Spark Capital's $75 million investment in Anthropic is now worth $7 billion

In 2023, Yasmin Razavi made a very bold decision.

She represented Spark Capital and invested $75 million in an AI company that had almost no revenue and no mature product at the time.

The company is called Anthropic.

In just three years, the paper value of that investment has changed dramatically. According to The Wall Street Journal on July 18, based on Anthropic’s latest valuation, Spark Capital’s stake in Anthropic is worth approximately $7 billion.

A $75 million investment generated nearly 100x in paper returns.

During China's internet era, people have seen similar stories before.

In 2010, when Zhang Lei’s Hillhouse Capital invested in JD.com, JD had not yet proven it could become an e-commerce giant. At the time, JD originally planned to raise about $75 million, but Zhang Lei believed that the core of future e-commerce competition would not just be traffic, but supply chain and logistics capabilities—ultimately leading Hillhouse to invest $300 million. This investment later became one of Hillhouse’s most famous cases.

Xu Xin’s investments in NetEase and JD.com also occurred before these companies were fully recognized by the market. When she invested in NetEase in 2000, the Chinese internet industry was in a trough following the burst of the dot-com bubble, and NetEase’s stock price had plummeted, with widespread skepticism about the future of internet companies. However, she believed in the long-term value of growing internet user bases and viable business models, ultimately reaping substantial returns.

Today, Anthropic is creating its own "Zhang Lei and Xu Xin."

Yasmin Razavi of Spark Capital is also betting on the same opportunity: whether a company that has yet to prove its business value can become the next-generation technology infrastructure.

At the beginning of 2023, the AI industry's attention was almost entirely focused on OpenAI.

At that time, ChatGPT had only been live for a few months and quickly became a global phenomenon. Microsoft announced additional investment in OpenAI, and Silicon Valley investors began reassessing the opportunities presented by generative AI.

In investors' eyes, OpenAI possesses nearly all the advantages of a market leader: it was the first to demonstrate the value of large language models to the public, boasts the strongest brand influence, and has received billions of dollars in support from Microsoft.

At the time, OpenAI was becoming synonymous with the AI era.

In contrast, Anthropic did not have the influence it has today—at the time, the company, founded by former OpenAI researchers in 2021, aimed to develop safer and more reliable large language models. But from an investor’s perspective, ideas do not equate to business value.

At the beginning of 2023, Anthropic was still a company that needed to prove itself. It had a top-tier research team and had developed its own large model, Claude, but it was still far from becoming a mature commercial company.

According to The Wall Street Journal, Yasmin Razavi conducted in-depth research on the AI industry before deciding to invest in Anthropic.

She is not focused on short-term model rankings, but rather on the longer-term industrial structure. In her view, large models may not just be ordinary software products, but could become important infrastructure for future technology industries.

If this judgment holds, then the key to competition among model companies is not just who has more users today. This also means that even if one model company is currently leading, it does not mean the market cannot accommodate other significant players.

After months of discussion, Yasmin Razavi, representing Spark Capital, led Anthropic’s Series C funding round, investing $75 million in the company.

In the following years, Anthropic's development gradually validated Razavi's assessment of the AI industry landscape.

In 2023, Anthropic launched Claude 2 and further opened its API services, beginning to explore the enterprise market; same year, Amazon announced a $4 billion investment in Anthropic, integrating Claude as a key component of AWS AI services; Google also continued to increase its investment and established a cloud computing partnership with Anthropic.

Continuous investment from capital and industry giants has provided Anthropic with the funding and computing power needed to train cutting-edge models.

Meanwhile, Anthropic has also taken a different commercial path from OpenAI.

If ChatGPT first demonstrated that large models could become consumer-grade products for the general public, then Anthropic is attempting to prove that AI can also become infrastructure within enterprise workflows.

Claude is increasingly being used in enterprise scenarios such as software development, knowledge management, and research analysis. Anthropic states that, as of October 2025, over 300,000 enterprise customers are using Claude; the number of large customers generating over $100,000 in annual revenue has grown nearly sevenfold over the past year.

This shift is changing how the market views Anthropic: it is now seen as a foundational model company establishing its position in the enterprise AI market, rather than a follower of OpenAI.

As Claude grows in the enterprise market, Anthropic's commercial value has been revalued by capital markets.

In May 2026, Anthropic completed a $65 billion Series H funding round, raising its post-money valuation to $965 billion, surpassing OpenAI’s approximate $852 billion valuation at the time and becoming one of the world’s most highly valued native AI companies.

According to The Wall Street Journal, based on Anthropic’s latest valuation, Spark Capital’s stake in Anthropic is worth approximately $7 billion—up from Anthropic’s valuation of about $4 billion when Spark made its investment in 2023.

A $75 million investment grew in value to nearly 100 times its original amount.

