In the year In 2020, Chinese startup Ziliz — which builds cloud-native software for AI applications and unstructured data analytics and is the creator of Milvas, a popular open-source vector database for similar searches — raised $43 million to grow and prepare its business. The company is about to relocate to the U.S. After nearly two years, today Zilize is announcing an additional $60 million in funding as it finally moves west, with a new headquarters in San Francisco to boost AI applications with an ever-expanding supply of unstructured data to drive demand for more efficient processing techniques. it is.
Led by Prosperity7 Ventures – a $1 billion venture fund created by Saudi oil giant Aramco (the name is a reference to the first commercial well to strike oil in the country) – the round includes previous Chinese backers Temasek Pavilion Capital, Hillhouse Capital, 5Y Capital and Yunqi Capital. The company isn’t disclosing its valuation, but it’s worth noting that this latest injection is being described as an extension to its $43 million Series B rather than a new round. We’ll update this if we learn more. The total revenue collected by the company has now reached $113 million.
The capital and relocation is not only a key time for the company, but also speaks to broader trends impacting the machine learning environment and Chinese-based startups.
First of these, Milvus, a Ziliz Spark product, has gone open source. The company says downloads have now passed the 1 million mark, up from 300,000 a year ago, while product users have grown by 300% over the same period – but it hasn’t disclosed its active user numbers. Clients include the likes of eBay, Tencent, Walmart, Ikea, Intuit and Compass.
In the year As we pointed out back in 2020, Milvus relied heavily on advertising and marketing spend, instead opting to use word of mouth on the places developers love to get ideas and inspiration, like GitHub and Reddit. That strategy worked: “Stargazers” on GitHub grew 200% to over 11,000, while the number of contributors doubled. (For comparison, it was starred about 4,400 times in 2020.)
The reason for the interest in Milvus – and then the Zilliz roadmap, based on the creation of additional products, most recently the service managed by the Zilliz Cloud, now in private preview; And Towhee, another open source framework, this is due to the increasing interest in vector databases for vector data ETL – how they can be used in AI applications.
Simply put, while information can (and often does) be processed by traditional databases, the complexity and structure of activities such as anomaly detection, recommendation, ranking, and other AI-driven tasks provide vector databases designed naturally and efficiently. How to represent AI data. (Ziliz says its vector database is “both cloud-native and capable of processing billions of vector data in milliseconds.”)
“Ziliz’s journey to this point began with the creation of Milvus, an open source vector database, and eventually joined the LF AI & Data Foundation as a top-level project,” Ziliz founder and CEO Charles Xie said in a statement. “Milves is now the world’s most popular open source vector database with over a thousand users. As a core contributor to Milvus, we continue to deliver on our commitment to delivering a fully managed vector database service on the public cloud with the security, reliability, ease of use and affordability that enterprises demand.
Others like Pinecone and Weaviate (competitors of Ziliz) are also building solutions to address this. Pincon raised money earlier this year and has some impressive names including Tiger Global and Menlo Ventures (for comparison, Pitchbook values Pincon at $168 million). Viviet’s parent, Semii Technologies, is based out of the Netherlands and has grown this year with the support of Friends of NEA.
Meanwhile, big cloud providers like AWS also have their own solutions. All of that points to a market opportunity that Ziliz is focused on addressing.
In the year It is interesting to note that in 2020, Ziliz said that more than half of the users of Milvis are outside of China. That speaks to where the company has positioned itself for the long term and where it sees its long-term growth. Xie – who goes by the nickname “Starlord” (yes) and previously worked as a software engineer at Oracle in the US before moving to start Ziliz in Shanghai – told TC that he envisioned the startup “going global from day one.” But he saw an opportunity to build in Shanghai first, because he could hire engineers, as well as the size of the Chinese market and the unstructured data at his disposal.
Of course, these days, many startups in the country are looking to move elsewhere to gain more freedom in how they grow their business, and to work with a wider set of clients. .
Prosperity7 has been playing a role in facilitating this migration. The fund only entered China last year and is actively hunting startups with international ambitions that can tap into the company’s extensive international network. We recently covered two investments, Jaka, a Beijing and Shanghai-based collaborative robotics startup, and Insilico, an AI medicine platform from Hong Kong.
Prosperity7’s investment in Ziliz appears to be in line with the investor’s mandate. Nowadays, it is not uncommon to see Chinese SaaS companies going international. Many of them were started by Chinese entrepreneurs with an international background. As the B2C space became saturated, they spent a few years testing the market in-house and growing from VCs who were very interested in B2B projects. But many find it difficult to monetize in China, where owners of small and medium-sized enterprises are still reluctant to pay for software subscriptions compared to their Western corporate counterparts.
“With his leadership at Milvus, Zilize is a global leader in vector similarity searches for massive amounts of unstructured data,” said Aisar Tayeb, CEO of Prosperity7 Ventures, in a statement. “We believe the company is in a strong position to build a cloud platform around Milvis, as data analytics platforms like Databricks and Snowflake have done with structured data. There is more than 4x more unstructured data than structured data, a gap that will continue to grow as AI, robotics, IoT and other technologies meld the digital and physical environments.”