rogo provides financial models, research, and pitch deck generation for investment professionals. Replit enables natural-language software development. GitLab covers DevOps and the full code lifecycle. Snowflake delivers governed data agents across structured and unstructured enterprise data. Lovable lets non-technical teams build and ship real applications from prompts.

The pattern is clear. Legal questions route to Harvey. Financial analysis routes to Rogo. Apps get built in Lovable. Data gets queried through Snowflake. Code ships through GitLab and Replit.

All powered by the same Claude intelligence layer, under one bill.

According to The Next Web, both Harvey and Replit are deeply dependent on Claude as their underlying model. For those partners, the Marketplace provides distribution into enterprise customer relationships that Anthropic has already cultivated, where budgets are committed, security reviews are complete, and contracts are signed.

Here is the detail that separates this from a standard marketplace play: Anthropic is taking zero commission on Marketplace transactions. According to Bloomberg via SiliconANGLE, Anthropic will not take a cut from purchases made through the platform.

That stands in sharp contrast to AWS Marketplace and Microsoft Azure Marketplace, which typically collect revenue shares from partner transactions. Anthropic is openly comparing its model to those cloud marketplaces, with one critical difference: it is forgoing the revenue that cloud giants have historically collected.

The reason is straightforward. Anthropic does not need marketplace revenue. What it needs is ecosystem lock-in.

Every partner tool purchased through Claude Marketplace deepens an enterprise's relationship with Anthropic, not with the underlying software vendor. The intelligence layer, Claude, is the constant. The partners are the variables. The more tools an enterprise runs through this layer, the harder it becomes to switch to a competitor.

This is platform strategy, not marketplace strategy. And in the context of the Pentagon dispute, it is also a resilience strategy. If government revenue dries up, a deeply embedded commercial enterprise base becomes the foundation Anthropic builds on instead.

The obvious comparison is OpenAI's ChatGPT App Directory, launched in December 2025. Both companies are building ecosystems around their AI models. The execution could not be more different.

OpenAI's App Directory targets consumers and individual users. Its early partners include Canva, Expedia, Figma, Spotify, and Zillow. Users invoke apps via "@" mentions in ChatGPT conversations. The focus is on retail, travel, and personal productivity. OpenAI is reaching 800 million ChatGPT users with an experience designed for individual workflows.

Anthropic's Marketplace targets procurement officers and CIOs. Its partners are enterprise software companies operating in legal, finance, data infrastructure, and software development. Purchases flow through existing spend commitments. The focus is on consolidating enterprise AI budgets under a single vendor relationship.

As The Next Web noted, OpenAI's integration model focused on consumer-facing workflows, while Anthropic's marketplace is positioned further up the enterprise stack.

The rivalry has an additional edge now. Hours after Anthropic was designated a supply-chain risk, OpenAI swooped in to take Anthropic's classified military contracts. OpenAI CEO Sam Altman claimed his company's Pentagon deal included the same two safety limitations Anthropic had insisted on, but phrased differently, agreeing to "all lawful purposes" while reportedly embedding the restrictions elsewhere in the agreement. The distinction reads more like legal engineering than a genuine policy difference.

The market will decide which bet pays off.

The bigger picture here is not about a marketplace. It is about who becomes the default intelligence layer for enterprise software.

The Snowflake partnership provides context. In late 2025, Snowflake and Anthropic announced a $200 million multi-year partnership that brings Claude to Snowflake's 12,600 global customers. Thousands of those customers already process trillions of Claude tokens per month through Snowflake Cortex AI.

That is not a pilot. That is infrastructure-grade adoption.

If the Marketplace model works, enterprises will not need to evaluate and procure 15 separate AI tools. They will adopt one AI platform and plug specialised agents into it as needed. Legal, finance, development, data, operations, all powered by the same intelligence layer, all billed through the same relationship.

That is what an operating system does. It provides the base layer that everything else runs on.

The Pentagon dispute, paradoxically, may have accelerated this thesis. Anthropic is now a company with a clear identity: the AI lab that prioritises safety, even when it costs them a $200 million government contract. For enterprise buyers in regulated industries like financial services, healthcare, and legal, where compliance and ethical AI are not optional, that positioning is not a liability. It is a selling point.

Four things will determine whether Claude Marketplace becomes a defining infrastructure play or a footnote.

First, partner expansion. Six launch partners is a proof of concept. The marketplace needs coverage across every major enterprise function, from HR to customer success to supply chain, to become indispensable.

Second, enterprise adoption velocity. The value proposition only works if enterprises actually redirect spend. Anthropic says the Marketplace is in limited preview, with interested customers reaching out through their account teams. The speed at which that pipeline converts will tell us everything.

Third, the legal outcome. Anthropic has said it will challenge the supply-chain designation in court. If it wins, the Marketplace becomes a growth engine layered on top of restored government revenue. If it loses, the Marketplace becomes the commercial moat that makes government revenue less critical.

Fourth, competitive response. Microsoft and Google both have deep enterprise relationships and their own AI models. OpenAI is now the Pentagon's preferred AI partner and has its own consumer app ecosystem gaining traction. The window for Anthropic to establish Claude as the default enterprise intelligence layer is open, but it will not stay open forever.

Anthropic just lost a $200 million Pentagon contract for refusing to remove AI safety guardrails, then launched a zero-commission enterprise marketplace the very next day. Is this the most calculated move in enterprise AI history, or the most reckless?