Ai

May 15, 2026

Exclusive: AHOY Thinks Montreal Is the Unlikely Capital of a New Kind of AI, Here's Why They're Right

AHOY, the deep-tech infrastructure company building the operational layer for Real-World AI, at Web Summit Vancouver and made a series of announcements that would have sounded like science fiction five years ago: $100 million in profitable annual recurring revenue, 330 times growth since 2021, and a nine-figure acquisition of a Canadian automation company called Wrk Technologies. And a new global headquarters, not in San Francisco, not in London, but in Montreal, Quebec.

by Kasun Illankoon, Editor in Chief at Tech Revolt

If you haven't heard of AHOY, that's part of the story. The company has spent years operating in the quieter corridors of government procurement and critical infrastructure, building AI systems for airports, water utilities, and various environments in roughly 40 countries. It doesn't make chatbots. It doesn't generate images. What it does, running AI at the edge, on-chip, in environments that can never touch the internet, is increasingly the most consequential question in the entire field.

And Montreal, it turns out, is exactly where that question is being answered.

What "Physical AI" Actually Means

The phrase gets thrown around a lot now, but AHOY has been building toward it since 2018. The idea is deceptively simple: most AI lives in the cloud. It takes your input, sends it to a server farm somewhere, processes it, and sends back a result. That works beautifully for writing an email or summarising a Zoom call. It falls apart completely if you're trying to make a real-time decision about a runway incursion at a major airport, or detect a pressure anomaly in a water distribution network before it becomes a rupture.

Physical AI, as AHOY defines it, is intelligence that perceives, decides, and acts on the physical world, under real-world latency, compliance, and resilience constraints. No cloud hop. No round-trip to a data centre. The system processes data where the data lives: on the chip, at the edge, on the premises, or in fully air-gapped environments with zero external network connection.

The production numbers AHOY has reported from its live deployments are striking: roughly a 30 percent increase in throughput in aviation environments, around a 40 percent reduction in losses in water distribution systems, and approximately a 60 percent improvement in safety enforcement alongside close to a 30 percent reduction in operational expenditure. These aren't benchmarks run in controlled conditions. They're outcomes already running in real systems, in real cities, in real countries.

The company calls its broader platform the "operational layer for Real-World AI", a stack that ingests live data from cameras, microphones, sensors, and telemetry systems, fuses it together, and turns it into operational decisions, all without sending a single byte to a foreign cloud.

That last clause is where the geopolitics begin.

The Sovereignty Problem Nobody Wanted to Name

For years, Canada's relationship with American cloud infrastructure was treated as a feature, not a bug. AWS, Microsoft Azure, and Google Cloud provided world-class computing at a price no domestic provider could match. Why build your own when you could rent theirs?

The answer arrived slowly, then all at once. In 2023, a Microsoft executive told a French Senate committee something that reverberated quietly through policy circles across the Western world: American technology companies would, when pressed, defer to the US government, regardless of what foreign laws said. The cloud you're renting isn't neutral. It belongs to someone, and that someone has a government, and that government has interests.

For Canada, this landed in a particularly uncomfortable way. The country's researchers, hospitals, public agencies, and security services had built enormous dependencies on infrastructure they didn't control, couldn't audit end-to-end, and couldn't guarantee would remain accessible if the bilateral relationship with the United States became more complicated. That relationship, as of 2026, has become more complicated.

Ottawa's response has been pointed. A Sovereign AI Compute Strategy announced in 2025 and 2026 committed up to $2 billion over five years to build publicly owned Canadian supercomputing infrastructure. Public consultations shaping the country's forthcoming national AI strategy produced, among thousands of submissions, a clear consensus: Canadians want domestic computers, sovereign data governance, and reduced dependence on foreign-controlled systems. The government's own language has grown less diplomatic: "Canadians want AI that is safe and sovereign."

This is the environment AHOY walked into. More precisely, it's the environment AHOY was built for, before Canada had fully articulated the problem.

