The growing demand for AI has given rise to a new category of digital services companies that sell computing power, access to models and infrastructure to developers. INTER drivers of this package IS Fireworks AIco-founded by ex Meta ExECUTIVE Lin Qiaowho led the creation of PyTorch, a popular open source machine learning framework, and a team of engineers from Meta and Google.
Fireworks AI is a platform for developers to build products faster and at lower cost than with proprietary models, using open source models. It has access to many of the most capable open source models on the market, such as Meta’s Llama series, Mistral, Qwen and DeepSeek. It also allows enterprises to upload their own data to train and adjust these models. Her clients include cursor, Harvey, Uber AND Shopifyamong others.
Lin describes Fireworks as “a specialized intelligence platform,” as opposed to general intelligence. Specialized intelligence was what AI researchers primarily relied on before general intelligence became viable. “Before generative artificial intelligence was a thing, there was no foundational model holding the world’s knowledge together. GenAI changed that,” Lin explained to the Observer. “Now, the foundational models learn from the public internet and massive labeled datasets, creating a deeper and more generalizable knowledge base that you can use directly as a black box API.”
But Lin believes that, amid the abundance of public data and general intelligence models, the most valuable uses of AI will come, counterintuitively, from specialization.
“Because the underlying models don’t have access to private data locked within applications and enterprises,” she said. “Most data is private, locked away within enterprises like proprietary IP and information that would never be shared outside of the company.”
Training and adjusting models with that private data creates an ongoing need for Fireworks services. “This is an ongoing process because applications continue to evolve, data distribution changes and the underlying models continue to improve,” Lin said. “We have customers who tune in once a week, once a day, or even once every few hours.” She predicted that this tuning process would soon be fully automated.
Once a model is fine-tuned, Fireworks helps optimize it for inference speed and cost. The company offers some of the fastest inference — the speed at which an AI generates an answer — in the industry. For example, the Cursor code editor uses Fireworks’ speculative decoding to provide code suggestions up to 13 times faster than traditional configurations.
Fireworks processes more than 30 trillion tokens in daily inference traffic (excluding training), more than OpenAI and Google’s Gemini, according to the latest data released.
The company makes money by charging users a flat fee per million tokens. Tokens are the basic unit of data that an AI reads, processes and generates; in English, a token is roughly four characters, or about three-quarters of a word.
“We offer a platform that covers the entire end-to-end spectrum of model development, from quality to speed and cost. The end result is that our customer gets better quality, much faster speed and five to ten times lower cost, allowing them to go to mass production quickly,” said Lin.
The new gap
These days, AI executives like to talk about “the gap,” or a competitive advantage that allows a company to stay ahead of the competition. At a time when it’s easier than ever to turn an idea into an app thanks to AI coding tools, the traditional product gap is disappearing.
“The data is moat because it can’t be duplicated,” Lin said. “The data collected to understand user intent, user preferences, and user engagement—what’s working well, what’s not working well, and where you need to optimize—is all your proprietary information, and that creates the asymmetry needed to compete. Anyone who can turn that data into their own proprietary intelligence can build on that. And it can get complicated.”
Fireworks competes with both closed model providers (such as OpenAI, Anthropogenic and Google) and infrastructure platforms like Together AIReplicate and AWS Bedrock. Its differentiation lies in its focus on open models while tightly integrating training, fine-tuning, and high-performance inference into a single system.
“We don’t need a Ferrari for grocery shopping.”
Besides the data gap, another argument for open models is unit economics. By allowing developers to choose from a wide range of open-weight models, platforms like Fireworks can match each task with the cheapest level of intelligence. This flexibility is increasingly important as companies look to deploy AI at scale. Using a single boundary model for each task quickly becomes prohibitively expensive.
“We don’t need to drive a Ferrari to go grocery shopping,” Lin said. “There are so many tasks we solve every day at varying levels of complexity. Some are extremely difficult, requiring intelligence beyond the human level to solve. Others are not so difficult. If you use a vendor that can help you automatically choose the best fit model for solving a given task, you get the quality you need at the lowest cost.”
When Lin founded Fireworks two years ago, the company initially focused on inference, treating it as “one size fits one.” Now, even training is doubling, driven by the rapid improvement and release cadence of open models. Open model quality has significantly narrowed the gap with closed models, while release cycles have accelerated from monthly to weekly. New models often top benchmarks and approach borderline performance.
“This makes training particularly attractive. With your private data and a little tuning, you can stay on top,” said Lin.
She went on to conclude, “We believe that specialized and generalized intelligence will coexist, but the world will not be dominated by a few generalized models. There will be millions of specialized models of intelligence—one for each use case.”