This is not simply a story of investment return. More importantly, Razavi correctly identified Anthropic’s position within the AI industry.

Spark Capital is not the only institution betting on Anthropic.

As Claude gradually becomes one of the leading foundational models outside of OpenAI, an increasing number of investors and industry giants are joining this AI gamble.

Their investment logic is not entirely the same: for venture capital firms, the key is assessing whether Anthropic has the potential to evolve from a technology company into a major platform in the AI era; for tech giants, investing in Anthropic simultaneously involves competition in computing power, cloud services, and the AI ecosystem.

The commonality is that they all invested in Anthropic before it had fully proven its commercial value, betting on its potential to become a major player in AI.

Most notably, another group of early venture capital firms—besides Spark Capital, Menlo Ventures was also one of the biggest winners in Anthropic’s rise, and its story is even more complex.

According to Business Insider, Menlo partner Matt Murphy was drawn to the team and technical direction after a brief conversation with Anthropic CEO Dario Amodei, but internal hesitation remains.

Ultimately, Menlo participated in Anthropic’s Series C round but chose not to lead it, while its competitor Spark Capital seized the opportunity and led the round.

Years later, Murphy admitted it was a regrettable decision: "I wish we had broken some rules during our Series C."

It turns out that Spark was right.

But Menlo didn’t miss Anthropic. A year later, during Anthropic’s Series D funding round, Menlo chose to increase its bet, leading the round and writing the largest check in the company’s history—even going so far as to raise additional capital through a special purpose vehicle (SPV) to expand its investment, due to limits on single-investment amounts.

Murphy believes that even if a market already has a clear leader, a highly valuable challenger can still emerge.

Today, this investment has become one of Menlo’s most significant bets in its history. As Anthropic’s valuation continues to rise, Menlo’s stake in Anthropic is now valued at approximately $14 billion.

For this type of VC, the return logic from Anthropic is similar to investing in companies like JD.com or Meituan during the internet era:

Acquire a portion of equity before the company becomes an industry giant. If your judgment is correct, early capital can achieve returns far exceeding market averages as the company grows.

For tech giants like Amazon and Google, investing in Anthropic carries even greater complexity—they are not merely betting on the growth of a single AI company, but also positioning themselves ahead in the competition for AI infrastructure.

In September 2023, Amazon announced an investment of $1.25 billion in Anthropic, acquiring approximately 10% of Anthropic’s equity; in March 2024, Amazon further increased its investment by $2.75 billion, bringing its total committed investment in Anthropic to $4 billion.

Meanwhile, Anthropic has selected AWS as one of its primary cloud platforms and uses Amazon-provided computing infrastructure to train and deploy the Claude models.

For Amazon, this is a dual investment. On one hand, it holds shares in Anthropic and can benefit from the growth of this AI company; on the other hand, by partnering with Anthropic, AWS gains a key customer and application ecosystem for the AI era.

The relationship between Google and Anthropic is more nuanced, as Google is itself one of the most important competitors in the AI field.

From DeepMind to Gemini, Google has made long-term investments in foundational model development and aims to maintain its technological leadership in the AI era. At the same time, Google is also investing in Anthropic and has become a key cloud computing and computing power partner for it.

In early 2023, Google announced an investment of approximately $300 million in Anthropic, acquiring about 10% equity. In November of the same year, the two parties expanded their collaboration, with Google committing additional investment to bring the total cumulative investment to $2 billion.

Meanwhile, the two parties are expanding their cloud computing collaboration: Anthropic will use Google Cloud infrastructure to train and deploy models, and will offer Claude services to enterprise customers through Google Cloud’s Vertex AI platform.

In the same month of November, Anthropic began using Google’s latest-generation TPU v5e accelerators for AI inference, while continuing to utilize related Google Cloud services. By 2025, Anthropic announced plans to expand its use of Google Cloud TPUs, aiming to connect up to one million TPUs to provide computing power for future Claude model training and deployment.

In other words, Google and Anthropic share a triple relationship: they are investor and investee, competitors, and key partners.

In the past, tech giants often built competitive advantages through in-house development or acquisitions; but in the era of large models, a more complex relationship has emerged among capital, computing power, and model capabilities.

For Amazon and Google, investing in Anthropic is both a financial investment and a strategic move to secure a position in the AI ecosystem amid the strong AI alliance between OpenAI and Microsoft.

Additionally, among Anthropic’s early investors is a particularly unique wealth story: the legacy of FTX.

In 2021, funds under FTX invested approximately $500 million in Anthropic.

At that time, the most popular narrative in the capital markets was still cryptocurrency. FTX was expanding rapidly, and Sam Bankman-Fried (SBF) had become one of the most prominent figures in the crypto industry, whereas AI foundation models had not yet become the focal point of global capital pursuit.