Why Montreal, and Why Mila

Montreal's claim to AI leadership is older and deeper than most people realise. The city has been a serious node in the global AI research network since the 1980s, when Yoshua Bengio — now a Turing Award winner and widely regarded as one of the godfathers of modern deep learning, began building what would eventually become Mila, the Quebec Artificial Intelligence Institute.

Mila has grown into the world's largest university-based deep learning research centre. It draws more than 140 affiliated professors from institutions including Université de Montréal and McGill University, and hosts over 1,400 researchers in total. Its open-science model, publishing research rather than hoarding it, has made it a different kind of institution than the closed AI labs attached to large technology companies. It produces foundational research and, crucially, a continuous stream of deployable innovation and deployable people.

AHOY's strategic partnership with Mila is the centrepiece of its Montreal announcement, and it goes well beyond a logo on a press release. The two organisations are launching joint research programmes focused specifically on edge AI, sovereign AI, and Physical AI, with a dedicated PhD cohort working on the architectural challenges AHOY is solving in production. That last phrase matters: this is applied research rooted in real systems, not theoretical exercises.

To lead it, AHOY made two senior appointments that signal just how seriously it is treating the Montreal operation. Dr. Hood Khizer joins as Chief Scientific Officer, taking ownership of AHOY's frontier research roadmap, including the company's internal R&D work through its Trouve Labs division and its expanded engagement with Mila's faculty.

Dr. Benoit Julien comes in as Chief Technologist, responsible for the developer and operator surface through which AHOY's Real-World AI stack is configured, deployed, and operated in the field. Together they bring scientific depth and engineering rigour directly into the Montreal headquarters, and will co-lead the company's ongoing relationship with Mila's research community. These are not symbolic hires parachuted in for a ribbon-cutting. They are the people responsible for making the partnership produce something.

Alongside the Mila partnership, AHOY is funding what it describes as Canada's first Physical AI Lab, a research facility dedicated to on-site work for on-chip, on-premises, and air-gapped systems. The lab is designed to open AHOY's research to the broader academic and developer community, turning a private R&D capability into a shared national resource. It is, in effect, an argument made in concrete and funding: that the next phase of AI research doesn't have to happen inside a hyperscaler's campus.

The Wrk Acquisition and What Comes Next

AHOY's announcement also included the acquisition of Wrk Technologies, a Montreal-based automation and execution platform. The deal, valued in the nine figures, brings Wrk's mature execution layer, built for APIs, legacy systems, AI models, and human-in-the-loop workflows, directly into AHOY's orchestration and perception stack.

The strategic logic is straightforward: AHOY's platform can perceive and decide; Wrk's platform can execute. Together they form a pipeline from sensor input to auditable action, without requiring customers to stitch together components from different vendors, each with their own cloud dependencies and data egress.

Jamil Shinawi, AHOY's CEO, framed the acquisition plainly: "Wrk is a force multiplier for our mission. Combining Wrk's execution engine with AHOY's orchestration, operational research and sovereign tooling lets customers move from design to production faster, more reliably, and without vendor lock in."

Mohannad El Barachi, CEO of Wrk Technologies and Chief Strategy Officer at AHOY, pointed to the shared philosophy as the basis for the deal: "Our plans to join AHOY are largely tied to a shared philosophy: decentralized, sovereign first AI built for real systems. Together we accelerate a future where intelligence is distributed, locally governed, and operationally reliable."

The Argument Being Made

Underneath the financing announcements and the headcount numbers is something worth taking seriously as an idea. The dominant narrative in AI over the past several years has been one of consolidation: larger models, more compute, more cloud, controlled by a smaller number of increasingly powerful companies. That narrative has produced genuine wonders. It has also produced a world in which most frontier AI capability is accessible to governments, cities, and critical infrastructure operators only on terms set by a handful of American corporations.

AHOY's thesis, the one its entire architecture was designed around, is that a specific and important category of AI cannot operate on those terms. The AI that runs airports and water systems and emergency response networks needs to run where the operations are, governed by the people responsible for those operations, with full auditability and no dependency on a data centre that could be switched off, subpoenaed, or simply made unavailable.

