Today, we are proud to announce the close of our $12M Series A round, led by Variant Fund and Polychain Capital, bringing our total funding to $20M. This milestone is meaningful because it brings us closer to Hyperbolic’s mission: making high-performance AI infrastructure more accessible to the developers, researchers, and teams building the next generation of AI.

Reflecting on this growth, I keep coming back to two principles that drove my co-founder Yuchen and me to build Hyperbolic: the stories of builders pushing the limits of AI, and the belief that access to infrastructure should expand opportunity rather than constrain it.

My Hyperbolic Story

For those who are new here, I’m Jasper Zhang. I earned my Math Ph.D. at UC Berkeley in just two years, after winning gold at the Alibaba Global Math Competition and the Chinese Mathematical Olympiad. Along the way, I explored different worlds, from the fast-paced halls of Citadel Securities to research at the frontier of distributed systems and verification. Now, as CEO of Hyperbolic, my focus is on something that feels deeply meaningful: making AI infrastructure more accessible to the people building the future.

I have always loved the simple elegance and beauty of mathematics: the ability to derive a complete world from just a few basic axioms. With mathematics as the foundation of so much of science and technology, I saw it as my entry point for making a meaningful contribution to human society.

Having focused on topology and hyperbolic geometry during my Ph.D., I found myself at a crossroads. The traditional academic path felt limiting. If I pursued a career in academia, I worried that I might not see the positive impact of my work on people’s lives. The finance industry offered a place where mathematics had immediate, tangible impact, so I explored quantitative research at Citadel Securities. After working in traditional finance, however, I found that the industry did not fully align with my values or with the kind of positive-sum impact I wanted to create.

I have always believed that life is a positive-sum game, where one person’s progress can contribute to the progress of others. That belief became even stronger as AI began to accelerate. I watched ChatGPT and other breakthroughs reshape the technology landscape, and I realized AI would have a profound impact on the future of humanity.

At the same time, some of the brightest people I knew working in AI were struggling to access the GPU resources and AI services they needed to move their work forward. Compute had become one of the biggest bottlenecks in AI progress.

That inspired me to start Hyperbolic Labs: to build an open-access AI cloud that gives developers and researchers access to reliable, scalable, and affordable GPU compute and inference. Our goal is to help more builders experiment, deploy, and scale without being blocked by infrastructure access.

The Team Behind Hyperbolic

It’s hard to overstate how important the human element has been in shaping Hyperbolic. When I first set out to build this company, I knew I wanted it to be more than a collection of talented individuals. It needed to feel like a team that genuinely believed in each other and in the promise of making AI more open and accessible. That kind of environment does not happen by accident. It grows when the right people converge around a shared purpose.

I still remember meeting my co-founder, Yuchen, at the tail end of 2022. We were two people who had grown tired of seeing GPU access locked behind barriers, and we believed the future should look different: more open, more efficient, and more accessible to builders. We found ourselves finishing each other’s thoughts on where technology could go if more people had access to the infrastructure they needed.

The energy between us was unmistakable. When a hyperbolic snowstorm unexpectedly stranded me in Seattle, we spent extra days mapping out what Hyperbolic might become. That extended stay remains one of my most cherished memories because it marked the moment we realized we were not simply dreaming. We were on our way to building something real.

From there, the team grew in the most organic way. For our first hire, I cold reached out to a talented frontend engineer, Matthew. Meeting face-to-face over coffee allowed us to share the vision person-to-person, not just founder-to-engineer. Matthew’s decision to leave a well-established company for this fledgling startup felt like a profound vote of confidence. It said to me: this is not just my dream. It is our dream.

Yuchen and I uprooted our lives in Seattle and New York to join Matthew in California because we believed that physically working together would forge a deeper sense of trust and cohesion. It did. We did not just work together; we shared in one another’s joys and challenges, supporting each other through every high and low.

Once we secured seed funding, our circle widened further. Matthew introduced us to his friends Viktor, Yoofi, and Kai, who were excited not only by the technology, but by our mission and team. Then we met Christian, who had long been thinking about how to make GPU infrastructure more accessible.

Bringing on Yuchen and my former teammates and friends like Lesheng, Peggy, Jeremy, Joelly, and Shouqiao felt like adding familiar notes to a melody we were still composing. They knew how we worked, we knew their strengths, and together we could fill in the gaps. Over time, we shaped a team dynamic that thrived not on rigid hierarchy, but on mutual respect, shared curiosity, and a willingness to help one another succeed.

