I just returned from the NVIDIA GTC conference where Salute was an exhibitor and where I spoke with a number of our customers about their accelerating AI deployments. The event’s nickname, “The Superbowl of AI” is well earned. It is a showcase for the greatest innovations happening in AI, and the event has an atmosphere of excitement and anticipation that is unique for technology conferences.
I have attended GTC many times in the past, and each time I come away with a greater appreciation of the creativity, passion and admirable goals of the people working on advanced computing technology. But this year’s event made an even bigger impression on me for a number of reasons. I want to share my observations because I believe we are at a really important moment in the history of technology where the work that all of us are doing takes on an even greater importance and urgency.
Significant Momentum on Impactful Use Cases:
One of the biggest takeaways for me from GTC is the progress that is being made on significant use cases for AI that will impact people’s lives in positive ways. There is exciting work being done to advance the science that will lead to treatments for life-threatening illnesses. There is important work being done to transform global telecommunications in a way that will raise up people across the world and drive economic opportunity. There are major initiatives that will modernize the power grid and drive cleaner energy. And major progress on many other important, impactful use cases. The speed of progress in just a year across all of these important applications of AI is breathtaking, and I’m very proud that Salute is playing such a key role delivering the operational model that makes these use cases possible..
A Revolution in Computing Infrastructure:
NVIDIA CEO Jensen Huang gave a wide-ranging keynote address that makes it clear how transformative AI technology is across every industry. One of the most profound things in his presentation was his vision for new computing infrastructure. He refers to it as an “accelerated computing fabric” comprised of next-generation data centers, networks and computing systems – all working together drive innovation and to help us make those ambitious AI use cases I mentioned above a reality. Data centers are the indispensable infrastructure for this vision, and Huang went into great detail about the AI factories that are needed to power GPUs and manufacture tokenized assets that enable AI applications to solve complex challenges.
Huang was purposeful in calling these facilities AI factories and differentiating them from traditional data centers because their purpose, design and operations are fundamentally different. The energy density of these facilities and the critical work being performed by GPUs require an operational model and a cooling strategy that is purpose-built for AI. Salute has been well ahead of the curve on this and has developed the first operational model for AI/HPC Direct-to-Chip cooling that meets the unique needs of AI factories. It’s clear from Huang’s keynote that operations are more critical than ever. The companies with the right operational model will be able to protect their investments in AI and accelerate their initiatives.
Accelerating Scale of Deployments:
One of the aspects of Huang’s keynote that will generate a lot of headlines is the volume of GPUs that are being manufactured. He reported that NVIDIA has already delivered 6 million Blackwells and has pre-sold $500 Billion in GPUs and Servers for 2025-2026. This is a remarkable pace of production that makes it even more critical for data center operators to be ready to successfully support those systems. That pace of production is an important reminder of how urgent it is for data center operators to ensure that they are truly ready for DTC liquid cooling operations. AI/HPC operations require a fundamentally new operational model that is designed for AI and that can scale at the pace that these numbers speak to.
It is equally important to have a team with the right skills. Throughout Huang’s keynote, one of the themes that come through is the importance of people with critical skills. He was speaking broadly across so many areas of the AI, but his point is particularly relevant for data center operations. Data center operations teams need entirely new skills and training to work in AI environments, and the time is now to be building those skills in order to be able to scale your team at the pace needed to support rapid deployments.
Huang’s keynote and all of the conversations I had with companies at GTC make it clearer than ever that data center operators need the right operational model and an operations team with the right skills to protect their investments in AI. Salute launched the industry’s first Direct-to-Chip Liquid Cooling Operations Service for AI data centers at the GTC event. The service is based on years of work to prepare for the moment that Huang describes in his keynote—work that involves countless conversations, collaborations and pilot projects with NVIDIA, the NVIDIA ecosystem, OEMs, liquid cooling companies, data center providers and other experts. We were preparing for the day when companies would need to take action on AI operations. GTC made it clear that day is today. And I’m proud that Salute is the partner who can help data center companies rapidly establish world class operations for their AI infrastructure.
About the Author:
John Shultz is the Chief AI Product and Learning Officer for Salute and a highly respected thought leader in the data center industry. He brings nearly three decades of experience in digital infrastructure to this role leading Salute’s AI-driven data center solutions, which enhance data design, build, and operations through AI innovation. Prior to joining Salute, Shultz was the founder and executive consultant at Advanced Data Center Consulting Group, which advised and trained Data Center Service Providers, investments managers on AI Market trends, demand and Direct to Chip Liquid operations.
In this role, he was also the industry’s foremost educator and trainer of investment teams, data center service providers and other executives on the impact of AI on the data center market. Prior to founding that successful consultancy, Shultz held senior positions as the global head of product management at the two largest data center providers in the world: Digital Realty Trust, the world’s largest wholesale data center company, and Equinix, the world’s largest retail data center company. At both organizations, his responsibilities included P&L management, new product development, customer interviewing, competitive intelligence and acquisition. He is a graduate of Texas A&M University.