Shinawi captured the company's positioning with unusual clarity: "The market is separating into systems that can only describe and systems that can reliably act, and Real World AI lives on the second side of that line. Today's announcements are how that thesis becomes a global rollout. Montreal anchors our North American base. The partnership with Mila compounds the research. Canada's first Physical AI Lab opens that research to the broader community. And the Perception Suite puts the tooling directly in the hands of operators, governments, and enterprises. Each step extends sovereign infrastructure into a new region, a new layer of the stack, and a new set of builders who need to deploy it."

He added something that reads less like marketing and more like a philosophical position: "AHOY is the only company in the world that has done the research and is shipping the tools that let customers build their own sovereign version of Real World AI. That is what democratizing the next phase of AI actually looks like."

Canada, in 2026, is asking the right questions about who controls its digital infrastructure. AHOY's move to Montreal is a bet that the country is ready to act on the answers, and that the world is watching to see what that looks like in practice.

Exclusive: AHOY Thinks Montreal Is the Unlikely Capital of a New Kind of AI, Here's Why They're Right

AHOY, the deep-tech infrastructure company building the operational layer for Real-World AI, at Web Summit Vancouver and made a series of announcements that would have sounded like science fiction five years ago: $100 million in profitable annual recurring revenue, 330 times growth since 2021, and a nine-figure acquisition of a Canadian automation company called Wrk Technologies. And a new global headquarters, not in San Francisco, not in London, but in Montreal, Quebec.

by Kasun Illankoon, Editor in Chief at Tech Revolt

If you haven't heard of AHOY, that's part of the story. The company has spent years operating in the quieter corridors of government procurement and critical infrastructure, building AI systems for airports, water utilities, and various environments in roughly 40 countries. It doesn't make chatbots. It doesn't generate images. What it does, running AI at the edge, on-chip, in environments that can never touch the internet, is increasingly the most consequential question in the entire field.

And Montreal, it turns out, is exactly where that question is being answered.

What "Physical AI" Actually Means

The phrase gets thrown around a lot now, but AHOY has been building toward it since 2018. The idea is deceptively simple: most AI lives in the cloud. It takes your input, sends it to a server farm somewhere, processes it, and sends back a result. That works beautifully for writing an email or summarising a Zoom call. It falls apart completely if you're trying to make a real-time decision about a runway incursion at a major airport, or detect a pressure anomaly in a water distribution network before it becomes a rupture.

Physical AI, as AHOY defines it, is intelligence that perceives, decides, and acts on the physical world, under real-world latency, compliance, and resilience constraints. No cloud hop. No round-trip to a data centre. The system processes data where the data lives: on the chip, at the edge, on the premises, or in fully air-gapped environments with zero external network connection.

The production numbers AHOY has reported from its live deployments are striking: roughly a 30 percent increase in throughput in aviation environments, around a 40 percent reduction in losses in water distribution systems, and approximately a 60 percent improvement in safety enforcement alongside close to a 30 percent reduction in operational expenditure. These aren't benchmarks run in controlled conditions. They're outcomes already running in real systems, in real cities, in real countries.

The company calls its broader platform the "operational layer for Real-World AI", a stack that ingests live data from cameras, microphones, sensors, and telemetry systems, fuses it together, and turns it into operational decisions, all without sending a single byte to a foreign cloud.

That last clause is where the geopolitics begin.

The Sovereignty Problem Nobody Wanted to Name

For years, Canada's relationship with American cloud infrastructure was treated as a feature, not a bug. AWS, Microsoft Azure, and Google Cloud provided world-class computing at a price no domestic provider could match. Why build your own when you could rent theirs?

The answer arrived slowly, then all at once. In 2023, a Microsoft executive told a French Senate committee something that reverberated quietly through policy circles across the Western world: American technology companies would, when pressed, defer to the US government, regardless of what foreign laws said. The cloud you're renting isn't neutral. It belongs to someone, and that someone has a government, and that government has interests.

For Canada, this landed in a particularly uncomfortable way. The country's researchers, hospitals, public agencies, and security services had built enormous dependencies on infrastructure they didn't control, couldn't audit end-to-end, and couldn't guarantee would remain accessible if the bilateral relationship with the United States became more complicated. That relationship, as of 2026, has become more complicated.

