In the fast-paced capital narrative of the artificial intelligence track, Anthropic is reshaping the market landscape with astonishing financial expectations. According to a独家 report from The Information, this startup known for "safe AI" is expected to achieve up to $70 billion in revenue by 2028 and generate $17 billion in operating cash flow — while its main competitor, OpenAI, will still be stuck in a trillion-dollar cash loss pit during the same period.
This contrast stems from Anthropic's clear and aggressive B2B commercialization path. According to insiders, the company is expected to earn $3.8 billion this year just from selling AI models through APIs, nearly twice the API revenue of OpenAI during the same period ($1.8 billion). Its enterprise product line is growing rapidly: Claude Code, designed for developers, has an annualized revenue approaching $1 billion, more than doubling from $400 million in July.
Enterprise collaborations are flourishing, with integration into workflows becoming key
Recently, Anthropic has accelerated embedding Claude deeply into global enterprise ecosystems: not only has it partnered with Microsoft to integrate the model into Microsoft 365 and Copilot, but it has also expanded its strategic alliance with Salesforce and plans to deploy dedicated AI assistants for tens of thousands of employees at giants like Deloitte and Cognizant. These moves indicate that Anthropic is transitioning from a "model provider" to an "enterprise intelligent infrastructure."
At the same time, the company continues to optimize its product matrix — over the past two months, it has launched lighter and more economical versions of Claude Sonnet 4.5 and Haiku 4.5, meeting the needs of large-scale deployment; it also released Claude for Financial Services and Enterprise Search features, allowing companies to connect internal documents, CRM, project management, and other systems to Claude, creating a unified intelligent entry point.
Financial model surprising reversal: Gross margin surges from -94% to 77%
Even more impressive is its dramatic turnaround in profitability. In 2024, Anthropic's gross margin was **-94%, but it is expected to jump to 50% this year, reaching 77% by 2028**. This means its unit economic model has been validated, and large-scale expansion no longer relies on endless burning of money.
In contrast, despite OpenAI's valuation reaching $500 billion and a user base of 800 million weekly active users, its aggressive infrastructure investments will lead to a cash consumption of $14 billion in 2026, and cumulative cash burn may exceed $115 billion before 2029, making profitable prospects difficult to see in the short term.
Valuation may reach $400 billion, but concerns remain
Anthropic's last round of financing has raised $13 billion, with a valuation of $17 billion. If it continues its current growth momentum for another round of financing, the market expects its valuation to reach between $30 billion and $40 billion. However, challenges remain: the company still carries a $2.5 billion credit facility and $1.5 billion in copyright litigation settlement costs, and long-term compliance costs cannot be ignored.
Nevertheless, Anthropic has proven that in the AI competition, it's not the one with the most users who wins, but the one who monetizes most steadily. While the industry is still debating technological boundaries, it has quietly built a high-margin, strong-cash-flow AI corporate empire — and this B2B revolution may have just begun.
