Video: Tech Talk: Inside the Toronto Data Center Market | Duration: 1928s | Summary: Tech Talk: Inside the Toronto Data Center Market | Chapters: Welcome to Tech Talk (29.220001s), Toronto's AI Rise (137.33s), AI Market Readiness (251.27s), Data Sovereignty Imperative (345.20000100000004s), Toronto's Connectivity Advantage (538.42s), Infrastructure as Strategy (732.88s), AI Use Cases (921.1800010000001s), AI and Sustainability (1095.240001s), Global Platform Strategy (1268.140001s), Infrastructure Mistakes (1528.740001s), Toronto's AI Future (1718.210001s)
Transcript for "Tech Talk: Inside the Toronto Data Center Market":
Good afternoon, and welcome to Tech Talk. In today's edition, we're going inside the Toronto data center market to discuss AI, data sovereignty, and Canada's next digital economy. As organizations move towards AI initiatives from experimentation to production, infrastructure strategy is becoming increasingly tied to data location, sovereignty, ecosystem connectivity, and access to innovation. Businesses are no longer asking whether AI will transform their operations. They're asking where AI should run, how data should be governed, and what infrastructure foundation will support long term growth. Joining us today, we have Rick Gallietta, principal solutions architect, and Evan Harrison, market director from our portfolio management group. Hey, Rick. How are you doing today? Doing well, Dan. Thank you. Appreciate it. Looking forward to the conversation. Thank you for joining us. Hey, Evan. How are you doing today? Doing well. Happy to be here. Thank you for being here. Rick, Evan, we visited markets across North America through this Tech Talk series, but Toronto stands apart for a few reasons. It's Canada's financial capital. It's home to the Toronto Stock Exchange. It's one of North America's fastest growing technology ecosystems, and it's become a global center for AI research and innovation. What's interesting isn't any one of these factors by itself. It's what happens when they all come together. Organizations today are trying to balance AI innovation, data governance, cloud strategy, security, and business outcomes all at the same time. Toronto sits at the center of that conversation. So what do you gentlemen think? Should we dive right into it? Yeah. Sounds great. Evan, start with you. Toronto has long been Canada's financial capital. Why is the city suddenly getting so much attention in these conversations around AI? Yeah. Great question, Dan. Vicinity has evolved beyond being just a banking and financial services hub. Toronto now sits at the intersection of several powerful trends, financial services, artificial intelligence, cloud adoption, and enterprise digital transformation. What makes Toronto unique is the concentration of innovation within a single market. You have world class universities producing AI talent, a thriving fintech ecosystem, leading AI researchers and start ups, and a deep base of enterprise customers that are actively investing in new technologies. As AI moves from experimentation to real world deployment, organizations across virtually every industry are rethinking how they operate, compete, and serve customers. Financial institutions, health care providers, technology companies, manufacturers, and government organizations are all exploring ways to leverage AI at scale. That shift is creating entirely new conversations around infrastructure, data, connectivity, and capacity. The demand we're seeing isn't limited to technology companies. AI related requirements are emerging across every sector of the economy. Toronto is particularly well positioned to benefit from this trend because it combines a highly educated workforce, a strong innovation ecosystem, and one of the largest concentrations of enterprise organizations in North America. As a result, the city is becoming an increasingly important market in the broader AI economy. Awesome. Yeah. Thank you, Evan. Yeah. Rick, let... Let's hear your perspective. Can you expand a little bit on some of the investments, that you're seeing in the market, and how is that translating into activity across the market? Yeah. Certainly. Great question, Dan. Cut... Customers quite frankly have moved beyond, basic awareness stages of AI and what we at Digital Realty call AI curious. The questions I hear now are much more operational. You know, where should the workload run? How do we protect sensitive data? How do we connect it to the right users, applications? And how do we govern it and scale? And so Toronto is differentiated because the pieces required for a strong AI market are coming together in one place. We're... As you said, world class, as Evan said, world class research, deep technology talent, major financial institutions, enterprise demand, cloud access, and organization that is serious about governance and data control. That combination makes Toronto a practical market for moving AI from experimentation and AI curious to secure governed, and production ready or, as we say, AI ready. And many of the enterprises I speak to would say, you know, they know they can benefit from the technology. They know what AI can do, but they're now asking how to operationalize private AI securely and effectively towards AI ready and to make AI reliable, governed, connected, and repeatable across the enterprise. And Toronto provides the ecosystem needed to support that. Thank you, Rick, for those insights. If I could stay on you for a second, we've we've heard a lot about data sovereignty. Why do you think these conversations are coming up in in boardrooms? Yeah. Another great question, Dan. As AI makes data location much more visible in traditional applications, data movement was often predictable. With AI, you have you have training data, prompts, retrieval databases, inference endpoints, logs, and