28.8 billion! A top female student from Fudan University founded a unicorn company in three years.

If we were to name the place where unicorns are born the fastest, it would probably be the AI startup circle.

Recently, foreign media reported that Fireworks AI, an emerging cloud service provider, is seeking a new round of financing at a valuation of $4 billion (approximately 28.8 billion yuan), including the financing amount.

Well - known venture capital firms such as Lightspeed Venture Partners and Index Ventures are having in - depth discussions about leading the investment. Once the financing is successful, the founder of Fireworks AI will create a miracle of a 7 - year valuation growth in 1 year and build a unicorn worth 28.8 billion yuan in 3 years.

Chinese - American female entrepreneur achieves annual revenue of over $200 million

This is the story of a Chinese - American woman starting a business in California.

In 2022, Lin Qiao, a former senior engineering director at Meta, and six co - founders established Fireworks AI in Redwood City, California. Lin Qiao, the co - founder and CEO, graduated from Fudan University with a bachelor's and master's degree in computer science and later obtained a doctorate in computer science from the University of California, Santa Barbara.

Before starting her business, Lin Qiao had 24 years of industry experience and was a senior technology expert.

Her first job was a research position at IBM, focusing on data infrastructure and database technology. Later, she served as a technology supervisor at LinkedIn. During her time at Meta, Lin Qiao's greatest achievement was leading the development and deployment of PyTorch on Meta's data centers, mobile devices, and AR/VR devices to support Meta's applications.

In addition, she built an AI framework and a developer platform, and promoted innovation through a powerful ecosystem and platform specifically built for PyTorch to improve the speed of model exploration and production time. To date, PyTorch has become one of the most mainstream open - source machine - learning frameworks globally.

In Lin Qiao's view, the success of PyTorch stems from its simplicity for researchers, which extends to the production environment. With this experience, Lin Qiao and her team founded Fireworks AI, building the system from scratch and starting at the bottom.

"We have to think about how to efficiently acquire data, how to perform distributed inference efficiently, and how to conduct large - scale training efficiently," said Lin Qiao. What Fireworks AI aims to do is to compress the training and inference time, enabling more people outside of ultra - large enterprises to use general artificial intelligence (GenAI).

Finally, Lin Qiao and her team integrated the capabilities of small, customizable models and the capabilities of open - source and closed - source artificial intelligence to provide developers with a simple, high - performance, and low - cost API. "This approach allows customers to focus on innovation and product design instead of struggling with complex technical issues, enabling them to deploy AI models efficiently in a short period."

With the two keywords of higher performance and lower cost, Fireworks AI has over 100 models and has achieved an annual revenue of over $200 million, which means a monthly revenue of nearly $17 million. The company told investors that it expects its annual revenue to reach $300 million by the end of this year.

As of March this year, its competitor Together had an annual revenue of $150 million, about $12.5 million per month, and a valuation of $3 billion.

Image source: Fireworks

3 - year valuation soars to 28.8 billion yuan, may compete with NVIDIA

Compared with traditional cloud service providers, Fireworks AI offers large models with lower cost, higher performance, and customizable open - source features. Under this differentiated approach, the company's valuation has risen rapidly.

Financing data shows that Fireworks AI, founded in 2022, completed its seed - round financing that year. In 2023, it completed its Series A financing, with a valuation reaching $552 million. Investors included top venture capitals such as Sequoia Capital and Benchmark, as well as industrial capitals such as NVIDIA, AMD, Databricks Ventures, and MongoDB Ventures.

Recently, the discussions with well - known venture capital firms such as Lightspeed Venture Partners and Index Ventures not only pushed its valuation towards $4 billion, making it a unicorn in the industry, but also increased the valuation of Fireworks AI by 7 times in 1 year.

Regarding NVIDIA's investment in the Series A round, Fireworks AI stated on its official website that it is not only financial support but also in - depth technical cooperation. Simply put, they are customers of each other.

Fireworks AI said that its platform can efficiently run and fine - tune open - source AI models such as DeepSeek and Llama, which perform well on NVIDIA GPUs. In addition, Fireworks AI also supports NVIDIA's NIM microservices, which offer a wide range of AI models covering multiple fields such as text generation, image processing, video processing, and 3D simulation.

