AiThority Interview with Hiren Majmudar, President, NeuReality

Hiren Majmudar, President, NeuReality, discusses more on today’s network infrastructure trends, AI across compute, CPU, NIC engines and a lot more in this AiThority quick-chat.

———–

Hi Hiren, welcome to our AiThority.com Q&A series, tell us about yourself and your hi-tech and AI journey.

My career has been focused on driving innovation and investment at the intersection of technology and society for over three decades. Prior to NeuReality, I held leadership roles at GlobalFoundries, where I led the data center, communications infrastructure, and compute businesses. Before that, I served as VP and Managing Director at Intel Capital and led software and engineering teams at Intel. My experience at Intel Capital gave me a unique perspective on the AI landscape, where I helped shape investment strategies in AI, compute, and networking silicon, and oversaw key acquisitions and partnerships.

Also Read: AiThority Interview with Jie Yang, Co-founder and CTO of Cybever

What are you most looking forward to in your new role at NeuReality?

I’m thrilled to join NeuReality at such a pivotal moment in the AI Age, the company’s history, and the market shift toward AI Inference. Where Training is how AI is created, Inference is how AI is used. The AI landscape is shaping up with ongoing innovation and daring moves – and now AI inference truly is a distinct market from AI training. NeuReality is uniquely positioned to drive positive change with our customers and across the industry, with the game-changing NR1 technology that unlocks the full potential of their GPU and AI accelerators. I’m eager to collaborate with industry leaders – hyperscalers, chip makers, system integrators, cloud service providers and the channel – to form strategic partnerships that marry our technology with theirs and drive widespread adoption to the benefit of end-customers across diverse industry sectors including finance, life sciences and government.

What trends in AI compute and network infrastructure are defining the market today?

The AI world is shifting from training to inference. This is creating a huge demand for specialized hardware and software because they are two distinct markets with different business and technical requirements. Many companies are building closed systems – whether it’s Nvidia, Cerebras, Groq or SambaNova by the sheer fact that they all rely on traditional CPU systems.

NeuReality is taking a different approach as the only open, agnostic, purpose-built AI Inference architecture designed to work with any GPU or AI accelerator by pairing them with our novel 7nm NR1™ server-on-chip. So we replace the CPU and NIC cards – we don’t compete with GPUs or any AI accelerator; we aim to make them and their underlying system architecture better.

The future of AI is open, adaptable, and optimized for real-world results aligned to business outcomes.

For providers trying to expand in this space: what strategies work in your view?

In the AI jungle, it’s easy to get lost chasing butterflies. Instead, focus on becoming the lion:

• Hunt the right prey: Identify a real customer pain point and solve it relentlessly. Don’t invent problems where they don’t exist. The AI savanna is teeming with genuine challenges.
• Pride yourself on partnerships: Find allies who complement your strengths. Build a coalition that can deliver a complete and compelling solution.
• Sharpen your claws: Develop a unique advantage. Make your partners’ solutions better, faster, or more efficient. Become indispensable.
• Defy the herd: Don’t blindly follow the pack. Lions are independent thinkers, forging their own paths. Challenge conventional wisdom to break down barriers and empower end-customers to achieve higher profit margins through new AI revenue streams while reducing the cost and complexity of AI deployments.
• Roar with results: Back up your claims with solid proof. In the post-hype AI era, trust and credibility are earned through demonstrable business ROI.

How are you seeing different advances in AI across compute, CPU, NIC engines and others change the entire IT architecture and infrastructure support game today?

The AI explosion is pushing today’s IT infrastructure to its breaking point. While we’ve seen incredible innovation across the spectrum of AI accelerators and compute engines, we haven’t seen that level of innovation in the systems that support those accelerators. AI still devours energy and budgets.

The newer AI inference entrants offer smart alternatives – where the AI systems must be as innovative and advanced as the tech giants or more so. To me, the ultimate measure isn’t just price/performance – but AI performance per dollar, per watt, and per query which can translate into customer acquisition and retention targets. For example, if I need to grow my retail business by 5000 net-new customers in 2025 who are AI-ready, by how much do I need to scale my AI/IT infrastructure and to what extent can I correlate the Customer Acquisition Cost (CAC) with Cost Per Token?

We need to shift our AI discussion from hype to practical reality. While tech companies treat AI as a capital investment, most businesses implement it as an operational expense. This fundamental difference requires a more pragmatic approach to AI deployment and economics.

Also Read: AiThority Interview with Robert Figiel, VP of Centric Market Intelligence R&D at Centric Software

Can you talk more about some of the biggest tech giants as well as innovative startups that are making waves in different areas of AI (how / why).

AI has the power to transform every industry, but excessive costs and uncertain ROI still hold the majority of organizations back – with 42% global and 33% U.S. companies having deployed AI as of September 2024 only an 8% gain from the year previous, according to Exploding Topics. Even among those deploying AI, the extent and sophistication of usage can vary. That’s why the most impactful innovators are those who break down market and technical barriers.

When I say AI Infrastructure, I don’t just mean hardware and networking. I’m talking broadly about memory, air and liquid cooling, security, software, and deployments across cloud, on premise and edge computing. I mean ethical and responsible AI, too. Look at cutting-edge cooling solutions like those from Asperitas and ZutaCore, who are pushing the boundaries of thermal management with immersion cooling and direct-to-chip liquid cooling, respectively. These advancements are crucial for maximizing AI performance, but they also aim to reduce environmental harm. Performance and protecting the planet – and not or.

Giants like AMD and Salesforce are making waves in AI hardware, cloud, CRM, and platform integration. Meanwhile, startups like Cerebras and SambaNova challenge the status quo with fresh approaches to AI memory bandwidth and data flow-as-a-service. And NeuReality reimagined AI Inference architecture to be CPU-free to boost GPU utilization to 100% maximum capability – by up to 4x what it is today with GPUs running idle a big proportion of the time. We do that while delivering up to 13-15x greater energy efficiency and 50-90% greater price/performance versus today’s CPU-reliant systems. For businesses and governments, that translates directly to lower operational costs and therefore higher profit margins.

I like to say, we’re the lions defying the herd, forging new paths to make AI systems more performant, energy efficient and affordable. This collaborative spirit, bringing together the strength of giants and the agility of startups, is what will unlock the full potential of AI across industries.

[To share your insights with us as part of editorial or sponsored content, please write to psen@itechseries.com]