Earlier this month, Garry Chan represented Ontario’s hardtech and AI incubator ventureLAB at the Global Partnership on AI Summit in Belgrade, Serbia, which convened from science, public policy, and industry to discuss the future of artificial intelligence. 

While the conference drew leading voices from government, academia, and Big Tech, Chan was there representing the perspective of early-stage startups.

“A lot of founders are doing the zero-to-one stage well, and we need to support that. But the real gap, and the big opportunity, is going from one to 100.”

Chan is the Chief AI Advisor at ventureLAB, offering insights from a career that spans entrepreneurship, advising startups, facilitating commercialization, and scaling tech businesses globally.

At ventureLAB, he aims to lead where AI intersects with hardware innovation. The incubator integrates expertise in semiconductors and hardware to help startups commercialize and scale AI-driven technologies.

Their state-of-the-art AI Compute Lab provides startups with the tools to test, optimize, and de-risk their solutions. Coupled with its AI commercialization approach for real-world testing and collaboration, ventureLAB seeks to accelerate the convergence of AI and hardware to drive breakthroughs in manufacturing, healthcare, and transportation.

From that vantage point, Chan believes that early-stage startups are driving some of the most groundbreaking advancements in AI, while operating under immense constraints and navigating a lack of resources, infrastructure, and governance support.

Chan sat down with BetaKit to discuss the role of early-stage companies within Canada’s AI ecosystem, how AI governance can be built into innovation from day one, and what Canada needs to do if it wants to become a global leader in the field.

The following Q&A has been edited for brevity and clarity.

You recently attended the GPAI Summit in Belgrade. What did you take away from the event?

I was on a panel during the two days that focused on partnerships between academia, industry, and government—looking at AI from multiple dimensions. It was fascinating to see perspectives from academics, some from the US, others from Canada, alongside people like me, who work within the early-stage startup ecosystem.

From my perspective, we need to figure out how to put all these pieces together. AI governance and the ethical use of AI are critical. It’s not just something abstract you read about in journals or hear from thought leaders. It’s something we can and should infuse into early-stage companies right from the start.

We may be thinking about AI governance at the government, institutional, or enterprise level, but I feel there’s clear applicability even within early-stage companies.

Why do you think it’s important to include early-stage companies in conversations about the future of AI?

Early-stage startups play an incredibly important role in this conversation, especially in a place like Canada. By nature, they are disruptive, they’re doing something new, something others may not have done or even thought of before. If they’re not at the table, we lose out on understanding what’s happening at the forefront of the technology. 

Canada is also a country of startups, and as a country that serves small businesses and is the birthplace of amazing companies, we have a special responsibility to support these innovators and entrepreneurs who are pushing the boundaries.

What does that look like in practice for early-stage companies?

I liken it to raising a child. You try to teach good habits early. Startups should learn to bake ethical AI, safety, and governance into their DNA from day one, so as they scale, they don’t suddenly have to go back to say, ‘What do we do now?’

If you’re an early-stage company living paycheck to paycheck, you know that financial governance is critical. If you don’t manage cash flow, you won’t exist next month. Companies understand this because financial management has been ingrained over centuries. But with AI governance, we’re just at the tip of the iceberg.

The best way to approach it is to start with data governance. Many software and AI-powered companies serve customers globally. Without the right structures, checklists, and safeguards in place, you can’t operate—whether in Europe, the US, or elsewhere. Issues like data residency, privacy, protection, and security are foundational, and AI governance is a natural extension of this.

Now, for early-stage companies, what does this mean? Imagine you’re productizing your offering. You might start with one customer and one dataset in Canada. But in 12 months, you’ll likely need to operate in Europe or the US. That requires building considerations for data governance, AI infrastructure, and scalability right from the start.

How is ventureLAB looking to address some of these challenges through its AI programming?

We’re tackling three core challenges through our Accelerate AI program. The first is our Deep Learning Training and Inference Program, where we help companies productize their solutions. Building a piece of software often involves lots of smart people, like PhDs, but what we aim to do is bring in a deep bench of advisors, which we call executives-in-residence. These are people with decades of experience who might not be super technical, and that’s a good thing, because you want to pair deep engineering expertise with business-savvy people.

The second is our High-Performance Computing and Data Analytics support, which is tied to our AI compute capabilities. Thanks to funding from multiple levels of government, we’ve built a hardware lab with a server farm of GPUs. This gives startups access to the resources they need to train their software in a secure, stable environment without worrying about huge costs or chasing cloud credits.

Finally, we have the Enterprise-Ready AI Solutions Program, which focuses on commercialization. So, once your product is ready, we want to create a collision of the potential buyers and users of your tech and the solution providers. We want to make sure that we have a safe platform where people can try out the technology, where they can experiment with your tech, and where startups can co-sell and potentially resell each other’s solutions.

What do you think is the biggest opportunity for AI in Canada?

The biggest opportunity is to do everything we can to help companies scale. A lot of founders are doing the zero-to-one stage well, and we need to support that. But the real gap, and the big opportunity, is going from one to 100, helping founders find customers and helping companies scale.

Our mission is clear: to drive transformative growth and position Canada at the forefront of the global AI economy.

How does compute pose a challenge for companies as they scale?

AI is very data and processing-intensive, which means you need a lot of horsepower. If you’re a university researcher, you might have access to some compute power. If you’re an established enterprise, there are plenty of commercial providers like AWS and Google Cloud. But for early-stage startups, you’re kind of in between. Many founders end up hopping from one vendor to the next, relying on incentives like cloud credits.

Imagine, you’re always shopping for the next month’s cloud credits to do your data inference or machine learning, and therefore you don’t have time to focus on your business, which is really what you should be doing. Imagine if startups had a stable environment or sandbox where they could experiment securely, build their AI, and position themselves to delight their customers.

What kind of role do you see ventureLAB’s programming playing in Canada’s AI ecosystem?

Looking back on when I was building my startup, I often wish I had a ventureLAB in my life. There are so many advantages—access to compute resources, product design support, scaling, and commercialization.

But I think there’s something that goes beyond that: having a community of 130 smart founders, mentors, and executives-in-residence all supporting each other. You could argue it’s a chaotic environment, but it’s also a very vibrant one where you don’t need to go very far to find someone to challenge you on your product or your commercialization strategy.

I spent almost 10 years in Silicon Valley, where people say you can find all the talent, money, and customers you need within an hour’s drive. We want to do something like that here as well. If we can create a vibrant community where all these different parties have a role to play, I think something amazing will come out of it.

Visit ventureLAB.ca to learn more about how we empower hardtech founders to build and scale globally competitive ventures, advancing Canada’s knowledge-based economy.

Feature image provided by ventureLAB.

The post Global AI gets a startup perspective: Insights from ventureLAB’s Garry Chan first appeared on BetaKit.