No one expected that years later, the situation would change dramatically,

At the end of 2022, FTX's empire collapsed and entered bankruptcy proceedings, while Anthropic rose rapidly amid the wave of generative AI.

This portion of Anthropic shares from FTX’s bankruptcy estate has also become one of the most valuable assets, evolving from an early venture investment.

According to court documents, FTX and its affiliate Alameda Research previously invested approximately $500 million and held about 7.84% of Anthropic. In 2024, when FTX’s bankruptcy estate sold these shares, the transaction was valued at approximately $1.3 billion.

This investment became one of the few assets in FTX’s bankruptcy estate to experience significant appreciation, even carrying a sense of generational shift: an investment in AI made by a company that rose during the cryptocurrency bubble and ultimately collapsed, yet gained tremendous value due to the next wave of technological advancement.

Of course, not all investors receive the same returns; the earlier you enter, the greater the potential return, but also the higher the uncertainty.

In 2023, when Spark invested in Anthropic, the company was valued at approximately $4 billion; later investors acquired shares at higher prices after Anthropic had demonstrated some of its value.

This is also the core logic of venture capital: the greatest returns often come from those willing to bet on a future that is still uncertain.

Anthropic's story therefore becomes special.

It not only created an AI company with a valuation nearing $1 trillion, but also generated wealth for a group of people around this company.

For Anthropic's investors, their story has not truly ended yet.

Because the most enticing numbers in venture capital are often the most misunderstood.

Book value.

It looks like wealth has been created, but there’s still a hurdle to overcome before it can be truly realized.

Based on Anthropic's latest valuation, Spark Capital's stake is worth approximately $7 billion, but this does not mean Spark has received $7 billion in cash.

In the private market, company valuation represents only the equity value calculated by investors based on the latest financing price; actual gains can only be realized upon an IPO, secondary market trading, or other exit opportunities.

In other words, the $7 billion represents the current market valuation of Anthropic’s shares, not the realized profit for Spark Capital.

And this is the question all AI wealth stories must confront: Can the high valuations bestowed by capital markets ultimately be supported by real business value?

The logic of venture capital is to bet on the future in advance, but when valuations approach the trillion-dollar mark, market bets are no longer just on company growth, but on the pace of the entire industry’s future development.

Over the past year and a half, the AI industry has experienced an unprecedented capital boom, with investors willing to believe that companies with advanced model capabilities will become the next-generation technology infrastructure. Anthropic is precisely the most representative example of this wave.

At the same time, the market has begun to reevaluate the growth rationale of the AI industry.

As mentioned in our previous article, Over the past few months, global AI-related assets have experienced significant volatility. Some AI-themed stocks that were previously favored by capital have pulled back after reaching highs, as investors shift their focus from “how much potential AI holds” to “how much real value AI can actually create.”

This does not mean the AI wave has ended; on the contrary, every technological revolution in history has gone through a similar process.

Capital once believed the internet would create numerous new giants, but in the end, only a few platforms truly built sustainable business models survived; the same held true for the mobile internet era—smartphones, app ecosystems, and mobile payments gave rise to new industry leaders, yet many companies reliant solely on traffic and funding for growth ultimately disappeared.

Even in the AI era, it will be no exception.

Of course, Anthropic's high valuation is not entirely based on speculation. Previously, Anthropic's internal financial projections disclosed to investors showed that the company expects to achieve operational profitability for the first time in the second quarter of 2026: quarterly revenue is projected to reach approximately $10.9 billion, a significant increase from $4.8 billion in the first quarter, with an estimated operational profit of about $559 million.

External institutions are even more optimistic: SemiAnalysis, based on Anthropic’s revenue growth, enterprise demand, and the commercialization progress of Claude, estimates that Anthropic’s operating profit in the third quarter could exceed $1 billion.

If these predictions materialize, Anthropic will become one of the few AI companies capable of achieving profitability at the operational level while continuously investing in cutting-edge model development.

For this reason, Anthropic stands out as a particularly unique case in discussions about the AI bubble: it represents both the high expectations capital has for the future and the commercial capabilities that only a few AI companies have demonstrated so far.

Anthropic has the capability, but whether its capabilities can justify the price the market has assigned to it remains to be seen.

For Anthropic investors, the real validation may be coming soon.

In June this year, Anthropic secretly filed its IPO documents with the SEC; in July, CNBC reported that the investment banks handling the listing were arranging meetings between Anthropic’s management and investors in preparation for a potential IPO. Reports suggest the company could list on the capital markets as early as this fall, with market rumors indicating an IPO could launch as soon as September, while other reports point to a target window in October.

If the IPO proceeds smoothly, Spark Capital's $7 billion book value will be tested by the public market.

This article is from the WeChat public account "Letter AI," authored by Yuan Xinyue and edited by Wang Jing.