telemetry. Each one raises questions about where the data is stored, which models, process it and where, and how to access it, and how it's governed. So for regulated industries, especially financial services, health care, government, and public sector adjacent organizations, that quickly becomes a boardroom topic because the risks are normal but purely technical. You know, they often touch compliance, brand trust, customer reputation, obligations, and business continuity. And data residency and data sovereignty are related but not identical. Residency is about where the data physically sits. Sovereignty is about which laws and jurisdictions apply, and customers need to understand both when they design AI platforms. The larger point is that AI strategy and data strategy are not... Now inseparable, and it's about privatizing AI, also known as Saba or private AI. And if you can't explain where your data lives and how it's governed, it's hard to scale AI with confidence. And organizations want more control over sensitive information, and data residency and governance are no longer optional but mandated strategic considerations. Yeah. Rick makes an important point. AI has elevated data strategy from an IT discussion to a business and board level priority. What we're seeing in conversations with Canadian organizations is a growing focus on long term data strategy. Companies aren't just asking where their data resides today. They're asking how their data architecture will support AI initiatives five and ten years from now. For many organizations, particularly those in regulated industries, there is a strong desire to keep sensitive data in country while still benefiting from speed of innovation happening across the global technology ecosystem. They want the confidence that comes with local governance and compliance, but they also want access to the latest AI capabilities and cloud services. The question is no longer whether organizations will adopt AI. The question is how they can do it in a way that aligns with their regulatory requirements, risk management frameworks, and customer commitments. That's why we're seeing data residency, governance, and infrastructure decisions become increasingly strategic. Organizations are looking for partners that can help them maintain control of their data while still enabling the flexibility and scalability needed to support AI adoption over the long term. Ultimately, the companies that can strike that balance, leveraging global innovation while maintaining strong governance and local control will be the best position to move quickly and confidently as AI continues to evolve. And the two of you make some strong points. Now Toronto sits between major financial and technology markets like New York and Chicago. Rick, why does that connectivity matter? Excellent point. You know, typically, modern businesses depend on ecosystems and not just isolated infrastructure. They need to reach cloud providers, NeoCloud, SaaS platforms, financial exchanges, business partners, customers, and internal users with performance and control. And Toronto's location and connectivity make it valuable because it can serve Canadian requirements while still giving organizations reach into major North American business metros and current such as you mentioned, New York, Chicago, and the borough ecosystems. You know, for AI, connectivity matters because the data, the model, the application, and the user are often not in the same place. And from an infrastructure perspective, customers are not just looking for compute. They need an architecture that connects AI to enterprise data, applications, clouds, hyperscalers, neo clouds, and partners across the technology stack and users in a secure and controlled way. And the architect that... The architecture has to move data securely, minimize unnecessary movement, and connect workloads to the right model and the right environment at the right time and tokenomics as we call it. You know, the customer conversations are increasingly about building an interconnected platform in private AI, infrastructure where control and governance is needed, cloud where elasticity is needed. And we say build a base and like the spike and network services and capabilities that make the whole design operate as one architecture in a private and secure way. And so private connectivity increasingly drives business agility. Yeah. I I agree with a lot of your points, Rick. From from my perspective, one of Toronto's greatest advantages is its position as a gateway between Canada and some of the world's most important financial and technology markets. Geographically and digitally, Toronto is very well connected to major US hubs like New York and Chicago. That connectivity matters because today's business operates across borders and rely on seamless access to customers, partners, cloud platforms, and financial markets. Toronto has built a very strong ecosystem around that foundation. The city combines Canada's largest concentration of financial institutions with the growing technology sector, a mature cloud adoption environment, and a deep base of enterprise organizations. Those elements reinforce one another and create a highly connected business landscape. For organizations deploying AI, modernizing applications, or expanding digital operations, connectivity is no longer just about network access. It's about being close to the people, markets, data, and ecosystems that drive business outcomes. Toronto provides that combination in a very compelling way. As a result, many organizations view Toronto not only as a key Canadian market, but as an important node with a broader North American digital economy. When organizations talk about AI, they often focus on the models. Why are infrastructure and data becoming bigger conversations today? Yeah. I