Shadeform Makes GPU Access Faster and Easier for AI Developers

The tech landscape in our modern society, whether in the fields of artificial intelligence or machine learning, is constantly evolving. Companies in the industry are persistently utilizing AI and ML* to better their products and services, causing a swift rise in GPU* demand that has led to major shortages.

The continuous need for machine learning training and inference these days makes it difficult for businesses to purchase GPUs. With the prices of GPUs from private companies getting more unaffordable, the units are getting more inaccessible than ever.

A 2023 Y combinator-backed startup named ShadeForm, founded by Ed Goode, Ronald Ding, and Zachary Warren, developed a unified, easy-to-use API and platform to solve the ever-unfortunate GPU shortage for any company in need.

*ML: Machine Learning, a branch of artificial intelligence focusing on the use of data and algorithms to imitate the way humans learn and gradually improve accuracy.

*GPU: Graphic Processing Units, a specialized processor originally designed to accelerate graphics rendering and process many pieces of data simultaneously.

Extensive Efforts

With our tech-heavy world being deeply dependent on machine learning these days, the tech becoming increasingly important year after year, players from multiple sources are ceaselessly trying to find a solution to the GPU shortage.

Smaller cloud providers have started going online to provide more GPUs, though their greatly intrepid efforts haven’t played too big of a role. Due to their unstable capacities, disparate interfaces, and non-standard OS images, not many companies have been able to adopt GPUs at scale, which is what most companies desperately need at the moment.

GPU manufacturers are also pushing to increase production to meet the demand. However, to hope that their efforts will deliver quick results might lead to disappointment. Aware that increasing production will take a long time, manufacturers are allocating GPUs to customers based on priorities, as well. At the moment, however, the available GPUs have mostly been allocated to essential industries like healthcare and education, meaning that the tech industry might not make the earlier cuts.

Also aware of the issue are several government bodies that are building an effort to support the semiconductor industry, which includes GPU manufacturing. With hopes of ending the shortage, government bodies are offering financial assistance, tax breaks, and various incentives to manufacturers. Unfortunately, not much of their efforts have come to fruition just yet.

The three founders of ShadeForm, having first-hand experience in building and operating critical cloud infrastructure services like AWS and Azure, have grown aware that a simple and easy-to-use cloud-based machine learning infrastructure could significantly increase GPU availability for companies. With the added knowledge of infrastructure management, the three started developing ShadeForm’s API* and platform.

*API: Application Programming Interface, code that enables two software programs to communicate and integrate.

A Unified Platform For All

Shadeform’s unified platform allows any ML inference and training jobs to be deployed to any cloud provider. All ML-friendly providers are grouped into a whole cluster, so anyone can access said providers and manage their GPU instances from anywhere through the platform as well as the startup’s generalized APIs.

The variety of providers that come with the platform makes it incredibly versatile. It is also pretty accessible and simple to use. Once users sign up with Shadeform and connect to their accounts, they can choose whichever provider they need straight away, without any cloud account management.

Once chosen, only one click is needed to launch and deploy any GPU instance and workload. Users will then be given an SSH key* to start working with any of the providers right away. Shadeform also offers transparency, as it supplies users with the pricing data of each provider clear as day.

*SSH Key: An authenticated and encrypted access credential for the secure shell network protocol, a protocol used for remote communication between machines on an unsecured open network.

There could be instances where the specific GPU type or provider companies need won’t be available upon checking. With every user’s best interest in mind, the startup also allows auto-reservations. Users can reserve the providers they need and receive a notification as soon as they’re available. Users are also able to set auto-reservation configurations, meaning Shadeform can deploy their workloads for them when the providers become available.

By creating this platform, the three founders aim to build a future where companies will never be hindered from deploying and completing their workloads due to their GPU providers’ limited capacity and features.

Meme & AI-Generated Picture

Job Posting

• Chime - Creative Director - San Francisco, CA (Hybrid)

• City National Bank - Reporting Analyst II - Los Angeles, CA (Hybrid)

• Cox Enterprises - Cyber Risk & Compliance Manager - Atlanta, GA (Hybrid)

• The PNC Financial Services Group - Security Engineer, OUD - Pittsburgh, PA (Hybrid)

Promote your product/service to Digger Insights’ Community

Advertise with Digger Insights. Digger Insights’ Miners are professionals and business owners with diverse Industry backgrounds who are looking for interesting and helpful tools, products, services, jobs, events, apps, and books. Email us [email protected]

Your feedback would be greatly appreciated. Send it to [email protected]