Hewlett Packard Pathfinder

News

April 4, 2025

alt text

Welcome to Pathfinder Insights! This year marks a special milestone — Hewlett Packard Pathfinder’s 10th anniversary. Over the past decade, we’ve partnered with visionary founders, backed groundbreaking technologies, and helped drive innovation across industries. In this issue, we reflect on this journey and what’s next for Pathfinder.

We also explore how Generative AI is revolutionizing software development, offering a fresh perspective on its transformative impact. In addition, we sit down with Punit Chiniwalla following his recent promotion to Senior Managing Director of Pathfinder. Over the past two years, Punit has played a key role in shaping investment strategies and strengthening partnerships. Now, with expanded responsibilities, he shares his immediate priorities, key lessons from his journey, and insights into the evolving shift from AI training to AI inference.

This issue also features a Q&A with Dragos' CEO and Co-founder Robert M. Lee, diving into the latest trends in industrial cybersecurity. Lastly, we break down key takeways from NVIDIA GTC, analyzing the event's broader implications for AI and computing.

Pathfinder is looking forward to its next decade of innovation and collaboration.

Celebrating a Decade of Innovation with Hewlett Packard Pathfinder

This year marks the 10th anniversary of HPE’s Pathfinder program, a milestone that underscores HPE's commitment to driving innovation and fostering transformative solutions. Since its inception in 2015, Pathfinder has played a critical role in identifying and investing in category-leading startups that align with HPE’s strategy. Over the past decade, the program has made 50 strategic investments — deploying $400 million to date — while supporting 14 companies that have achieved successful exits and 13 that have reached unicorn status.

Discover how Pathfinder has driven innovation in enterprise technology and what’s next for the program. Read the full story here.

How Generative AI is Revolutionizing Software Development

Generative AI is transforming the way developers work, with AI coding tools delivering significant productivity boosts and accelerating software development.

These tools fall into two main categories: (1) Copilots, like Codeium, which work alongside developers to enhance existing workflows, and (2) AI software engineers, such as Devin from Cognition Labs, designed to perform tasks like a human engineer and potentially serve as a full-scale replacement. By automating repetitive, manual tasks, these tools free up developers to focus on more complex problems and achieve higher-quality results.

Click here to explore how AI tools are reshaping software engineering and transforming developer workflows.

Pathfinder One-On-One: Punit Chiniwalla

In this edition of Pathfinder 1x1, we sit down with Punit Chiniwalla following his recent promotion to Senior Managing Director of Pathfinder. Over the past two years, Punit has been deeply involved with our portfolio, helping to shape investment strategies and strengthen partnerships. Now, with expanded responsibilities, he’s focused on new opportunities to further strengthen Pathfinder’s impact.

In our conversation, Punit shares his immediate priorities, key lessons from his journey, and insights into the evolving shift from AI training to AI inference.

1. Congratulations on your promotion to Senior Managing Director of Hewlett Packard Pathfinder! With your expanded responsibilities, could you share your immediate priorities you plan to focus on?

Thank you! I'm excited to lead the strong team at Pathfinder! Pathfinder has an established track record of working with the venture community and delivering meaningful value to both HPE and our portfolio companies.

Our team meets with well over 1,000 new startups each year at various stages of their lifecycle. My immediate priorities are to focus this pipeline to drive impact through:

• Strategic Alignment with HPE's Long-Term Vision – Working closely with HPE leaders to identify disruptive technologies, complement business lines, and unlock new revenue streams.

• Strong Thematic Sourcing and Smart Investment Strategy – Leveraging a thesis-based investment approach with robust due diligence and financial discipline.

• Active Value Creation for HPE through Portfolio and Ecosystem Engagement – Going beyond capital to drive HPE collaboration with startups, venture capitalists, and industry thought leaders.

2. Can you share a key lesson or insight from your previous roles that you believe will be particularly valuable in your new position as Senior Managing Director?

I have been fortunate to have a vantage point of seeing technologies both succeed and fail through multiple economic cycles. I started my career in R&D at IBM Watson and then spent the better part of two decades in venture capital and corporate venture roles, including positions with a large Silicon Valley VC, running a CVC out of Asia, and serving as Director of Corporate Development at AWS during the early Gen AI days.

Predicting the future is nearly impossible, but these experiences help guide the team in finding the balance between strategic value and financial returns while staying at the cutting edge of disruptive technologies.

3. In our last issue, you mentioned looking forward to the shift from AI training to AI inference and its potential widespread impacts. Even though it’s only been a couple of months, what early activities or initiatives have you been most excited to see unfold in this area?

Managing AI inference is a major opportunity for HPE. As inference workloads scale exponentially, a significant portion will be distributed across enterprise and edge deployments. However, capturing value in this fast-moving landscape is challenging—GPUs are advancing faster than hardware cycles, inference systems require better middleware coordination, and business models are shifting from SaaS to usage-based (token) pricing. I am particularly interested in tracking opportunities that:

Make the Infrastructure More Efficient

• Inference Optimization – Reducing computational costs via model pruning, quantization, and distillation. Workload Management – Enhancing scheduling and orchestration to maximize resource efficiency across hybrid environments.