will say, you know, from a frontier per... Frontier model perspective, you know, frontier models are important for the right workflows and for more efficient token management and cost efficiencies, but they're only part of the AI value chain. And the real differentiation comes from proprietary sovereign private data, domain domain expertise, governance, integration, and the ability to deliver AI for real workflows. That's why infrastructure is becoming the bigger conversation. If the data is trapped in silos, if the network path isn't efficient, if compliance is unclear, or if the compute platform can't scale, the model doesn't deliver business value. And, you know, customers also need flexibility to select and source the right frontier models for each of those workloads, manage those token, consumption costs and those costs more efficiently and effectively and access ecosystem of partners across and up and down that stack. That flexibility is difficult to achieve when data, compute and models and connectivity are designed in separate environments. And so customers are realizing that they need an AI ready foundation for high performance environments and compute intensive workloads, secure connectivity to clouds and business partners, and governance that controls and support, that supports sensitive private enterprise data. And the companies that get this right will treat infrastructure as part of the AI strategy from day one and not something to solve after a proof of concept. Again, data access, governance, and infrastructure strategy create a competitive advantage. And at the end of the day, AI outcomes depend on how effectively organizations can connect the data to applications and users. Yeah. And and that's exactly why we're seeing this conversation move well beyond the IT organization and into the boardroom. Historically, infrastructure decisions were often viewed as technical decisions. Today, they're increasingly business decisions because they directly influence how quickly an organization can adopt AI, create new products, improve customer experiences, and generate value from its data. The organizations that are leading in AI aren't necessarily those with access to the most advanced models. They're the ones that have built the right foundation to operationalize AI at scale. That means having a clear data strategy, strong governance, secure connectivity, and infrastructure that can support growth over the long term. When executives evaluate AI investments today, they're thinking about competitiveness, risk management, innovation, and speed to market. Infrastructure sits at the center of all of those priorities because it determines how effectively an organization can turn AI from a promising technology into measurable business outcomes. Ultimately, infrastructure is no longer just supporting the business. Increasingly, it's helping shape business strategy itself. Now about the question everybody's asking, what are customers actually doing with AI? Yeah. So I can grab that one then. Great question again. And, yes, the AI journey. You know, we're seeing customers move from experimentation, we call AI curious to AI ready into practical business focused AI use cases. And the first wave is often productivity. Enterprise co pilots, knowledge search, document summarization, service desk, automation, and developer productivity. In financial services, those conversations often include fraud detection, risk modeling, transaction analysis, compliance workflows, customer personalization, and internal knowledge management. And what's changing is the scale and seriousness of those deployments. You know, customers are asking how to secure the data pipeline, where inference should happen, how to control access, how to make AI available across the business without creating new compliance problems. And many customers are also thinking in stages. You know, start with lower risk productivity use cases. You know, gain some quick wins and learn how to govern them, then expand it to more mission critical AI workflows once the operating model is proven. Again, knowledge management, customer experience enhancements, fraud detection, risk analysis, and AI assisted software development. Again, quick wins, to move the program forward. Evan? Yeah. I know, personally, goofy pictures of my wife and my pets. Professionally, what I find most interesting is that we're seeing organizations move from pilot programs and isolated experiments into production deployments with clear business objectives. A year or two ago, many of the conversations were centered on what AI might be able to do. Today, customers are much more focused on how they can deploy it at scale, measure outcomes, and integrate it into day to day operations. That's an important shift because it signals growing confidence in the technology. Organizations are moving beyond testing individual use cases and are starting to think about enterprise wide adoption. They're asking how AI can improve productivity across the business, enhance customer experiences, strengthen decision making, and create sustainable competitive advantages. As that transition happens, success becomes less about the model itself and more about execution. Organizations need the right data, governance, infrastructure, and operating model to support AI across multiple business functions. What we're seeing in the market is that AI is no longer viewed as an innovation project. Increasingly, it's becoming a core business capability, and customers are investing accordingly. Sustainability and AI. Often, these these topics are not discussed together. Why do you think they should be? I'll I'll take this one, Dan. I mean, what I'm what I'm seeing is organizations today aren't choosing between innovation and sustainability. They're pursuing both simultaneously As they invest