OXMIQ Labs Steps Into One of the Most Ambitious AI Infrastructure Plays of the Decade
In March 2026, OXMIQ Labs announced a strategic technology partnership with AM Intelligence Labs (AMI Labs), a business division of AM Group, to co-architect what could become one of the world’s largest renewable-powered AI compute platforms. The target: 2 GW of AI compute capability across India by 2030, with an initial 1 GW AI Compute Hub already in active development in Noida, Uttar Pradesh.
The partnership brings together two entities that each solve a different piece of a very large puzzle. AMI Labs brings the energy. OXMIQ brings the silicon, systems, and software expertise to put that energy to work efficiently at a scale most compute platforms have never attempted.
Why India, Why Now
Raja Koduri has been building hardware and systems for the world’s most demanding computing environments for decades. The decision to direct OXMIQ’s expertise toward India reflects just how significantly the country’s position in the global AI economy has shifted. India is now emerging as the world’s second-largest market for AI usage and token consumption, driven by a dense developer ecosystem, an expanding digital economy, and rapid enterprise adoption of AI across industries.
AM Group, parent company of Greenko, India’s largest green energy producer, sits at the center of this shift. Greenko operates 50 GW of renewable capacity across solar, wind, and hydro, supported by 100 GWh of intelligent energy storage. The group supplies roughly 2% of India’s total power, and that energy is owned, operated, and priced carbon-free at 50 to 70% below conventional data center power costs.
That cost structure matters enormously at scale. One of the most persistent challenges in large scale AI infrastructure has been securing reliable, affordable, and clean power. AMI Labs has addressed exactly that.
What OXMIQ Brings to the Platform
Architecture From the Ground Up
OXMIQ will serve as the architecture and engineering partner for the entire compute platform. That means working with AMI Labs on systems architecture, the hardware roadmap, and the supply chain strategy that will underpin the facility. The scope of OXMIQ’s involvement spans the full compute stack:
- Transistor-level GPU architecture and advanced packaging
- Rack-scale systems and high-performance interconnects
- Orchestration software for AI workloads at massive scale
- Liquid cooling, interconnect topology, and accelerator selection
- Workload orchestration across a unified, end-to-end system
Every layer is being engineered as a single integrated platform rather than a collection of independently sourced components. This unified approach is what OXMIQ describes as optimization from photons to outcomes, or in more technical terms, electrons-to-tokens economics. The goal is to deliver dramatically lower-cost AI compute at gigawatt scale by removing inefficiencies at each layer of the stack.
Speed and Modular Execution
Phase 1 of the Noida Compute Hub is already in active execution. Bringing the initial compute capacity online by the end of 2027 is identified as a key near-term milestone. OXMIQ is working closely with AM Group to optimize modular execution delivery, ensuring the platform can be deployed at speed without sacrificing efficiency or long-term scalability.
A Fully Vertically Integrated Platform
AMI Labs is developing the 1 GW AI High Performance Compute Hub in Noida as a fully vertically integrated platform. That integration spans several layers:
Owned carbon-free power generation feeds directly into advanced data center infrastructure, which in turn supports high-performance accelerators, a complete software stack, and a range of applications. On the consumption side, AMI Labs is building flexible service models, including AI Pods-as-a-Service and Tokens-as-a-Service, giving customers multiple ways to access compute depending on their workload requirements.
The vertical integration is what makes the economics work. When the same organization controls the power source, the data center, the hardware, and the software, every optimization across those layers compounds. That is the core premise of what AMI Labs and OXMIQ are building together.
What the Leaders Are Saying
Raja Koduri was direct about why the partnership made sense: “AMI Labs is the ideal partner for OXMIQ. They have solved the hardest constraint in large-scale AI infrastructure: access to reliable, carbon-free power at global scale. Our team has spent decades building silicon, systems, and software that power the world’s most advanced computing platforms. Bringing that expertise into AMI Labs’ infrastructure from the first architectural decisions means every rack, every interconnect, every storage and cooling system is designed around the workloads and economics required for the AI era.”
Anil Chalamalasetty, Group Chairman of AM Group, framed OXMIQ’s value to the platform from the other side of the table:
“OXMIQ gives AMI Labs access to some of the deepest hardware and systems expertise in the industry. Their team’s experience across leading Silicon Valley companies is exactly what we need to architect infrastructure that can compete globally. Together we are laying the foundation for AMI Labs to become a full-stack AI compute platform.”
About Raja Koduri and OXMIQ Labs
Raja Koduri founded OXMIQ Labs, headquartered in Campbell, California. His career spans leadership roles at Apple, AMD, Intel, and ATI Technologies. OXMIQ delivers licensable chiplet-based AI hardware and software solutions built for the age of inference.
The company’s product portfolio includes OxCapsule and OxPython, which deliver immediate optimization across heterogeneous hardware, and OxCore and OxQuilt, which provide the chiplet-native roadmap to zettascale efficiency. With this partnership, the engineering depth OXMIQ has built under his leadership is being applied to one of the most consequential compute infrastructure projects currently underway.