News
DealsCorvex agrees USD 103 million take-or-pay deal to perform AI inference with Mid-Atlantic GPUsSovereignNeysa scales AI infrastructure in India with AMD Instinct MI350X GPUs on TytanInferenceMicrosoft Foundry adds Baseten, DeepSeek AI inference (MSFT:NASDAQ)ClustersDataOne and Rowan University sign engineering co-op program agreement for DataOne's Vineland AI factoryClustersLook inside Cal Poly's $3 million AI Factory (Video)DealsOpenAI Taps AMD Chips for 457 MW of AI Capacity in Six-Year TensorWave DealClustersCal Poly launches an AI data center on campus. Take a look insideInferenceArista takes aim at the next AI network battleground: distributed inferenceDealsOpenAI Signs 457 MW AI Data Center Deal as AMD Expands Its Challenge to NvidiaClustersBehind Dell AI Factory鈥檚 NVIDIA & AMD Enterprise ExpansionDealsCorvex agrees USD 103 million take-or-pay deal to perform AI inference with Mid-Atlantic GPUsSovereignNeysa scales AI infrastructure in India with AMD Instinct MI350X GPUs on TytanInferenceMicrosoft Foundry adds Baseten, DeepSeek AI inference (MSFT:NASDAQ)ClustersDataOne and Rowan University sign engineering co-op program agreement for DataOne's Vineland AI factoryClustersLook inside Cal Poly's $3 million AI Factory (Video)DealsOpenAI Taps AMD Chips for 457 MW of AI Capacity in Six-Year TensorWave DealClustersCal Poly launches an AI data center on campus. Take a look insideInferenceArista takes aim at the next AI network battleground: distributed inferenceDealsOpenAI Signs 457 MW AI Data Center Deal as AMD Expands Its Challenge to NvidiaClustersBehind Dell AI Factory鈥檚 NVIDIA & AMD Enterprise Expansion
EveryNeocloud
Fal

Fal

Serverless GPU platform for generative AI inference

Fal is a developer platform founded in 2021 that provides fast inference APIs for generative media. Headquartered in San Francisco, it offers serverless GPU hosting and model endpoints for image, video, and audio generation.

Share on WhatsApp

  • Founded in 2021 by Burkay Gur and Gorkem Yurtseven
  • Offers serverless GPU hosting and model APIs
  • Headquartered in San Francisco with global team
  • Supports image, video, audio and 3D generation
  • Billing per output for APIs, per second for serverless

What it is

Fal operates as a generative media platform designed for developers. It provides application programming interfaces for creating images, videos, audio, and three-dimensional content. The service also offers serverless graphics processing unit hosting for custom models. Users access hundreds of models through a single key and client software development kit. Each model functions as an endpoint identifier, allowing developers to fetch live schemas and pricing data directly. This structure ensures that code snippets remain accurate to the current platform state. The company aims to reduce barriers to creative expression by optimising inference speeds and costs.

How it came about

The company was established in 2021 by Burkay Gur and Gorkem Yurtseven. Both founders previously worked at Coinbase and Amazon, where they observed challenges in artificial intelligence infrastructure. Their shared passion for technology led them to address issues such as slow inference speeds and high operational costs. Since its inception, the firm has developed rapid inference capabilities for models like Stable Diffusion XL and Whisper. It claims to support hundreds of millions of customers, demonstrating significant scale. The team is headquartered in San Francisco but includes global talent, with a preference for in-person work in downtown offices alongside remote options.

The compute

The platform runs on graphics processing unit infrastructure that supports autoscaling from zero to thousands of runners. It specialises in inference for generative models, focusing on speed and efficiency. Developers can deploy their own models, pipelines, or hypertext transfer protocol servers using the serverless offering. Billing for these custom deployments occurs per second of runner time, based on machine type. For standard model APIs, charges are applied per output generated. The system handles long-running tasks through a queue mechanism, supporting polling and webhooks. Some models allow streaming of partial results or real-time interaction via web sockets.

Where the GPUs are

Fal is headquartered in San Francisco, California. The company states that its team spans the globe, comprising international talent. While it prefers employees to work in person at its downtown location, it provides remote opportunities for specific candidates. The technical infrastructure relies on cloud-based graphics processing units, though specific data centre locations are not detailed in the provided material. The platform serves users globally through its application programming interface endpoints.

Customers and deals

The platform states it supports hundreds of millions of customers. It empowers developers to build creativity tools and scalable applications. By providing fast inference, it enables responsive and immersive experiences for end users. The service is used by those creating visual content, including images and videos. Developers integrate the platform using client libraries for languages such as JavaScript, Python, Swift, and Java. The company advocates for open source sharing to benefit the wider technical community.

Money and ownership

Fal is backed by investors and angels from Silicon Valley. Specific funding amounts or valuation figures are not disclosed in the available information. The business model involves prepaid credits for model API usage, billed per output. Serverless deployments are charged per second of compute time. The company does not charge for server errors or time spent waiting in queues. Pricing details for each model are available through machine-readable text files associated with each endpoint.

Its place in AI

Fal positions itself at the forefront of artificial intelligence inference, particularly for generative media. It aims to amplify human creativity by making powerful models accessible and cost-effective. The company views generative artificial intelligence for media as a key future domain. By optimising performance, it helps developers navigate graphics processing unit shortages. Its mission is to serve as a catalyst for a new generation of creators, similar to how previous platforms enabled musicians and video makers.

Questions people ask about Fal

What services does fal provide?

Fal offers a generative media platform for developers. It provides fast application programming interfaces for image, video, audio, and three-dimensional generation. It also offers serverless graphics processing unit hosting for deploying custom models.

How is billing structured on fal?

Model API usage is billed per output generated, using prepaid credits. Serverless deployments are charged per second of runner time based on the machine type. Users are not charged for server errors or queue wait times.

Who founded fal and when?

The company was founded in 2021 by Burkay Gur and Gorkem Yurtseven. They previously worked at Coinbase and Amazon, aiming to solve infrastructure challenges they observed in those roles.

Where is fal based?

Fal is headquartered in San Francisco, United States. The team works primarily in person at its downtown office but also offers remote work opportunities for exceptional candidates globally.

Good reading and sources

    Where this page comes from

    There is no English Wikipedia article on Fal yet, so the facts on this page come from the sources listed above: fal.ai/about, fal.ai/. Some are the subject's own pages, so its claims about itself are reported as its claims. EveryNeocloud wrote the text in its own words, with AI assistance, and compares it with the sources to make sure no passage is copied.

    It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.

    Photo credits on this page

    The small pictures above come from Wikimedia Commons under free licences. Some are cropped to fit.

    The weekly digest

    Every Monday: the provider of the day, new pages on EveryNeocloud, and the conferences and anniversaries of the week ahead.