Ottawa's response has been pointed. A Sovereign AI Compute Strategy announced in 2025 and 2026 committed up to $2 billion over five years to build publicly owned Canadian supercomputing infrastructure. Public consultations shaping the country's forthcoming national AI strategy produced, among thousands of submissions, a clear consensus: Canadians want domestic computers, sovereign data governance, and reduced dependence on foreign-controlled systems. The government's own language has grown less diplomatic: "Canadians want AI that is safe and sovereign."

This is the environment AHOY walked into. More precisely, it's the environment AHOY was built for, before Canada had fully articulated the problem.

Why Montreal, and Why Mila

Montreal's claim to AI leadership is older and deeper than most people realise. The city has been a serious node in the global AI research network since the 1980s, when Yoshua Bengio — now a Turing Award winner and widely regarded as one of the godfathers of modern deep learning, began building what would eventually become Mila, the Quebec Artificial Intelligence Institute.

Mila has grown into the world's largest university-based deep learning research centre. It draws more than 140 affiliated professors from institutions including Université de Montréal and McGill University, and hosts over 1,400 researchers in total. Its open-science model, publishing research rather than hoarding it, has made it a different kind of institution than the closed AI labs attached to large technology companies. It produces foundational research and, crucially, a continuous stream of deployable innovation and deployable people.

AHOY's strategic partnership with Mila is the centrepiece of its Montreal announcement, and it goes well beyond a logo on a press release. The two organisations are launching joint research programmes focused specifically on edge AI, sovereign AI, and Physical AI, with a dedicated PhD cohort working on the architectural challenges AHOY is solving in production. That last phrase matters: this is applied research rooted in real systems, not theoretical exercises.

To lead it, AHOY made two senior appointments that signal just how seriously it is treating the Montreal operation. Dr. Hood Khizer joins as Chief Scientific Officer, taking ownership of AHOY's frontier research roadmap, including the company's internal R&D work through its Trouve Labs division and its expanded engagement with Mila's faculty.

Dr. Benoit Julien comes in as Chief Technologist, responsible for the developer and operator surface through which AHOY's Real-World AI stack is configured, deployed, and operated in the field. Together they bring scientific depth and engineering rigour directly into the Montreal headquarters, and will co-lead the company's ongoing relationship with Mila's research community. These are not symbolic hires parachuted in for a ribbon-cutting. They are the people responsible for making the partnership produce something.

Alongside the Mila partnership, AHOY is funding what it describes as Canada's first Physical AI Lab, a research facility dedicated to on-site work for on-chip, on-premises, and air-gapped systems. The lab is designed to open AHOY's research to the broader academic and developer community, turning a private R&D capability into a shared national resource. It is, in effect, an argument made in concrete and funding: that the next phase of AI research doesn't have to happen inside a hyperscaler's campus.

The Wrk Acquisition and What Comes Next

AHOY's announcement also included the acquisition of Wrk Technologies, a Montreal-based automation and execution platform. The deal, valued in the nine figures, brings Wrk's mature execution layer, built for APIs, legacy systems, AI models, and human-in-the-loop workflows, directly into AHOY's orchestration and perception stack.

The strategic logic is straightforward: AHOY's platform can perceive and decide; Wrk's platform can execute. Together they form a pipeline from sensor input to auditable action, without requiring customers to stitch together components from different vendors, each with their own cloud dependencies and data egress.

Jamil Shinawi, AHOY's CEO, framed the acquisition plainly: "Wrk is a force multiplier for our mission. Combining Wrk's execution engine with AHOY's orchestration, operational research and sovereign tooling lets customers move from design to production faster, more reliably, and without vendor lock in."

Mohannad El Barachi, CEO of Wrk Technologies and Chief Strategy Officer at AHOY, pointed to the shared philosophy as the basis for the deal: "Our plans to join AHOY are largely tied to a shared philosophy: decentralized, sovereign first AI built for real systems. Together we accelerate a future where intelligence is distributed, locally governed, and operationally reliable."