in AI to drive productivity, improve customer outcomes, and unlock new business value, they're also under increasing pressure to meet corporate ESG commitments and operate more efficiently. What we're hearing from customers is that sustainability has become a key consideration in infrastructure planning. They're evaluating not only performance, scalability, and security, but also how infrastructure choices align with their environmental goals and long term business objectives. Toronto is particularly well positioned in this regard. The market has built a strong reputation for supporting technology growth while maintaining a focus on sustainability, making it attractive to organizations that wanna scale responsibly. The most successful AI strategies will balance innovation, operational performance, and sustainability. Those priorities are becoming increasingly interconnected, which is why sustainability is now a strategic consideration in infrastructure decision making rather than a separate conversation. Yeah. I agree.... You, Evan. And... Yeah. Rick, please. Yeah. Expand on expand on what customers are looking for from a sustainability perspective and and how infrastructure providers can help organizations meet both their AI objectives and their ESG goals. Yeah. I know. Great great points, Evan and Dan. You know, AI scale and growth has a real infrastructure footprint. The compute requirements, the power densities, the cooling profiles, and utilization patterns are significantly different from many traditional enterprise workloads. And that makes sustainability part of the architecture conversation. Customers want performance, scalability, but they also need efficient design that align with previously stated and committed corporate ESG commitments, sustainability goals, and long term operating costs. You know, the best conversations are not AI versus sustainability. They're about designing AI environments intelligently by placing workflows where they make the most sense, avoiding unnecessary data movement, improving utilization, and using infrastructure partners that can support efficient growth and the ability to future poise them. In my view, sustainability, and AI is not only about the data set of facility, it's also about the architecture, choices that customers make across compute, storage, network cloud and data governance. And long term strategy requires balancing that performance efficiency and growth as well. Awesome. Yeah. Rick, maybe you can walk us through some real customer conversation scenarios that you're having today. Sure. Great great question. Great point. I I... You know, most of the conversations... You know, a very common customer conversation starts with a business team that has proven an AI use case in a pilot environment, but then realizes production is a completely different challenge. Their data may live across on premise systems, public cloud platforms, SaaS applications, and partner environments. And they wanna use that data for AI, but they need to preserve security, access control, auditability, and residency requirements and governance. And that quickly turns into an infrastructure design discussion. What data needs to stay local? What can move to the cloud, albeit, you know, hyperscalers or neos? Where should inference happen, where do we connect users' data and applications without introducing latency, risk, or complexity. And our role is to simply, you know, simplify that path by thinking through the architecture holistically, data placement, connectivity, cloud access, security resilience, and future scale instead of treating AI as just a standalone deployment. And the outcome customers want is speed with control. They wanna move faster, but not at the expense of governance, resilience, risk, or any long term flexibility. But great question, yeah. Yeah. And the conversations certainly sound complex. Indeed. Evan, Toronto organizations often operate globally. Why does a global platform matter more now than it did a few years ago? Sure. I think what's changed over the last few years is that global reach has gone from being a nice to have to a business requirement. Toronto based organizations may operate locally, but their customers, employees, partners, and even data, are increasingly distributed across multiple regions and countries. Whether they're serving financial markets, supporting global enterprises, or deploying AI enabled applications, they need infrastructure and connectivity that can operate consistently wherever business takes place. What customers are looking for is consistency. They want the same levels of performance, security, governance, and operational standards, whether they're deploying in Toronto, New York, Chicago, London, or APAC. That consistency reduces complexity, accelerates deployment timelines, and helps organizations scale more effectively. We're also seeing customers become much more focused on flexibility. As business priorities evolve, they want the ability to expand into new markets, connect to new ecosystems, and support growth without without having to rethink their entire infrastructure strategy every time. Ultimately, a global platform provides optionality. It allows organizations to align infrastructure decisions with business strategy and growth objectives rather than being constrained by geography. Rick, what do your customers expect from a global platform today and how interconnected infrastructure and ecosystems help support those increasingly global operating models? Yeah. Great question. You know, a global platform matters because customers are rarely selling for one location anymore. Their users, data applications, partners, and compliance requirements are often... Are... Span multiple markets. And if every market is designed differently, the environment becomes harder to operate, harder