Interestingly, NVIDIA may form a competitive relationship with Fireworks AI.

In an interview, Lin Qiao said that NVIDIA is interested in entering any profitable market and will trigger competition. But from an industry perspective, the market doesn't like monopolies. So, it's an economic issue. As for when this competition will come, Lin Qiao said it should be soon.

In fact, in terms of business expansion, NVIDIA has achieved business implementation in multiple scenarios through investment.

For example, in the field of artificial intelligence and cloud computing, CoreWeave, in which NVIDIA invested in 2023, has become a direct cloud computing power provider for NVIDIA. NVIDIA has also established a partnership with Hugging Face, an invested company, to provide generative AI super - computing services for developers to help them build large - language models.

In the healthcare field, NVIDIA has invested in several companies to support AI drug R & D through computing power provision. In industries such as quantum computing, NVIDIA has cooperated with Google and Oracle to jointly develop quantum chips and test and develop hybrid quantum computing services for enterprise customers respectively.

Therefore, some analysts believe that it is worth paying attention to whether NVIDIA will acquire Fireworks AI in the future, given their cooperation and potential competitive relationship.

It's worth mentioning that the new investor Index Ventures recently obtained a high investment return from the acquisition of Scale AI, which makes startups like Fireworks AI promising in the eyes of established venture capital firms.

Chinese entrepreneurs become "money magnets" in the AI startup circle

The entrepreneurial success of Lin Qiao, a top student from Fudan University, is not accidental.

In the AI startup circle, Chinese entrepreneurs are very active, covering multiple aspects from algorithm innovation at the model level to vertical industry solutions at the application level. For example, Scale AI was founded in 2016 by Chinese - American teenagers Alexandr Wang and Lucy Guo, focusing on data annotation services, and has quickly grown into a unicorn enterprise with a valuation of $29 billion.

Furthermore, Fei - Fei Li, a professor at Stanford University, founded World Labs, a "spatial intelligence" company, focusing on enabling AI to have advanced visual reasoning capabilities. The video - generation AI company Pika, co - founded by Wenjing Guo and Chenlin Meng, has completed a Series B financing of $80 million.

In terms of fundraising ability, Chinese entrepreneurs are also very strong. According to incomplete statistics, the total valuation of Chinese - founded AI companies is at least over $30 billion (approximately 210 billion yuan), and unicorns such as Scale AI, Cognition AI, Imbue, and Cresta have emerged.

Besides Scale AI, Cognition AI was created by Chinese - American youth Scott Wu. Its AI programmer Devin received financing a few weeks after its launch, and the parent company's valuation increased 6 times to $2 billion within half a year. To date, Cognition AI has completed 3 rounds of financing, and the latest round was led by 8VC, with a valuation of $4 billion.

The latest news shows that Cognition AI is in talks for a financing round of over $300 million, which will increase the company's valuation to $10 billion.

In addition, Imbue has received a total financing of $220 million, with a valuation of over $1 billion. Cresta, co - founded by Chinese entrepreneurs, has also raised over $270 million, with a valuation of $1.6 billion.

Regarding the entrepreneurial success of Chinese entrepreneurs in the AI field, some people believe that they generally have top - notch educational backgrounds and rich industry experience.

Take Alexandr Wang, the founder of Scale AI, as an example. He was an outstanding student at the Massachusetts Institute of Technology. Wenjing Guo, one of the founders of Pika, studied at Harvard and Stanford. The Chinese entrepreneurs behind Cognition AI, Imbue, and Cresta also have excellent educational backgrounds.

"These entrepreneurs not only have profound technical skills but also have a keen insight into the market. For example, Alexandr Wang, with his deep technical accumulation in the field of machine learning, keenly discovered the key role of data annotation in the development of artificial intelligence, thus creating an efficient and high - quality data annotation service model that has promoted the rapid development of the AI industry," an investor commented.

More importantly, the perseverance and innovative spirit shown by Chinese entrepreneurs during the entrepreneurial process are admirable. Especially in the face of many challenges such as technical difficulties, market competition, and financial pressure, they always maintain a persistent pursuit of their goals.

In short, in this era of "innovation and opportunity," Chinese entrepreneurs are creating miracles.

This article is from the WeChat official account "China Venture Capital". Author: Mei Chen. Republished by 36Kr with permission.