The Argument Being Made

Underneath the financing announcements and the headcount numbers is something worth taking seriously as an idea. The dominant narrative in AI over the past several years has been one of consolidation: larger models, more compute, more cloud, controlled by a smaller number of increasingly powerful companies. That narrative has produced genuine wonders. It has also produced a world in which most frontier AI capability is accessible to governments, cities, and critical infrastructure operators only on terms set by a handful of American corporations.

AHOY's thesis, the one its entire architecture was designed around, is that a specific and important category of AI cannot operate on those terms. The AI that runs airports and water systems and emergency response networks needs to run where the operations are, governed by the people responsible for those operations, with full auditability and no dependency on a data centre that could be switched off, subpoenaed, or simply made unavailable.

Shinawi captured the company's positioning with unusual clarity: "The market is separating into systems that can only describe and systems that can reliably act, and Real World AI lives on the second side of that line. Today's announcements are how that thesis becomes a global rollout. Montreal anchors our North American base. The partnership with Mila compounds the research. Canada's first Physical AI Lab opens that research to the broader community. And the Perception Suite puts the tooling directly in the hands of operators, governments, and enterprises. Each step extends sovereign infrastructure into a new region, a new layer of the stack, and a new set of builders who need to deploy it."

He added something that reads less like marketing and more like a philosophical position: "AHOY is the only company in the world that has done the research and is shipping the tools that let customers build their own sovereign version of Real World AI. That is what democratizing the next phase of AI actually looks like."

Canada, in 2026, is asking the right questions about who controls its digital infrastructure. AHOY's move to Montreal is a bet that the country is ready to act on the answers, and that the world is watching to see what that looks like in practice.

Exclusive: AHOY Thinks Montreal Is the Unlikely Capital of a New Kind of AI, Here's Why They're Right

AHOY, the deep-tech infrastructure company building the operational layer for Real-World AI, at Web Summit Vancouver and made a series of announcements that would have sounded like science fiction five years ago: $100 million in profitable annual recurring revenue, 330 times growth since 2021, and a nine-figure acquisition of a Canadian automation company called Wrk Technologies. And a new global headquarters, not in San Francisco, not in London, but in Montreal, Quebec.

by Kasun Illankoon, Editor in Chief at Tech Revolt

If you haven't heard of AHOY, that's part of the story. The company has spent years operating in the quieter corridors of government procurement and critical infrastructure, building AI systems for airports, water utilities, and various environments in roughly 40 countries. It doesn't make chatbots. It doesn't generate images. What it does, running AI at the edge, on-chip, in environments that can never touch the internet, is increasingly the most consequential question in the entire field.

And Montreal, it turns out, is exactly where that question is being answered.

What "Physical AI" Actually Means

The phrase gets thrown around a lot now, but AHOY has been building toward it since 2018. The idea is deceptively simple: most AI lives in the cloud. It takes your input, sends it to a server farm somewhere, processes it, and sends back a result. That works beautifully for writing an email or summarising a Zoom call. It falls apart completely if you're trying to make a real-time decision about a runway incursion at a major airport, or detect a pressure anomaly in a water distribution network before it becomes a rupture.

Physical AI, as AHOY defines it, is intelligence that perceives, decides, and acts on the physical world, under real-world latency, compliance, and resilience constraints. No cloud hop. No round-trip to a data centre. The system processes data where the data lives: on the chip, at the edge, on the premises, or in fully air-gapped environments with zero external network connection.

The production numbers AHOY has reported from its live deployments are striking: roughly a 30 percent increase in throughput in aviation environments, around a 40 percent reduction in losses in water distribution systems, and approximately a 60 percent improvement in safety enforcement alongside close to a 30 percent reduction in operational expenditure. These aren't benchmarks run in controlled conditions. They're outcomes already running in real systems, in real cities, in real countries.

The company calls its broader platform the "operational layer for Real-World AI", a stack that ingests live data from cameras, microphones, sensors, and telemetry systems, fuses it together, and turns it into operational decisions, all without sending a single byte to a foreign cloud.

That last clause is where the geopolitics begin.

The Sovereignty Problem Nobody Wanted to Name

For years, Canada's relationship with American cloud infrastructure was treated as a feature, not a bug. AWS, Microsoft Azure, and Google Cloud provided world-class computing at a price no domestic provider could match. Why build your own when you could rent theirs?

