Opinions expressed by Digital Journal contributors are their own.
In the current gold rush around artificial intelligence, the conversation tends to center on models and capabilities. But behind the scenes lies a more foundational problem: access to infrastructure. Chips are scarce, power contracts are inefficient, and much of the available GPU inventory sits idle or is captured early by the largest players. Niraj Yagnik sees this imbalance not just as a technical issue but as a market failure.
Yagnik is the co-founder and CTO of FPX AI, a company developing a real-time exchange for AI infrastructure. Alongside CEO and lifelong friend Dhyay Bhatt, he’s building a system that treats GPUs, rackspace, and power as standardized resources that can be bought, sold, and deployed in an open marketplace. Sellers can list idle inventory. Buyers can access it quickly. Pricing is clear, and terms are consistent. According to Yagnik, this kind of liquidity does more than improve efficiency. It creates a fairer system for innovation.
Today, access to compute is often determined by size, influence, and timing. Yagnik believes it should be determined by need and speed. The current system, he explains, leaves billions of dollars’ worth of hardware sitting unused while startups and research teams wait in line. The result is wasted capacity on one end and restricted progress on the other.
Before launching FPX, Yagnik led product at BitOoda, a digital infrastructure investment bank and brokerage where he worked on cross-functional projects spanning AI, software, legal, technical, and compliance domains. It was during this time — and through his experience working with AI systems and datacenters — that he began to recognize the scale of demand for compute infrastructure. What stood out most was the absence of transparency. While traditional commodities benefit from mature marketplaces, AI infrastructure still operates in a fragmented, informal system. FPX was built to change that.
Yagnik’s background blends engineering with academic research. He grew up in central Mumbai, started coding at age 12, and eventually pursued a Bachelor’s in Computer Engineering. He recently earned his Master’s of Science in Computer Science with a focus on Artificial Intelligence from UC San Diego, graduating in 2024. He has published multiple peer-reviewed papers in areas like explainable AI and evolutionary algorithms. His interest in research, he says, came from a desire to contribute to the foundation of the field, not just build on top of existing tools. That mindset has carried over into his work today, where building systems from scratch often requires navigating ambiguity and solving problems without a clear precedent.
While FPX focuses on digital infrastructure, Yagnik and Bhatt are also working on another project aimed at the construction industry. Through a platform called Aurexis, he and his team are building AI-native software for engineering, procurement, and construction firms. Construction is one of the largest industries in the world but also one of the least digitized. Most software tools in the sector are designed for administrative reporting, not day-to-day execution. Aurexis is different. It provides AI-driven support for tasks ranging from site selection and scheduling to safety tracking and field operations.
According to Yagnik, the idea is not to layer automation on top of broken systems. It is to reimagine the entire workflow so that frontline workers, project managers, and executives are all supported by the same intelligence-driven platform. The company is already working with contractors in Massachusetts, supporting multimillion-dollar projects and helping teams execute with greater speed and visibility.
In both ventures, Yagnik focuses on making infrastructure more accessible. At FPX, that means giving smaller players a path to GPU access without needing insider connections. At Aurexis, it means giving small to mid-sized contractors access to tools traditionally built only for the largest firms. In both cases, the work is about lowering the barriers that keep important technology out of reach.
Yagnik also remains active in education. He advises the Generative AI for Value Creation program at Southern Connecticut State University, where he helps design a curriculum that bridges technical knowledge with real-world business needs. He has delivered guest lectures at UC San Diego and Manipal Institute of Technology, often speaking about AI ethics, infrastructure economics, and translating research into market solutions.
Much of his work is shaped by his early life in Mumbai, where he was raised in a tight-knit community that valued resourcefulness and grit. His partnership with Bhatt goes all the way back to childhood. That long-term relationship, Yagnik says, brings a level of trust that is essential when building under pressure. He also credits his parents and family for encouraging his interest in technology and for supporting his decision to take a less conventional career path. Currently based in Stamford, Connecticut, Yagnik continues to build at the intersection of AI, infrastructure, and real-world impact.
Yagnik does not pretend that startup life is easy. He describes it as intense, unpredictable and filled with risk. But he also says it is what he has always wanted to do. His long-term goal is to make FPX the global backbone for AI infrastructure, a neutral platform where GPUs and power can be traded with clarity and confidence. With Aurexis, he hopes to modernize construction software in a way that prioritizes execution, safety, and long-term impact.
In a sector filled with promises about what AI can do, Yagnik’s focus remains on what it requires. Compute, power, coordination. Systems that work in the background but decide who gets to participate in the first place. His work does not chase attention. It tries to fix the parts of the machine that no one sees but everyone depends on.
