AI is one of the most powerful technologies of our generation, but in order for it to reach its fullest potential, it needs to be developed in the open. Limited training resources, rising costs and access to compute, and monopolization by a handful of tech companies, has resulted in second order effects that hinder progress.

We are here to offer a new solution for AI development and contribution. The team has built a decentralized platform on Base that makes AI accessible to everyone, and rewards users for their contributions. This includes providing compute to the network, building applications with inference powered by our Developer Platform, and using our ecosystem consumer applications.

AI has limitless scalability and is more performant with global participation. This should be the standard regardless of which entity is fighting for market share.

Our belief revolves around making AI a collaborative technology where users are fairly compensated for their efforts. The Function team’s years of experience building in web3 helps us adapt in a rapidly-evolving industry. This is our story.

Function’s Early Days 

Prior to joining forces, software engineers Erick Ho and Alex Mo were each building in the crypto space, gaining firsthand experience with the challenges of scalability and decentralization. Erick began his career at AWS, where it provided him with a unique vantage point to witness the rising wave of artificial intelligence and the challenges associated with centralized AI systems. He witnessed firsthand the computational demands of AI and its limiting access to only a select major companies.

Erick crossed paths with Alex during their time at Base at the height of the NFT boom in 2021. Alex, who had joined Coinbase as a developer, quickly made his mark by creating an NFT index that enhanced Coinbase’s NFT marketplace. It wasn’t until a year later when Erick was brought on to help scale its operations. Their shared drive and determination created a partnership that would lay the groundwork and later become the foundation for Function Network. 

Challenges with AI

Artificial intelligence began to dominate headlines in early 2022 with companies like OpenAI, Google, and Microsoft quickly becoming household names. The advent of large language models (LLMs) like ChatGPT by OpenAI sparked a revolution in how the world interacted with artificial intelligence, while megacap Google and Microsoft raced to integrate AI into their existing products and services. 

As AI adoption accelerated, the duo saw the challenges presented from their time at AWS and Coinbase materialize. Inference costs were becoming increasingly expensive, as so few compute providers were able to become successful. What’s more, these centralized AI systems created fundamental issues: a lack of seamless AI integration into existing blockchain infrastructures, cost inflation, and censorship-driven outputs.

• AI and blockchain integration challenges: Centralized AI models rely on massive datasets stored in servers, making them more vulnerable to data breaches and cyberattacks that can compromise sensitive information globally. As well, centralized systems provide a siloed approach to AI development, raising concerns about transparency as there’s often no visibility to see how these datasets are trained.

• Cost inflation: The growing demand for AI is creating a stark divide in the tech landscape. Costs are spiraling out of control while major tech companies are dominating critical resources. These large cap behemoths are forecast to spend $325 billion in infrastructure this year alone, fueling a cloud compute market that’s already worth $675 billion in 2024. This massive influx of capital is driving up costs and creating a supply crunch that is stifling innovation and competition. It is also providing significant strain on centralized cloud providers, leading to increased latency. As demand outpaces supply, many businesses are feeling the pinch of being priced out of the AI race and find difficulty scaling their operations. 

• Censorship vulnerabilities: Power and control are concentrated in the hands of a few large corporations, making AI systems more susceptible to censorship by authorities or governing bodies. This can restrict access to information and cause manipulation of global narratives. It is possible for centralized systems to filter and remove information, potentially suppressing legitimate discourse and stifling freedom of speech. Users also often remain unaware of why specific content has been removed or altered, fuelling mistrust.

These challenges set Erick and Alex on a path to create a network that is more open, transparent, and inclusive for AI usage and development.

The Birth of Function Network

Function Network launched on Base as a platform for people to use AI, build with AI, and power AI while also being rewarded for it. The team has built a full-stack managed AI cloud where any type of user can access an array of open source LLMs, powered by a global infrastructure network. It is the first decentralized inference network that supports the next wave of AI powered applications in both web2 and web3.