What really stood out were all the unexpected, purely human moments: the small experiences that still make me smile whenever I think about them. During our San Francisco onsite, we tried Mexican tacos, watched election updates on TV, and played poker. The next night, we switched gears: authentic Hunan food followed by a screening of Rush Hour. Somehow, we accidentally invented what we now lovingly call the Hyperbolic handshake, and it became a secret signature of our team. One afternoon, we tackled an escape room together and managed to crack every code and solve every puzzle, only to realize at the end that the door had already been unlocked.

And of course, there were culinary adventures. We brought every new teammate to our favorite Xiao Long Bao restaurant, Seven Grams in Irvine, more than ten times and playfully insisted everyone try pig ears. We also fell in love with collecting idioms from one another’s hometowns and cultures.

These moments — handshakes, escape rooms, pig ears, and idioms — taught us as much about each other as any grand vision or strategy session. They are the stories that will live on long after the prototypes and pitch decks are outdated, and they remind me that the very best parts of building something new often happen between the formal meetings and planned milestones.

Looking back at the people who have joined Hyperbolic, I see not just colleagues, but friends who helped me shape a vision into reality. Each person brought their own conviction and their own reason for believing that making AI infrastructure more open and accessible would matter. Without them, Hyperbolic would not be what it is today.

It is a rare privilege to work with people who care deeply about both the vision and the team. They push me to think differently, consider perspectives I might have overlooked, and approach each challenge with empathy and optimism. We might not have all the answers yet, but we know the kind of environment that produces them: one founded on trust, guided by shared purpose, and energized by the possibility of making something better for builders everywhere.

That is the team behind Hyperbolic, and I could not be more grateful for each of them.

Hyperbolic Today: Building High-Performance AI Infrastructure

Today, Hyperbolic is an open-access AI cloud providing high-performance GPU compute and inference for developers, researchers, startups, and AI teams. Our journey from concept to reality has been guided by a core principle: AI infrastructure should be reliable, accessible, and efficient enough for more builders to participate in the next wave of AI innovation.

Drawing from Yuchen Jin’s experience in AI systems and my own background in mathematics and verification, we have built a platform focused on making powerful AI infrastructure easier to access and use. Hyperbolic gives teams a way to access GPU compute, run open-source models, and build AI applications without the long procurement cycles, high costs, or operational complexity that often slow teams down.

What the Close of Our Series A Means for Hyperbolic

The close of our Series A marks a real turning point for Hyperbolic. After two years of testing ideas, facing challenges head-on, and refining our vision, it feels like we are stepping from the scrappy, unpredictable early days into a new chapter defined by clearer direction and a more confident stride.

In those first two years, we tackled some of the toughest barriers holding back AI innovation:

  • Inaccessible compute

  • Limited access to optimized open-source models

  • Fragmented AI development workflows

We have also built live products to solve these problems:

  • Hyperbolic’s GPU Marketplace

  • Hyperbolic’s Inference Services

Instead of just talking about these problems, we got to work. We built the GPU Marketplace to make high-performance compute easier to access. We created our Inference Services so developers could run powerful open-source models without having to assemble every piece of infrastructure themselves. Each product was born out of necessity and shaped by feedback from the builders using Hyperbolic.

Now, with this new round of funding, we are ready to double down. We will expand our engineering team, strengthen our go-to-market efforts, and continue improving the infrastructure developers and researchers rely on to build AI applications.

My hope is that Hyperbolic becomes the place where AI developers and researchers come not just to solve problems, but to create. I want them to build their ideas on our infrastructure: services and applications that are accessible, performant, and ready to grow alongside them. It is about building fertile ground where ideas can take root, flourish, and spread to benefit more people.

At Hyperbolic, we have always believed that the real magic happens when we join forces. We go further faster when we go together. Now is your chance to be part of that. Let’s keep building together.

With gratitude,

Dr. Jasper Zhang, Ph.D.
Co-Founder & CEO, Hyperbolic

Take Your Ideas Hyperbolic

Visit app.hyperbolic.ai to access the infrastructure you need to bring your ideas to life.

About Hyperbolic

Hyperbolic is the Open-Access AI Cloud, giving researchers, startups, developers, and AI-native companies fast, flexible access to high-performance GPU capacity. The platform helps teams start on demand, scale programmatically, and grow into reserved infrastructure without long waitlists, rigid contracts, or complex procurement cycles.

Founded by award-winning math and AI researchers from UC Berkeley and the University of Washington, Hyperbolic is committed to making advanced AI infrastructure more accessible to builders around the world.

Website | X | LinkedIn | YouTube | GitHub | Documentation