The Infrastructure Conversations That Will Shape Digital Markets

Share this article

A few days at NVIDIA GTC leaves little doubt that the scale of what is underway in compute, infrastructure, and applied AI is difficult to fully grasp until you are inside it. 

The conference offered a full day workshops ranging from the most complex topics like robotics and twinning to building LLM applications with prompt engineering and was followed by sessions that assumed participants were prepared to stretch technically, conceptually, and operationally. Across the conference, there was very little passive content. Nearly every room required concentration, curiosity, and a willingness to absorb unfamiliar ideas quickly. It felt, in many ways, like a data enthusiast’s version of a world fair.  

The case studies alone made clear why the conversation around compute has become so central. AI deployment is now touching almost every sector: healthcare, robotics, industrial systems, climate modelling, logistics, media generation, digital twins, autonomous systems, language interfaces, and scientific research. The volume of processing required to support this breadth has moved computational power from technical consideration to strategic necessity. 

That reality explains why infrastructure received as much attention as applications. NVIDIA’s latest rack-scale systems are engineered to be liquid cooled using warm water rather than energy-intensive chilled systems, reducing power requirements associated with thermal management while allowing far greater density inside data centre environments. Sustainability was not treated as a secondary concern; it was embedded directly into how next-generation compute is being built. This demonstrated how quickly scientific and engineering barriers move when the market behind them becomes large enough. Once major industrial demand forms around a problem, whether cooling, interconnect efficiency, memory bandwidth, or inference speed, progress accelerates at remarkable speed. Entire ecosystems appear to reorganize around solving what, only recently, looked like hard constraints. 

Sovereignty was another constant theme, although the language around it differed depending on who was speaking. In Silicon Valley, sovereignty often referred to corporate control: where proprietary models sit, who governs access to data, how organizations retain control over internal intelligence layers, and how dependent they become on external platforms. This is an important distinction for Canadian businesses because corporate sovereignty deserves more attention than it often receives in our own market conversations. 

At home, sovereignty usually points first to infrastructure: domestic capacity, independent compute environments, trusted hosting, and control over critical digital pipelines. At GTC, both versions of sovereignty were clearly visible. International buyers are demanding environments that preserve control, comply with local legal frameworks, and still operate within globally connected technical standards. That demand is influencing architecture directly. 

What is emerging may be a new kind of ecosystem shaped by geopolitics as much as engineering. Markets are beginning to define their own infrastructure preferences while still relying on hardware and systems built for interoperability. That balance of local control without technical isolation appeared repeatedly across discussions involving governments, enterprise leaders, and global infrastructure players. 

For digital media, this matters more than it may first appear. Understanding how these systems work, where compute is being concentrated, how inference layers are being structured, and what safeguards are still under development will increasingly shape how media products are designed, bought, measured, and governed. 

The technical weaknesses are well understood inside these conversations. Prompt injection, hallucinations, model reliability, provenance, security vulnerabilities, and orchestration risks were not avoided topics; they were active areas of work. That matters because sustainable adoption depends less on excitement and more on building carefully enough that major systems do not require wholesale reconstruction two years from now. 

There is real value in spending time close to this level of development. Not because every industry participant needs to become technical, but because the future structure of digital markets is being influenced directly by infrastructure decisions now underway.

Related Articles

Corus (NextScreen) | Shark Ninja

Corus delivered a first-to-market shoppable innovation in Canada that extends brand storytelling through a custom second-screen experience - NextScreen. Built to support Linear, Connected TV, and Web, the format enables audiences to shop, engage, or learn more directly from video creative, reducing friction and helping drive conversion at the moment of inspiration. By combining upper-funnel impact with direct response capability, NextScreen allows brands to connect attribution to specific creative executions, supported by enhanced analytics and creative-level reporting through its custom portal.

Captivate | Dell Product Launch

Dell’s programmatic campaign delivered measurable gains in ad recall (68%), consideration (53%), and customers action (2 in 5). Captivate’s programmatic guarantee capabilities highlights DOOH’s ability to deliver predictable and reportable results. Watch the video here.

Read the Case Study