Claude mania is real

At HumanX 2026, one of the AI industry's marquee conferences in San Francisco, CNBC spoke with 19 executives and investors. The question was simple: which AI company has the momentum right now? The answer, almost unanimously, was Anthropic. Not OpenAI. Not Google. Anthropic. The phrase "Claude mania" kept surfacing in conversations across the conference floor. And the thing driving it wasn't a chatbot, a benchmark, or a flashy demo. It was a command-line coding tool called Claude Code. This post explores what it means when one company captures developer mindshare in the agent era, and whether this moment represents a durable shift or just another hype rotation.

The HumanX vibe check

HumanX isn't a developer conference. It's where enterprise buyers, VCs, and C-suite executives gather to figure out where the AI money is actually going. That makes the "Claude mania" signal particularly telling. This wasn't developers on Reddit gushing about a new tool. These were the people signing seven-figure contracts. Anthropic's early strength in the enterprise positioned it to ride the wave of AI coding agents. While OpenAI kicked off the generative AI boom with ChatGPT in 2022, Anthropic appears best positioned to win the contracts from the biggest spenders. The company crossed a $30 billion annualized revenue run rate by early April 2026, up from roughly $9 billion at the end of 2025. Claude Code alone has a run rate above $2.5 billion, having more than doubled since the start of the year. Business subscriptions to the tool have quadrupled since January, and enterprise users now represent more than half of Claude Code revenue.

How Claude Code shifted the conversation

Claude Code launched in early 2025 as a terminal-based coding agent. Unlike traditional autocomplete-style tools like GitHub Copilot, Claude Code takes a delegation-first approach. You describe what you want, and it plans, executes, and iterates across your codebase. Within 10 months of launch, it captured a 46% "most loved" developer rating, while GitHub Copilot, with a two-year head start and 20 million users, sat at just 9% developer satisfaction. According to one analysis, 71% of developers who regularly use AI coding agents use Claude Code. An estimated 4% of all GitHub public commits are now authored by Claude Code, with projections of 20%+ by end of 2026. The impact rippled beyond Anthropic. OpenAI scrapped Sora, its AI video generation app, just six months after launch. The reason, according to multiple reports: refocusing resources on developers and enterprise customers, partly in direct response to Claude Code's momentum. As one OpenAI executive reportedly put it, "We cannot miss this moment because we are distracted by side quests." OpenAI then rushed to acquire Windsurf, an AI coding startup, for $3 billion, and launched a new $100/month ChatGPT Pro subscription tier specifically to compete on coding. The entire industry pivoted.

The Pentagon paradox

Here's where the story gets interesting. At the very moment Anthropic was dominating HumanX, it was also embroiled in an unprecedented legal battle with the U.S. Department of Defense. In late February 2026, the Pentagon blacklisted Anthropic and labeled it a "supply chain risk," a designation normally reserved for companies from adversarial nations like Huawei. The reason? Anthropic refused to lift safety guardrails on Claude for military use, specifically declining to allow its AI for mass surveillance or autonomous weapons systems without human oversight. A California federal judge temporarily blocked the order in March, calling it "classic First Amendment retaliation." But a D.C. appeals court separately refused to pause the blacklisting, creating a split legal landscape. The case is still unfolding. The paradox is striking. A company banned by the Pentagon is simultaneously the hottest name in enterprise AI. If anything, the confrontation appears to have helped. Anthropic's main app topped Apple's download charts during the dispute, riding a wave of public sympathy. Revenue kept accelerating through the entire ordeal, jumping from $14 billion to $19 billion to $30 billion in annualized run rate within weeks. It turns out that standing up to the government on AI safety principles, whatever you think of the politics, is remarkably good for brand credibility with enterprise customers who care about responsible deployment.

