Customer Stories

How SPREEAI cut storage costs by more than 30x and doubled model FLOPs utilization with Lambda

Lambda — Tue, 08 Sep 2026 19:02:00 GMT

How Kodiak doubled AI training performance on Lambda

Lambda — Fri, 05 Jun 2026 00:39:26 GMT

How Lambda built a Hyperscaler cluster in 90 days

Lambda — Tue, 18 Nov 2025 05:10:19 GMT

Creating beyond limits

Lambda — Tue, 18 Nov 2025 05:02:31 GMT

Challenge

Pika’s Mission to Unlock People’s Creativity

Pika, a fast-growing Al startup, is on a mission to democratize video creation by lowering barriers to entry and expanding creative opportunities for anyone with an imagination. While their goal is ambitious, the impact they hope to make is to bring more small, everyday creative ideas to life.

Pika empowers creatives to transform their ideas into videos that push the bounds of storytelling, with captivating realism and special effects.

Whether Pika's users want to create a funny video to share with friends, extend a family video, or make promotional content for their small business, they have the creative freedom to instantly turn their ideas into a reality. Pika's easy-to-use Pika 1.5 platform uses advanced generative Al technology to make it so that artists no longer need technical skills to make videos, democratizing access and allowing people to have fun.

“We are a very technologically advanced and focused company. We have the most brilliant minds in the world working on our models. What we’re trying to do with it now is move AI video creation to a place that is much more fun, much more accessible, and much less serious.”

AI-driven drug discovery

Lambda — Tue, 18 Nov 2025 05:02:31 GMT

Challenge

Providing the clinical-stage biotech company continuous access to GPUs for scaling its state-of-the-art drug discovery models

AI presents exciting possibilities across industries, particularly in drug discovery, where AI has the potential to change lives, transform health outcomes, and improve the quality of medicine and the pace at which they are developed. Innovation in the field relies on the ability to access and train models on massive amounts of data from a vast array of sources. Recent developments have seen researchers gain the ability to use AI/ML tools to predict the structure of proteins, virtually on demand, while profoundly expanding their ability to search for novel chemical mechanisms that might lead to new and better medicines. The paradigm of discovery based on traditional laboratory methods is rapidly being complemented by in silico platforms that have the potential to increase the speed and decrease the cost of bringing new medicines to patients.

Giving back to the open-source community

Lambda — Tue, 18 Nov 2025 05:02:30 GMT

Challenge

Demanding access to extensive NVIDIA GPU resources

The size of the model demanded extensive NVIDIA GPU resources, which was challenging, but the real hurdle fal faced was how to consume those resources—mainly, when training and experimenting with their models, how could they access what they needed, when they needed it. Initially, fal trained version 0.1 of AuraFlow using a onemonth contract with a cloud provider. While this allowed them to start development, it took weeks to negotiate a short-term contract that worked for them, and those were cycles fal didn’t want to spend every time they trained a new version of the model.

This prompted fal to seek a more agile infrastructure that can scale up for experimentation and training within seconds, then instantly scale down when not in use. They needed a system that provided on-demand resources, without tying them down to long-term agreements or paying for idle GPUs. As the fal’s Founding Engineer—Batuhan Taskaya—put it, “the first question we had was how the hell are we going to train this in terms of the GPUs we need, where are we going to find the GPUs?” With future iterations in mind, including versions 0.2 and 0.3, they sought asystem capable of adapting resource levels in real-time.

Powering industry-leading generative AI innovations

Lambda — Tue, 18 Nov 2025 05:02:28 GMT

For 3D artists, generative AI is closing the gap between creatively conceptualizing new ideas and executing immersive 3D content frame-by-frame. Creatives can now use generative AI to take an idea from a text or image input to a 3D model in minutes, allowing them to create at the same pace that they envision new ideas. Since its introduction, the pace of generative AI innovation has been staggering, and there are countless more transformative uses waiting for startups with the ideas and resources necessary to bring them to market. In competitive generative AI spaces like 3D model generation, leading players stay ahead by continuously driving innovation week after week. To remain at the forefront, constant improvement is not just an advantage—it's a necessity.

Targeting complex diseases with AI

Lambda — Tue, 18 Nov 2025 05:02:27 GMT

Overview

Genesis Therapeutics is at the forefront of AI-driven drug discovery, leveraging diffusion models, LLMs, and physical ML simulation together in a single platform to create next-generation protein- based therapeutics. Their mission is to accelerate the identification of treatments for complex diseases, particularly targeting “undruggable” proteins—proteins that are currently too large or intricate to be targeted effectively.

However, as Genesis expanded its research, they encountered the same bottleneck other AI innovators face: their models required exponentially more processing power to analyze and predict protein interactions, leading to long training times and slow iteration cycles.

To scale their research and accelerate discoveries, Genesis partnered with Lambda, leveraging high-performance computing to dramatically reduce training times, improve model accuracy, and push the boundaries of AI-driven drug discovery.