The AI landscape is witnessing a remarkable surge. According to a Bloomberg report, AI inference platform Fireworks AI is in discussions for a new funding round that would value the company at $150 billion (approximately 1 trillion yuan). This marks the emergence of another AI unicorn valued in the trillions.

This milestone was achieved in just four years. Fireworks AI was founded in 2022, providing a platform for businesses to run open-source AI models with a pay-per-token model. The company now processes 30 trillion tokens daily, fueling its rapid growth through token sales.

The driving force behind this success is a familiar figure: Lin Qiao. A graduate of Fudan University with both bachelor's and master's degrees in computer science, she began her journey into the AI world there. Riding the wave of the AI boom, she has led the team to build a company with a valuation in the trillions.

The story of Fireworks AI begins with Qiao. After earning her Ph.D. in computer science from the University of California, Santa Barbara, she held significant technical roles at IBM, LinkedIn, and Meta. During her seven years at Meta, she managed a team that grew from 5 to over 300 people. She co-founded the well-known AI frameworks Caffe2 and PyTorch and led the comprehensive deployment of PyTorch across data centers, mobile devices, and AR/VR devices. This experience laid the groundwork for her later founding of Fireworks AI.

In 2022, Qiao identified a market gap: while AI infrastructure was mature, many enterprises and developers lacked the means to deploy AI technology into production quickly, efficiently, and at low cost. She recognized the need for an "AI cloud" that would provide users with a convenient, reliable, and scalable deployment environment.

Thus, in October 2022, just before the launch of ChatGPT, she founded Fireworks AI with the mission to "spread the fire of AI to the world."

Qiao quickly assembled a "dream team" of co-founders. Five of the seven co-founders were senior engineers and experts she had worked with at Meta, and another came from Google's AI division. Besides Qiao, the team includes two other Chinese professionals: Benny Chen and Chenyu Zhao.

In retrospect, Qiao's timing was impeccable. The launch of ChatGPT at the end of 2022 ignited the generative AI market, and Fireworks AI quickly found itself at the forefront.

The company acts as a facilitator in the AI era. Simply put, Fireworks AI provides a high-performance AI cloud platform. On this platform, developers can easily access leading global open-source large language models like DeepSeek and Kimi without the need for costly self-built infrastructure and maintenance. Compared to traditional cloud service providers, Fireworks AI offers lower-cost, higher-performance, and customizable open-source large models.

The company's revenue comes from the vast number of tokens consumed by users on its platform.

With just a single line of code, users can deploy models in seconds. The platform supports models such as DeepSeek-V4-Pro, Kimi K2.6, MiniMax M2.7, GLM 5.1, Qwen3.6 Plus, and Gemma 4 31B IT, covering text, image, audio, and multimodal domains, and supporting model fine-tuning, reinforcement learning, and evaluation.

The burgeoning inference market has led to an exponential increase in token consumption. Recently, Qiao shared on social media that Fireworks AI currently processes approximately 30 trillion tokens per day. Clients include AI programming tool Cursor and major enterprises like Uber, Samsung, Notion, and Shopify.

Every token translates into real revenue. Data from research firm Sacra shows that as of February 2026, Fireworks AI's annualized revenue had reached $315 million (approximately 2.2 billion yuan), a year-on-year increase of 416%.

Looking back, Fireworks AI's rise has been astonishing. In its founding year, the company secured a $25 million seed round led by Benchmark. In 2023, it completed a Series A round with investors including Sequoia Capital, NVIDIA, and AMD.

Fireworks AI quickly attracted attention from Silicon Valley luminaries. Supporters include former Snowflake CEO Frank Slootman, former Meta COO Sheryl Sandberg, Airtable CEO Howie Liu, and Scale AI CEO Alexandr Wang.

In 2024, Fireworks AI completed a $52 million Series B round led by Sequoia Capital, with continued participation from NVIDIA and AMD, bringing the company's valuation to $552 million. NVIDIA CEO Jensen Huang praised the company highly, calling it "the TSMC of artificial intelligence."

By 2025, a Series C round arrived as scheduled, led by Lightspeed Venture Partners, Index Ventures, and Evantic, with Sequoia Capital continuing to participate, valuing the company at $4 billion.

The underlying logic behind Fireworks AI's formidable momentum is that as AI applications deepen, its value becomes more apparent. "Every enterprise differs in its business model, goals, and operations. Fine-tuning becomes a crucial link in leveraging AI, and this is where Fireworks AI comes into play," Qiao once explained.

With substantial funding in hand, Fireworks AI began expanding its footprint. In March of this year, the company announced the acquisition of game server hosting company Hathora. Hathora's core technology in gaming scenarios—low-latency, large-scale distributed computing orchestration capable of supporting millions of simultaneous online players—is crucial for AI model inference at scale.

Now, a new funding round for Fireworks AI is imminent. According to Bloomberg, the company is in talks for a new round that would value it at $150 billion (approximately 1 trillion yuan).

"I've worked in the tech industry for over 20 years, witnessing wave after wave of industry-level transformations—from cloud computing to mobile applications—but this AI transformation is a complete tectonic shift," Qiao once remarked on her entrepreneurial motivation. Now, riding the AI wave, she has cultivated a trillion-yuan unicorn.

The AI Boom: An Outbreak of Trillion-Yuan Companies

Globally, the valuation ceilings for AI companies are being shattered at an unprecedented pace.

Just this week, the global AI community witnessed a valuation feast for the history books: Anthropic announced the completion of a $65 billion Series H funding round, with a post-money valuation soaring to $965 billion (approximately 6.5 trillion yuan).

Three months prior, during its Series G round, Anthropic's valuation was $380 billion. In just three months, its valuation surged by $600 billion. Extending the timeline, it took Anthropic only five years from its founding to become the world's highest-valued AI unicorn.

This spectacle is breathtaking.

In response to Anthropic, OpenAI is mounting a counteroffensive. According to foreign media reports, OpenAI is intensively advancing its IPO process, planning to complete its listing by September this year with a target valuation exceeding $1 trillion.

Across the global AI landscape, well-known AI unicorns like Databricks and Surge AI have long surpassed the trillion-yuan valuation threshold.

Turning our focus to China, more than half of the original "Six Tigers of Large Models" have already achieved valuations or market capitalizations of 1 trillion yuan. Among them, Zhipu AI's market cap once approached 900 billion Hong Kong dollars yesterday, surpassing the combined market cap of Meituan and JD.com.

The spectacle continues: MiniMax's Hong Kong market cap exceeds 250 billion Hong Kong dollars while it also initiates listing guidance for the A-share market. DeepSeek's pre-money valuation for its first round of financing was quoted at $45 billion (approximately 300 billion yuan). Kimi's latest valuation exceeds $20 billion. StepFun is sprinting towards a trillion-yuan IPO... One cannot help but marvel: AI is mass-producing trillion-yuan companies.

Discussing the current AI valuation feast, a16z co-founder Marc Andreessen offered a historically profound interpretation: "This is the biggest technological revolution of my lifetime. In terms of magnitude, it's clearly larger than the internet, comparable to things like the microprocessor, the steam engine, and electricity."

He pointed out that the customer revenue growth rates being achieved by this new generation of AI companies have already surpassed any previous wave he has witnessed.

Therefore, even though AI has not yet created traditional "hundred-billion-dollar products," the world has already given birth to many companies with valuations in the trillions and even tens of trillions—this financial miracle is rewriting the rules of the tech industry.

Whether this represents a new species or a super bubble, only time will tell.