• Edge AI & Microservices – Enabling real-time inference at the edge with lightweight, scalable architectures.

Drive More Usage

• Vertical AI Applications – High-performance inference unlocking real-time decision-making in healthcare, finance, and manufacturing.

• Agentic Platforms – AI-driven agents that optimize workflows, automate processes, and improve enterprise intelligence.

Enable New Hardware/Software Collaborations

• Co-Designing Hardware & AI Systems – Optimizing inference workloads for GPUs, TPUs, and custom ASICs.

• Leveraging HPE’s Go-to-Market Strength – Partnering with AI providers to simplify enterprise adoption and deployment.

Inside Our Portfolio: Q&A with Robert M. Lee of Dragos

Welcome to the latest edition of our series, "Inside Our Portfolio", where we engage in insightful conversations with leaders from our portfolio companies. This series aims to provide you with firsthand perspectives from those at the forefront of innovation and growth.

In this issue, Todd H. Poole, Managing Director of Hewlett Packard Pathfinder, interviews Robert M. Lee, CEO and Co-founder of Dragos. Dragos is a cybersecurity startup specializing in protecting industrial control systems (ICS) and operational technology (OT) from cyber threats through threat intelligence, incident response, and security solutions.

TP: When people think of cybersecurity, they typically imagine traditional IT assets – servers, laptops, cellphones, Inboxes – how do OT environments differ, and how does Dragos address these differences compared to traditional IT cybersecurity solutions?

RL: Operational Technology environments are different from traditional IT environments in that they control and monitor physical processes and equipment rather than manage data. When OT systems are disrupted, we have energy outages, product shortages, water safety hazards, and halted manufacturing. Because OT runs the revenue-generating side of the business that produces goods and essential services, cyber threats can cause significant financial loss and reputational damage. Dragos provides specialized OT cybersecurity technology and services that protect industrial operations from disruption, maintain safety, protect intellectual property, and build resiliency. The Dragos Platform identifies what assets are in the OT network, monitors the full OT environment, and provides OT vulnerability management and threat detection.

Click here to read the full interview.

Pathfinder at NVIDIA GTC: Key Takeaways on AI’s Next Era

Imagine a company creates a product that makes its predecessor effectively obsolete. Well, that's what NVIDIA did. Pathfinder recently attended NVIDIA’s GTC conference, where we immersed ourselves in a dynamic environment of innovation, connected with emerging startups, and participated in VC-specific events. We attended the full keynote, connected with emerging startups, and participated in VC-specific events. In addition, we were actively involved in HPE-focused sessions, including VP of Sales Enterprise & Global Accounts Monica Gille's Neoclouds+VCs reception and the Unleash AI Partner Program gathering for ISV partners related to HPE Private Cloud AI. These engagements set the stage for sharing Pathfinder’s perspective on the takeaways relevant to HPE.

The keynote delivered by CEO Jensen Huang was a highlight, where he outlined an ambitious vision to not only dominate AI compute but also extend leadership into networking and storage solutions. Among the most compelling announcements were the next-generation chips—Blackwell, which is already in production, and Rubin, expected by 2026. These new models have eclipsed the previous Hopper series, offering performance leaps that are nothing short of revolutionary. With Blackwell providing almost 68 times higher performance at 8 times lower cost-per-performance and Rubin promising a staggering 900 times improvement with a 30-fold cost efficiency, it’s clear that NVIDIA is setting new benchmarks in chip technology.

Equally fascinating was NVIDIA’s advanced scale-up strategy. Jensen showcased how the company has dramatically compressed 1 exaFLOP of capacity into a single rack, moving from a 4U design to a standard 2U/1U configuration with direct liquid cooling managing 120kW of power per rack. The announcement of an “Ultra Rubin” system slated for 2027, featuring 15 exaFLOPs per rack with a 600kW power envelope, pushed the envelope even further—even if it raised some eyebrows among experts regarding the physics of such efficient cooling. Additionally, the introduction of desktop-scale DGX systems, including the DGX Spark and DGX Station, positions NVIDIA to potentially disrupt lower-end compute markets and possibly compete with HPE Private Cloud AI products.

Other notable announcements further illustrate NVIDIA’s expansive vision. The launch of DynamoOS—a new operating system designed for orchestrating workloads across data centers—brings together innovations in scheduling, data management, and caching to achieve dramatic cost and performance improvements. The unveiling of a dedicated quantum computing lab in Boston signals a strategic push toward hybrid quantum-classical computing solutions. In the realm of networking, NVIDIA introduced ScaleX and Quantum-X800 platforms, marking a significant step with next-generation photonic interconnects, and announced a strategic partnership with Cisco to integrate NVIDIA AI into Cisco’s networking gear. On the storage front, Jensen underscored a shift toward semantics-based systems, emphasizing how optimized data organization and caching technologies are set to revolutionize inferencing. Overall, the conference showcased a robust lineup of advancements that are poised to reshape the landscape of AI and computing infrastructure.

Pathfinder: What's New in April 2025