• Provide Compute: At the core of Function’s full-stack decentralized AI ecosystem is Function Network, the foundational layer where anyone can contribute compute to run open source models. Unlike centralized platforms, Function’s decentralized structure distributes inference across different GPU and other hardware providers that not only lowers computing costs, but introduces fault tolerance to guarantee model uptime. Providers will be granted the right to offer compute and earn additional rewards by staking Function’s native token (FUNC). This serves as a reliable threat model that also enhances network security. 

• Provide Models: Function empowers anyone to contribute their own models. Model creators can upload, host, and share their open-source LLMs directly on Function. In return, they gain visibility, usage, and also rewards. Function levels the playing field for model creators, while making room for participation beyond the “Big 4” (Llama, Mistral, Qwen, and DeepSeek).

• Developer Platform: Function Network provides developers with an intuitive platform that delivers a one-stop shop for all AI development needs. It offers a suite of Open-AI compatible APIs (application programming interfaces) that helps streamline the integration of AI capabilities into other applications. This provides easy access to various open source models and allows Function to empower developers to create sophisticated AI-powered solutions, for free.

• Function Chat: Function Chat represents the user-facing application of Function’s infrastructure. It features a seamless, conversational AI experience that’s available on both the web and mobile app, while providing a familiar chat interface—much like what people use today with ChatGPT or Claude. By decentralizing a network of node operators, users can interact with a variety of AI models, reinforcing Function’s core belief in censorship-resistant technology.

Function’s decision to launch on Base is a natural fit, by leveraging Coinbase’s powerful distribution network and its success as the fastest growing L2. By building as an appchain L3, Function gains dedicated blockspace and more transaction throughput enabling the platform to scale efficiently. The move reinforces the team’s vision of decentralized, accessible infrastructure for AI innovation.

Join Function Network

Function Network is not just a platform, it’s a movement towards a more open, efficient, and accessible AI ecosystem. We invite businesses and developers to join on testnet and explore the endless possibilities of decentralized AI computing. 

The story of Function Network is far from over. In fact, it’s just the beginning. Start building here

AI is one of the most powerful technologies of our generation, but in order for it to reach its fullest potential, it needs to be developed in the open. Limited training resources, rising costs and access to compute, and monopolization by a handful of tech companies, has resulted in second order effects that hinder progress.

We are here to offer a new solution for AI development and contribution. The team has built a decentralized platform on Base that makes AI accessible to everyone, and rewards users for their contributions. This includes providing compute to the network, building applications with inference powered by our Developer Platform, and using our ecosystem consumer applications.

AI has limitless scalability and is more performant with global participation. This should be the standard regardless of which entity is fighting for market share.

Our belief revolves around making AI a collaborative technology where users are fairly compensated for their efforts. The Function team’s years of experience building in web3 helps us adapt in a rapidly-evolving industry. This is our story.

Function’s Early Days 

Prior to joining forces, software engineers Erick Ho and Alex Mo were each building in the crypto space, gaining firsthand experience with the challenges of scalability and decentralization. Erick began his career at AWS, where it provided him with a unique vantage point to witness the rising wave of artificial intelligence and the challenges associated with centralized AI systems. He witnessed firsthand the computational demands of AI and its limiting access to only a select major companies.

Erick crossed paths with Alex during their time at Base at the height of the NFT boom in 2021. Alex, who had joined Coinbase as a developer, quickly made his mark by creating an NFT index that enhanced Coinbase’s NFT marketplace. It wasn’t until a year later when Erick was brought on to help scale its operations. Their shared drive and determination created a partnership that would lay the groundwork and later become the foundation for Function Network. 

Challenges with AI

Artificial intelligence began to dominate headlines in early 2022 with companies like OpenAI, Google, and Microsoft quickly becoming household names. The advent of large language models (LLMs) like ChatGPT by OpenAI sparked a revolution in how the world interacted with artificial intelligence, while megacap Google and Microsoft raced to integrate AI into their existing products and services. 

