March Madness Also Brings Hyperbolic Growth
Every month, more developers are discovering our AI Cloud, built to be accessible, automated, and affordable. Balancing all three isn’t easy, and we’re not claiming perfection but we’re committed to building with developers in mind as we ship new features.
We’re proud to share this month's stats, outcompeting our stats from previous months
165,000 developers
GPU rentals growing 2x MoM
Total GPU hours up 2.6x MoM
Unique inference users up 50% MoM
None of this is possible without our developer community pushing the limits of what’s possible. Have a feature request or idea? Jump into our Discord and let us know by submitting a ticket.
DeepSeek-V3-0324 Now Live on Hyperbolic
At Hyperbolic, we’re focused on delivering the latest models as soon as they’re available. This is our promise to the developer community. This past month, we rolled out support for DeepSeek-V3-0324, one of the first to do it, and also the first inference provider to support the model on Hugging Face.
Alibaba Qwen QwQ-32B now live on Hyperbolic in BF16
This past month, Alibaba released Qwen-QW-32B, a reasoning model that outperforms DeepSeek-R1 (671B) and OpenAI’s o1-mini. We’re proud to be one of the first inference providers to serve it, with support for BF16 for optimal precision and efficiency. If you’re a Hugging Face user, you can also try it on Hugging Face directly.
ICYMI Hyperbolic Has Low Compute Prices On High Performing NVIDIA GPUs
You can stop wasting time chasing GPU rentals. Hyperbolic gives you direct access to NVIDIA GPUs, cheaper than your usual provider. Spin up in seconds, scale on demand, and run your workloads without waiting on approvals or dealing with sales reps. Just code, deploy, and go.
Hyperbolic Was Recently Featured In Lex Fridman's Podcast
While this didn’t happen earlier this month, we recently realized that Hyperbolic was featured on Lex Fridman's latest episode with Dylan Patel (from SemiAnalysis) and Nathan Lambert (from AI2 and Interconnects). In the episode titled “DeepSeek, China, OpenAI, NVIDIA, xAI, TSMC, Stargate, and AI Megaclusters,” Hyperbolic was mentioned as one of the cost-efficient inference providers for DeepSeek, alongside players like Together AI and Fireworks AI.
Small wins and recognition like this mean everything to our small but mighty team. Thanks for including us.
Dive Deeper with Hyperbolic’s Research Articles
The AI industry moves fast. At Hyperbolic, we move faster. While our core audience is developers, we stay ahead by tracking the evolving landscape across AI. Our team is rooted in research and constantly learning, we build based on where the space is heading, not where it’s been. Each month, we will publish our own research to help our community grow with us. Catch the latest on our blog or X. This month we covered:
Looking Forward
We’ve been running user interviews with developers and are actively incorporating feedback into our roadmap. Expect updates like Auto Top Ups, a revamped Model Playground, and Custom Port Configurations for GPU instances.
On the company side, a major brand update is coming in May, built to inspire the AI community to push what’s possible.
Got feedback or requests? Join our developer community on X, LinkedIn, or Discord.