Enterprise vs. consumer mindshare

The competitive landscape in April 2026 looks very different from a year ago. ChatGPT's U.S. app market share fell from 69.1% in January 2025 to 45.3% in January 2026. Google Gemini rose from 14.7% to 25.2%. Even Grok climbed from 1.6% to 15.2%. But market share in consumer apps tells only part of the story. Anthropic is winning where the money is: enterprise contracts. The company now serves over 300,000 business customers. Goldman Sachs is using Claude to automate banking tasks. Cognizant has equipped 350,000 employees with Claude. Claude Code is embedded in engineering workflows at companies that would never show up in an app download chart. OpenAI still has the stronger consumer brand. ChatGPT is a household name. But Anthropic has carved out a different position: the serious enterprise AI company that developers trust. And in the long run, enterprise revenue is what funds the next generation of models.

The developer darling trap

History offers reasons for caution. Being the developer darling of the moment doesn't guarantee long-term dominance. Heroku was the beloved deployment platform of the early 2010s. Developers loved its simplicity, its developer experience, its opinionated design. Then AWS got good enough, Kubernetes emerged, and Heroku stagnated under Salesforce's ownership. It still exists, but it lost the narrative. Parse was Facebook's mobile backend-as-a-service, and developers built thousands of apps on it. Facebook shut it down in 2016. The pattern repeats across tech: developer affection is necessary but not sufficient for durability. The question for Anthropic is whether Claude Code is sticky in the way that matters, or whether it's riding a wave that competitors can replicate in six months. Cursor already offers a compelling alternative with a different philosophy (control-first vs. delegation-first). Google has Gemini and deeper distribution through Android and Cloud. OpenAI has more capital and a massive user base to convert. The counterargument is that Claude Code isn't just a feature. It's the wedge product for an entire enterprise relationship. Once a company builds its engineering workflows around Claude Code, switching costs compound. The coding agent becomes the entry point for financial analysis, sales automation, cybersecurity, and scientific research, all areas where Anthropic is now expanding.

The "one agent, one job" thesis

There's a broader implication to Claude Code's rise. We may be entering an era where AI companies win not by building the best general-purpose chatbot, but by owning specific, high-value workflows. Coding is arguably the perfect first wedge. Developers are early adopters by nature. The feedback loop is immediate (does the code work or not?). The economic value is obvious and measurable. And once you've proven you can handle complex, multi-step reasoning in code, the same capabilities extend naturally to other domains. Anthropic's Claude Cowork, which launched in early 2026, exports Claude Code's agentic capabilities beyond the IDE and into general business workflows. It connects to Google Drive, Gmail, DocuSign, and FactSet, with customizable plugins for finance, engineering, and HR. One LinkedIn post noted it was "built entirely by Claude Code" in just 10 days, a recursive proof that the system that reasons about work can produce systems that reason about work. This is the real strategic bet. Claude Code isn't just a product. It's proof of a capability that Anthropic can deploy across every knowledge work category.

Is this a durable shift?

The honest answer: it's too early to know. But several factors suggest this isn't just another hype rotation. First, the revenue is real and accelerating. Going from $9 billion to $30 billion in annualized run rate in about four months is not a vibes-based phenomenon. Enterprise customers are signing real contracts. Second, the competitive response confirms the threat. OpenAI doesn't kill Sora and pivot to enterprise unless it sees an existential challenge to its business model. Google doesn't rush Gemini Code unless the market is moving. Third, Anthropic's cost structure appears more sustainable. OpenAI projects $14 billion in losses for 2026 and expects to spend $121 billion on compute. Anthropic projects positive free cash flow by 2027 while spending roughly a quarter of that on training. But there are real risks. Claude Code's dominance could be model-dependent, and model advantages are famously temporary. The Pentagon situation, while boosting the brand short-term, could limit government revenue long-term. And Anthropic is still private, meaning the growth story hasn't been tested by public market scrutiny. The most honest framing might be this: Claude mania is real in the sense that Anthropic has genuinely captured the enterprise AI zeitgeist in a way no one predicted two years ago. Whether it's permanent depends on whether they can convert developer love into infrastructure-level lock-in before the competition catches up. If history is any guide, the window for that conversion is probably 12 to 18 months. The clock is ticking.