Two people familiar with the company’s financials said Anthropic is projecting revenue of approximately $190 billion to $200 billion by 2028. The figure has not previously been reported and represents a dramatic increase from the company’s publicly disclosed revenue run rate of about $47 billion as recently as May.
The projection highlights the extraordinary growth expectations investors would be taking on if they participate in an eventual IPO. Rather than relying primarily on current earnings, bankers and investors are reportedly using enterprise-value-to-revenue multiples based on future forecasts, according to four sources familiar with the discussions.
Revenue multiples are widely used to value high-growth software companies that have yet to establish a mature and consistent profit profile. However, basing the valuation on projections as far out as 2028 is less common and reflects both the pace of Anthropic’s expansion and the difficulty of determining what its financial performance could look like at scale.
AI Spending Complicates Valuation
A major challenge for investors is Anthropic’s enormous spending on the infrastructure required to develop and operate advanced AI models. The company is investing heavily in computing power, GPUs, model training, inference and hiring, expenses that can put significant pressure on margins even as revenue rises rapidly.
That spending has also weighed on sentiment toward some of the technology companies seen as potential comparisons for Anthropic. Concerns about the huge capital requirements of the AI boom have contributed to pullbacks in several highly valued technology stocks in recent months.
Still, there are precedents for investors using distant financial projections when valuing fast-growing technology companies. Backers of Cerebras Systems, for instance, cited expectations for 2028 revenue ahead of its IPO, while SpaceX investors considered projections extending to 2029 before the company went public at a record valuation in June.
For Anthropic, the underlying bet is that revenue will eventually grow faster than the costs associated with producing it. Investors are therefore looking beyond the company’s current margins and focusing on the possibility that computing and model-training costs become more efficient while personnel and other expenses represent a smaller percentage of revenue as the business scales.
Palantir, Cloudflare and SpaceX Among Comparisons
As Anthropic prepares for its analyst day and potential public listing, investors are also examining comparable companies to establish a framework for its valuation.
Cloud infrastructure provider Cloudflare, enterprise software company Palantir and SpaceX are among the businesses being considered as reference points, according to people familiar with the process.
Comparable companies are an important part of IPO valuation because they provide investors with a benchmark for determining how businesses with similar growth characteristics or exposure to emerging technologies are priced. They can also help establish the revenue or earnings multiples that should be applied to a company’s projections.
Palantir currently trades at about 53 times its expected 2026 revenue, making it one of the most highly valued companies in the public market on a revenue basis. SpaceX and Cloudflare both trade at roughly 41.6 times expected 2026 revenue, according to LSEG data.
Each company offers a different comparison for Anthropic. Palantir has become a closely watched benchmark for businesses combining rapid growth with artificial intelligence exposure. Cloudflare provides a comparison with a high-growth software and infrastructure company, while SpaceX demonstrates how investors can place significant value on a company’s anticipated future scale rather than its current financial performance.
Anthropic’s Revenue Growth Drives Investor Optimism
Traditional measures such as earnings and EBITDA are less useful when assessing a company like Anthropic at its current stage because its financial profile is heavily influenced by the cost of expansion.
The company’s spending on GPUs, computing capacity, model training, inference and employees is essential to supporting its growth. However, investors are betting that these costs will eventually become less burdensome relative to revenue as Anthropic expands and AI infrastructure becomes more efficient.
Anthropic’s recent financial trajectory illustrates the speed of that expansion. Its revenue run rate stood at approximately $9 billion at the end of 2025 before climbing to more than $47 billion by May, according to the company.
The company has also projected at least $10.9 billion in revenue for the second quarter of 2026, more than twice the previous quarter’s figure, while targeting its first quarterly operating profit of $559 million.
Anthropic has said its revenue run rate increased by more than tenfold annually in each of the three years leading up to early 2026, a pace of expansion that helps explain why investors are willing to consider projections several years into the future.
The central question for investors is whether the company’s current spending represents an investment that will eventually produce significantly higher revenue and stronger margins. Improvements in AI technology could make training and inference more efficient, while operating and personnel costs could decline as a proportion of revenue.
David Merkel, a principal at investment firm Aleph Investments, said Anthropic could potentially reach a valuation of $2 trillion, although he questioned whether such a valuation would remain sustainable.
“Could they (Anthropic) get a $2 trillion valuation, yeah they could and I just wonder if it would stay there over time,” Merkel said.
He also urged investors to consider whether AI would ultimately generate enough additional productivity to justify the enormous valuations being placed on companies in the sector.
“Does it (AI) really produce so much additional productivity... These are just questions that we have to ask if we were thinking of pricing this, buying this.”
Anthropic did not immediately respond to a request for comment.