Anthropic CEO Dario Amodei revealed that while the company initially planned for tenfold growth, its first-quarter revenue and usage surged eightyfold on an annualized basis. This explains the company's struggle to meet demand. Speaking at Anthropic's developer conference in San Francisco on Wednesday, Amodei stated, "This is why we are experiencing difficulties with computing capacity." He added that the company is "working to provide more computing power as quickly as possible" and will make it available to users "as fast as we can."
Amodei noted that software engineers are the quickest to adopt new technologies, suggesting how AI will transform the entire economic system in the future. He described the current growth level as "simply insane" and "difficult to manage," expressing hope for a "more normal" expansion pace in the future.
Last month, Anthropic acknowledged that demand for its Claude AI model had placed "inevitable strain on infrastructure," affecting the "stability and performance" of the user experience, particularly during peak hours. In response to the computing power crunch, Anthropic adjusted the pricing model for its enterprise product, Claude Enterprise. Under the new structure, enterprise clients pay a base fee of $20 per user per month plus additional charges based on actual computing power consumption, replacing the previous fixed subscription model that charged up to $200 per user monthly for a set amount of discounted token usage.
Anthropic explained that the old model caused frequent work interruptions for heavy users hitting usage limits, while lighter users paid for unused capacity. A company spokesperson said the new pricing "better reflects how customers actually use Claude, as workloads shift from fixed-seat productivity tools to agent applications." Fredrik Filipsson, co-founder of Redress Compliance, which helps companies negotiate software licenses, commented that the new pricing could double or even triple costs for heavy users.
The explosive growth is primarily driven by the popularity of Anthropic's Claude AI models, particularly since the launch of Claude Code last year. Beyond existing partnerships with Amazon and NVIDIA, Anthropic is actively acquiring additional computing resources. Hours before Amodei's remarks, Anthropic announced a computing partnership with SpaceX, gaining full access to the computing capacity of SpaceX's Memphis Colossus 1 data center. This collaboration is expected to add over 300 megawatts of new capacity within a month, equivalent to 220,000 NVIDIA GPUs.
Additionally, recent reports indicate Anthropic has committed approximately $200 billion to Google Cloud over the next five years for training and deploying Claude. This investment reflects multiple agreements with Google, including an October announcement expanding Anthropic's use of Google Cloud TPU chips, with Google providing over 1 gigawatt of computing power this year. Last month, Anthropic secured a new agreement with Google and Broadcom for approximately 3.5 gigawatts of TPU resources, scheduled to come online starting in 2027. Amazon also announced a deepened strategic collaboration with Anthropic last month, locking in up to 5 gigawatts of current and next-generation Trainium chip capacity for training advanced AI models.
Strong demand for Claude is directly reflected in Anthropic's revenue. The company announced last month that its annualized revenue exceeded $30 billion, a significant increase from $9 billion at the end of 2025. Third-party research firm Semi Analysis reported in early May that Anthropic's annualized revenue had risen to approximately $44 billion, far surpassing OpenAI's growth rate during the same period. Enterprise clients contribute over 80% of revenue, with eight of the Fortune 10 being stable customers. The number of clients with annual payments exceeding $1 million has surpassed 1,000, doubling from previous figures. The Claude Code programming assistant alone generates nearly $2.5 billion in annualized revenue, a more than fivefold increase from $400 million in mid-2025, with enterprise use accounting for over half of this revenue. This indicates Claude Code is evolving from a coding tool into a serious enterprise software solution requiring budget, permissions, audit, and organizational management considerations. For upstream providers, this shift means more sustained demand reliant on stable delivery rather than sudden spikes around model releases.
Enterprise spending data from early 2026 shows Anthropic capturing 73% of the market share among companies purchasing AI tools for the first time, overtaking OpenAI. Anthropic's advantage in the enterprise market is linked to its "safety-first" product positioning. As a leading AI model developer focused on AI safety and alignment research, Anthropic's models demonstrate superior compliance and accuracy, leading to rapid adoption in heavily regulated sectors like finance, law, and healthcare.
According to informed sources, Anthropic has initiated its latest funding round, requesting investors to submit allocation proposals promptly. The round is expected to raise approximately $50 billion and close by mid-May. Anthropic is targeting a valuation of around $900 billion, but strong investor demand could push the final valuation higher. The company's valuation growth trajectory is remarkable: rising from $183 billion in September 2025 to $380 billion in February 2026, a 107% increase. The current target of $900 billion represents more than a doubling from February, potentially matching or surpassing its main competitor OpenAI, which achieved an $852 billion post-money valuation after a record $122 billion financing round in March 2026.
However, analysis suggests three significant risks in Anthropic's growth narrative. First, customer concentration remains a structural risk. Despite having over 300,000 enterprise clients, revenue is highly concentrated among top customers. Besides Amazon and Google contributing over 20% of revenue, Anthropic's business model is deeply tied to cloud platforms through long-term agreements, including up to $30 billion in Azure compute purchases with Microsoft, over $100 billion in AWS compute purchases with Amazon over the next decade, and up to $40 billion in investment and 5 gigawatts of TPU capacity with Google. While these agreements come with substantial equity investments and reciprocal benefits, potential price increases or supply fluctuations from core cloud providers could impact profit margins. Reduced purchasing by major clients, a shift to in-house model development, or support for competitors could also directly affect revenue stability.
Second, Anthropic's unique governance structure may concern investors. Organized as a Public Benefit Corporation (PBC), its core mechanism is the Long-Term Benefit Trust (LTBT), which holds special Class T shares with the power to elect a majority of the board. Even after going public, strategic decisions will prioritize "humanity's long-term benefit" over shareholder returns. While this structure provides policy advantages amid increasing AI regulation, public market investors may view it as a "mission discount"—the trust's veto power over certain business decisions could subordinate profits to ethical considerations, limiting investor voting rights and returns. Balancing the trust's mission with shareholder interests will be a core challenge.
Third, the high valuation premium places immense pressure on Anthropic to deliver strong performance. Based on the $900 billion target valuation and approximately $44 billion in annualized revenue, the price-to-sales ratio is about 20 times, significantly above the SaaS industry average of 8-12 times. To justify the current valuation, Anthropic needs to achieve $70 billion in revenue and $17 billion in cash flow by 2028, requiring at least 50% annual growth over the next three years alongside continuous margin improvement. Although Semi Analysis reports that Anthropic's inference infrastructure gross margin has improved from about 38% twelve months ago to over 70%, indicating better unit economics, persistently high model training and inference costs, coupled with price competition from rivals like OpenAI and Google's Gemini, make achieving these targets challenging.
Notably, Anthropic's IPO process is in direct competition with OpenAI. After its March 2026 financing round valued it at $852 billion, OpenAI also plans an IPO in the second half of 2026. If Anthropic goes public first, it could absorb substantial pent-up investor demand for AI stocks, potentially weakening market response to OpenAI's IPO. This competitive pressure is accelerating both companies' listing preparations. Anthropic has already appointed Goldman Sachs, JPMorgan, and Morgan Stanley as lead underwriters, with long-term legal counsel Wilson Sonsini handling IPO compliance, indicating a slightly faster pace than OpenAI.
To truly establish a market position rivaling OpenAI, Anthropic must overcome challenges in ecosystem development, cost control, and consumer market expansion. Market acceptance of Anthropic's IPO will largely depend on whether investors are willing to pay a premium for its "safety-first" differentiation and unique governance structure. Even if the IPO is successful, maintaining technological leadership and increasing market share in an industry where model iterations occur in under twelve months will remain key focus areas for the market.
Comments