Anthropic CEO Dario Amodei said the AI company had planned for 10x growth, but instead, its revenue and usage grew 80-fold in the first quarter on an annualized basis, a surge he called \u201cjust crazy\u201d and \u201ctoo hard to handle.\u201d\u00a0<\/p>\n
The company is growing so fast that its infrastructure has struggled to keep up, forcing it into an unlikely partnership with an industry rival: Elon Musk.<\/p>\n
\u201cThat is the reason we have had difficulties with compute,\u201d Amodei said during Anthropic\u2019s developer conference in San Francisco Wednesday, CNBC reported<\/a>.\u00a0<\/p>\n Anthropic\u2019s annualized revenue has climbed<\/a> to $30 billion, a three-fold increase compared to last year. Much of this increase has been fueled by a wave of corporate customers including Uber and Netflix<\/a> which are increasingly using Claude Code. Customers prefer Claude Code in some cases over competitors like Cursor because of its agentic abilities. Companies have also leveraged Claude\u2019s API to embed the large language model into everything from customer service platforms to task automation.\u00a0<\/p>\n \u201cSoftware engineers are the ones who are fastest to adopt new technology,\u201d Amodei said at the developer\u2019s conference. \u201cIt\u2019s a foreshadowing of how things are going to work across the economy, and how the economy is going to be transformed by AI.\u201d<\/p>\n Yet, in recent months, some Claude Code users have complained<\/a> about increasingly strict usage limits that have hurt their experience and the model\u2019s performance. Anthropic admitted<\/a> in a postmortem from late April that three bugs had affected Claude Code since March 4. The company\u2019s internal tests hadn\u2019t caught them, which is what led to several weeks of degraded performance. Yet, when users first flagged the issues in early March, the company said it could not reproduce them and didn\u2019t make an effort to fix the problem.<\/p>\n On social media, some users over the past several weeks have said they were cancelling their Claude subscriptions because of performance issues.\u00a0<\/p>\n \u201cWhatever happened in the last 1-2 months is a significant regression,\u201d wrote<\/a> one user in a post on X<\/a>.<\/p>\n Separately, Anthropic\u2019s skyrocketing growth has turned into an infrastructure emergency, and as a result the company struck an agreement with an unlikely ally. Anthropic announced<\/a> earlier this week it would rent the entire compute capacity of Colossus 1, the Memphis data center built by Musk\u2019s xAI, would allow for higher usage limits for Claude Code across all paid plans and the Claude API. The company is doubling Claude Code\u2019s five-hour rate limits for its Pro, Max, Team, and seat-based Enterprise plans. It\u2019s also doing away with peak hour limit reductions on Pro and Max accounts, according to the announcement<\/a>.<\/p>\n The deal gives Anthropic access to 220,000 NVIDIA GPUs<\/a> and 300 megawatts of new compute within the month, but it is now also at least partly reliant on Musk, a competitor in the AI race. One analyst told Fortune the deal could bring in between $3 billion and $4 billion<\/a> in annual revenue for SpaceX<\/a>.\u00a0<\/p>\n Anthropic\u2019s deal with SpaceXAI comes even as it has struck several deals previously to expand its compute. In early April, Anthropic signed a joint deal with Google<\/a> and Broadcom<\/a> that will give it access to five gigawatts of compute capacity. Less than a month later, it struck a deal with Amazon<\/a> Web Services to add up to another five gigawatts of compute in exchange for an agreement to spend more than $100 billion on AWS services over the next decade.\u00a0<\/p>\n The agreement with Anthropic is good news for SpaceXAI, which has aimed to diversify its business since acquiring xAI<\/a> in February. On Wednesday, Musk said in a post<\/a> on X that xAI will no longer exist as a separate company, and will simply serve as the AI arm of the newly named SpaceXAI. The company is preparing for an IPO later this year that is aiming to raise between $1.75 and $2 trillion.<\/p>\n And yet, Anthropic is turning to SpaceXAI despite the previous attacks by Musk in which he has labeled the company \u201cmisanthropic and evil.<\/a>\u201d In SpaceXAI\u2019s announcement of the deal with Anthropic, the company said Anthropic expressed interest in partnering with it to develop orbital data centers, which Musk has argued could provide boundless low-cost compute capacity, despite the engineering challenges<\/a>.<\/p>\n