Complexity, not capacity, is what kills projects.
Clusters sprawl across clouds and data centres. Developers queue for environments. Data scientists hack together notebooks. Finance teams can’t explain where the money went.
And the runway?
It vanishes.
More GPUs don’t fix it. More engineers don’t either. What’s missing is a way to collapse the chaos into something usable.
There’s one company building exactly that.
Who are they?
Company Background
Rafay’s founding team has made a career of being early.
At Soha Systems, they built the first clientless, zero-trust application access service years before “zero-trust” became a Gartner buzzword. Akamai acquired the company and folded it into its security stack.
Instead of sticking around, the team went back to their roots.
They started Rafay Systems with a sharper thesis: simplify how applications and compute are consumed, not just how they’re deployed.
The initial target was Kubernetes. In the late 2010s, Kubernetes took off, but it was still a nightmare to operate at scale. Enterprises sought agility without operational risk. Rafay built a platform that let developers get the benefits of containerisation without touching the complexity underneath.
That obsession with workflow simplification never went away, because Rafay’s customers don’t just stand up clusters.
They stand up entire environments.
Kubernetes for modern apps. SLURM for HPC. Jupyter Notebooks for researchers. GenAI workbenches for data scientists. Models-as-a-service for internal teams. All wrapped up with chargeback and billing, so CFOs can see where every GPU hour goes.
Then, sovereign AI and neocloud initiatives began to scale.
And Rafay’s platform found its sweet spot.
Operators needed a turnkey orchestration layer that could abstract multi-cloud sprawl while enabling governed self-service. Rafay turned any stack of compute in public clouds, sovereign deployments, or private data centres into a multi-tenant, monetisable cloud.
The approach is consistent.
Simplify the complexity. Wrap it in workflows. And make it consumable.
Executive Team
Haseeb Budhani, CEO - Former CEO of Soha Systems and CPO at Infineta. Veteran of building and scaling infrastructure platforms to acquisition.
Hanumantharao Kavuluru, SVP Engineering - Engineering leader at Soha and MobileIron. Brings depth in enterprise systems and cloud security.
Mohan Atreya, CPO - Product veteran from Okta and Neustar, with a track record of scaling complex platforms into enterprise adoption.
The same team has built and exited multiple companies. Rafay is where those experiences converge: simplifying orchestration while keeping governance intact.
The Edge
Rafay positions itself where few others can.
The orchestration backbone for Sovereign AI Clouds - Rafay goes beyond Kubernetes management, enabling multi-model tenancy, inventory management, SKU definition, and billing. It’s a full stack for operators who need to run GPU clouds at scale, across multiple regions, with enterprise guardrails baked in.
From racks to revenue in weeks, not years - Rafay compresses the GPU deployment and integration cycle, reducing time to market from months to weeks. For operators, that means a faster ROI, higher revenue per GPU, lower operating costs, and the ability to layer on software services for additional margin.
User experience as strategy - Rafay believes that user experience (UX) is the most important thing, regardless of the product being delivered. That’s why they invest heavily in delivering “best in class” UX for both their customers, and their customers customers.
Proof points that matter - Customers like Verizon or the US Air Force see measurable outcomes: 63% lower cloud costs, 4x more frequent deployments, and 76% lower MTTR.
Data scientists shouldn’t wait weeks for infra. With Rafay, CSPs and enterprises deliver secure, on-demand Jupyter, Ray, and Kubeflow environments behind their own firewall.
Learn more today: buff.ly/YlU1h42
— #Rafay Systems (#@rafaysystemsinc)
8:00 PM • Jul 7, 2025
Recent Moves
Rafay has been busy turning credibility into partnerships.
Reference Architecture with NVIDIA - Published a joint RA for sovereign AI deployments with NVIDIA for enterprises and national operators wanting to maximise their GPU investments. Vikram Sinha, CEO of Indosat Ooredoo Hutchinson, publicly backed it as the orchestration model for their NVIDIA Cloud Partner rollout.
Global teaming with Accenture - Signed a broad teaming agreement with Accenture to co-deliver sovereign AI projects worldwide, and named as the orchestration layer that makes secure deployment possible in a joint NVIDIA/Accenture whitepaper.
Ecosystem integrations - Signed partnerships with Netris, Data Direct Networks, and other ISVs to extend Rafay’s reach into storage, networking, and AI infrastructure optimisation.
Customer validation - Competed against the hyperscalers to land Verizon, Samsung, Amgen, Guardant Health, and multiple US federal agencies.
Rafay CEO @haseebbudhani explains why Accenture saw Sovereign AI Clouds as a global opportunity
Rafay powers the orchestration layer in Accenture’s Sovereign AI architecture:
• Multi-tenant infra
• Self-service AI envs
• Hybrid + sovereign-ready— #Rafay Systems (#@rafaysystemsinc)
9:00 PM • Jul 29, 2025
What’s Next
Acceleration will define Rafay’s next 12 months.
The company expects to be powering sovereign AI clouds on nearly every continent. Governments and enterprises want GPU clouds operational in months, not years.
Rafay’s pitch, and execution so far, align perfectly with that urgency.
The platform is also set to deepen.
Smarter scheduling is coming. Tighter billing integrations with systems like Monetize360 are underway. Support for emerging accelerators beyond NVIDIA is on the roadmap.
As the hardware market fragments, Rafay’s neutrality could become its most valuable differentiator.
But execution risk is real.
Hyperscalers won’t cede ground easily. Startups promising GPU orchestration are multiplying. And sovereign buyers? Notoriously difficult to convert.
Rafay must balance speed with stability, proving it can handle a global scale without diluting its value proposition.
Inside enterprises, the opportunity is equally complex.
CIOs are under pressure to repatriate workloads, cut cloud costs, and prove sovereignty. Rafay positions itself as the enabler of “cloud experience, on-prem economics.” But getting conservative enterprises to change entrenched habits will take more than technology.
It will take flawless delivery.
Where does that leave Rafay?
Potentially as the standard PaaS layer for GPU clouds worldwide. Not the chipmaker. Not the data centre operator.
But the orchestration layer in the middle makes infrastructure consumable and profitable.
The market is crowded, the stakes are high, and Rafay’s ability to simplify complexity, at scale, is about to be tested.