to secure, and slower to scale. Standardized architecture gives customers a repeatable way to expand without reinventing the model each and every time. And for a parallel based organization, that means they can keep Canadian requirements in focus while still connecting into global ecosystems and supporting business growth outside of Canada. And the value is consistency, global consistency, consistency controls, consistency connectivity, and consistent operational practices and the flexibility in the price workflows where they make the most business sense. Reducing complexity accelerates future growth. So what's the biggest infrastructure mistake organizations make when planning for AI? That's a great question. I would say that, you know, most importantly, most of them feel as if they're behind on the initiative, but I reassure them that they're not. We're in the early innings. Recognize that the wave is more than encrusted. It's more like a tsunami, and it's not a trauma. It's here. And anecdotally, most folks that we talk to, are not sure if they need a five bedroom house or a three bedroom house, but they wanna buy furniture now and recognize that the biggest mistake is treating AI as a point solution instead of an enterprise architecture decision. But, certainly, a consideration for those that operate their own data center on premise, trying to keep up with the increasing power cooling data center floor loading needs on an on premise data center is simply one that's not sustainable, continuously upgrading, saving capital into potentially aging legacy and the life on premise data center. Chip churn is accelerating. Keeping keeping up is not sustainable from a capital expense perspective. Sometimes organizations start with the model or the GPU requirement and only later ask where the data is how it will be governed, how users will access it, and how the environment will scale. That approach creates expensive rework. You may get a successful pilot, but then struggle to operationalize it because the network security data and governance and a placement strategy were not designed upfront. And so my advice is to plan backwards from the business outcome, what data is needed, who needs access, what risk must be managed, what performance is required, and what operating model will support the workload over time from that infrastructure perspective. Good advice. I mean, I... I've certainly paid the consequences of purchasing furniture too early in advance. Evan, what do you say? What do you think some of the biggest mistakes are? Yeah. Kinda echoing Rick's sentiments. If I had to point to one common mistake, it... It's waiting too long to build the foundation that AI ultimately depends on. Many organizations start with the model or the use case, which is understandable. But if the underlying data and connectivity and infrastructure strategy aren't developed in parallel, it becomes much harder to scale successful pilots into enterprise wide outcomes. The organizations gaining the most traction today are the ones treating AI as a business transformation initiative from the beginning. They're aligning infrastructure planning, data strategy, security requirements, and AI objectives early rather than trying to retrofit those capabilities later. That creates a meaningful advantage. It accelerates deployment, reduces risk, and gives organizations the flexibility to adopt new technologies as AI continues to evolve. Ultimately, AI is moving too quickly for infrastructure planning to be an afterthought. The companies that will lead over the next several years are the ones building the right foundation today. So they're ready not only for the opportunities they see now, but also for the ones that haven't emerged yet. Thank you. Great. Last question, guys. One line each. Feel free to punch it in the AI if you like. Why should organizations think about Toronto differently than they did a year ago? Yeah. I can start it here. I mean, great... That's a great question. You know, Toronto is strategically important because it's not only a source of AI innovation. It's a market where enterprises can operationalize private AI responsibly into production by bringing together infrastructure, connectivity, governance, cloud access, a strong partner ecosystem. And for us at Digital Realty, it's where we're building a foundation of AI driven business transformation and accelerating business outcomes that drive their returns on their investments. And quite simply, business outcomes equal business value. And what I say in this new world of AI is that you need to disrupt or be disruptive. Yeah. Very brief. Toronto has become the convergence point for Canada's financial technology, and AI ecosystems. Thank you, Rick. Thank you, Evan. As we bring it to a close, Toronto has always been one of Canada's most important technology and financial centers. What's changed is that AI has elevated the importance of data, connectivity, and ecosystem access. Organizations are realizing that AI success is driven as much by infrastructure strategy as it is by the model itself. As we've heard today, the conversation is no longer simply about where where workloads run. It's about where data lives, how organizations connect to partners and clouds, and how they create a foundation for long term innovation. Thank you, Rick and Evan, for sharing their insights, and we thank you for joining, Tech Talk Toronto. As we wrap up the session, I'd like to mention there is a meet the team button, for the guests to click to be connected to, one of our team members. We also welcome you to a, visualizer, demo highlighting Toronto, metro connectivity and ecosystem. Or if you'd like, we'd be happy to take you around a tour around one of our data centers here in Toronto. Thank you for joining us today. See you at the next Tech Talk. Thanks.