As AI adoption accelerated, the duo saw the challenges presented from their time at AWS and Coinbase materialize. Inference costs were becoming increasingly expensive, as so few compute providers were able to become successful. What’s more, these centralized AI systems created fundamental issues: a lack of seamless AI integration into existing blockchain infrastructures, cost inflation, and censorship-driven outputs.

• AI and blockchain integration challenges: Centralized AI models rely on massive datasets stored in servers, making them more vulnerable to data breaches and cyberattacks that can compromise sensitive information globally. As well, centralized systems provide a siloed approach to AI development, raising concerns about transparency as there’s often no visibility to see how these datasets are trained.

• Cost inflation: The growing demand for AI is creating a stark divide in the tech landscape. Costs are spiraling out of control while major tech companies are dominating critical resources. These large cap behemoths are forecast to spend $325 billion in infrastructure this year alone, fueling a cloud compute market that’s already worth $675 billion in 2024. This massive influx of capital is driving up costs and creating a supply crunch that is stifling innovation and competition. It is also providing significant strain on centralized cloud providers, leading to increased latency. As demand outpaces supply, many businesses are feeling the pinch of being priced out of the AI race and find difficulty scaling their operations. 

• Censorship vulnerabilities: Power and control are concentrated in the hands of a few large corporations, making AI systems more susceptible to censorship by authorities or governing bodies. This can restrict access to information and cause manipulation of global narratives. It is possible for centralized systems to filter and remove information, potentially suppressing legitimate discourse and stifling freedom of speech. Users also often remain unaware of why specific content has been removed or altered, fuelling mistrust.

These challenges set Erick and Alex on a path to create a network that is more open, transparent, and inclusive for AI usage and development.

The Birth of Function Network

Function Network launched on Base as a platform for people to use AI, build with AI, and power AI while also being rewarded for it. The team has built a full-stack managed AI cloud where any type of user can access an array of open source LLMs, powered by a global infrastructure network. It is the first decentralized inference network that supports the next wave of AI powered applications in both web2 and web3.

• Provide Compute: At the core of Function’s full-stack decentralized AI ecosystem is Function Network, the foundational layer where anyone can contribute compute to run open source models. Unlike centralized platforms, Function’s decentralized structure distributes inference across different GPU and other hardware providers that not only lowers computing costs, but introduces fault tolerance to guarantee model uptime. Providers will be granted the right to offer compute and earn additional rewards by staking Function’s native token (FUNC). This serves as a reliable threat model that also enhances network security. 

• Provide Models: Function empowers anyone to contribute their own models. Model creators can upload, host, and share their open-source LLMs directly on Function. In return, they gain visibility, usage, and also rewards. Function levels the playing field for model creators, while making room for participation beyond the “Big 4” (Llama, Mistral, Qwen, and DeepSeek).

• Developer Platform: Function Network provides developers with an intuitive platform that delivers a one-stop shop for all AI development needs. It offers a suite of Open-AI compatible APIs (application programming interfaces) that helps streamline the integration of AI capabilities into other applications. This provides easy access to various open source models and allows Function to empower developers to create sophisticated AI-powered solutions, for free.

• Function Chat: Function Chat represents the user-facing application of Function’s infrastructure. It features a seamless, conversational AI experience that’s available on both the web and mobile app, while providing a familiar chat interface—much like what people use today with ChatGPT or Claude. By decentralizing a network of node operators, users can interact with a variety of AI models, reinforcing Function’s core belief in censorship-resistant technology.

Function’s decision to launch on Base is a natural fit, by leveraging Coinbase’s powerful distribution network and its success as the fastest growing L2. By building as an appchain L3, Function gains dedicated blockspace and more transaction throughput enabling the platform to scale efficiently. The move reinforces the team’s vision of decentralized, accessible infrastructure for AI innovation.

Join Function Network

Function Network is not just a platform, it’s a movement towards a more open, efficient, and accessible AI ecosystem. We invite businesses and developers to join on testnet and explore the endless possibilities of decentralized AI computing. 

The story of Function Network is far from over. In fact, it’s just the beginning. Start building here