The answer arrived slowly, then all at once. In 2023, a Microsoft executive told a French Senate committee something that reverberated quietly through policy circles across the Western world: American technology companies would, when pressed, defer to the US government, regardless of what foreign laws said. The cloud you're renting isn't neutral. It belongs to someone, and that someone has a government, and that government has interests.

For Canada, this landed in a particularly uncomfortable way. The country's researchers, hospitals, public agencies, and security services had built enormous dependencies on infrastructure they didn't control, couldn't audit end-to-end, and couldn't guarantee would remain accessible if the bilateral relationship with the United States became more complicated. That relationship, as of 2026, has become more complicated.

Ottawa's response has been pointed. A Sovereign AI Compute Strategy announced in 2025 and 2026 committed up to $2 billion over five years to build publicly owned Canadian supercomputing infrastructure. Public consultations shaping the country's forthcoming national AI strategy produced, among thousands of submissions, a clear consensus: Canadians want domestic computers, sovereign data governance, and reduced dependence on foreign-controlled systems. The government's own language has grown less diplomatic: "Canadians want AI that is safe and sovereign."

This is the environment AHOY walked into. More precisely, it's the environment AHOY was built for, before Canada had fully articulated the problem.

Why Montreal, and Why Mila

Montreal's claim to AI leadership is older and deeper than most people realise. The city has been a serious node in the global AI research network since the 1980s, when Yoshua Bengio — now a Turing Award winner and widely regarded as one of the godfathers of modern deep learning, began building what would eventually become Mila, the Quebec Artificial Intelligence Institute.

Mila has grown into the world's largest university-based deep learning research centre. It draws more than 140 affiliated professors from institutions including Université de Montréal and McGill University, and hosts over 1,400 researchers in total. Its open-science model, publishing research rather than hoarding it, has made it a different kind of institution than the closed AI labs attached to large technology companies. It produces foundational research and, crucially, a continuous stream of deployable innovation and deployable people.

AHOY's strategic partnership with Mila is the centrepiece of its Montreal announcement, and it goes well beyond a logo on a press release. The two organisations are launching joint research programmes focused specifically on edge AI, sovereign AI, and Physical AI, with a dedicated PhD cohort working on the architectural challenges AHOY is solving in production. That last phrase matters: this is applied research rooted in real systems, not theoretical exercises.

To lead it, AHOY made two senior appointments that signal just how seriously it is treating the Montreal operation. Dr. Hood Khizer joins as Chief Scientific Officer, taking ownership of AHOY's frontier research roadmap, including the company's internal R&D work through its Trouve Labs division and its expanded engagement with Mila's faculty.

Dr. Benoit Julien comes in as Chief Technologist, responsible for the developer and operator surface through which AHOY's Real-World AI stack is configured, deployed, and operated in the field. Together they bring scientific depth and engineering rigour directly into the Montreal headquarters, and will co-lead the company's ongoing relationship with Mila's research community. These are not symbolic hires parachuted in for a ribbon-cutting. They are the people responsible for making the partnership produce something.

Alongside the Mila partnership, AHOY is funding what it describes as Canada's first Physical AI Lab, a research facility dedicated to on-site work for on-chip, on-premises, and air-gapped systems. The lab is designed to open AHOY's research to the broader academic and developer community, turning a private R&D capability into a shared national resource. It is, in effect, an argument made in concrete and funding: that the next phase of AI research doesn't have to happen inside a hyperscaler's campus.

The Wrk Acquisition and What Comes Next

AHOY's announcement also included the acquisition of Wrk Technologies, a Montreal-based automation and execution platform. The deal, valued in the nine figures, brings Wrk's mature execution layer, built for APIs, legacy systems, AI models, and human-in-the-loop workflows, directly into AHOY's orchestration and perception stack.

The strategic logic is straightforward: AHOY's platform can perceive and decide; Wrk's platform can execute. Together they form a pipeline from sensor input to auditable action, without requiring customers to stitch together components from different vendors, each with their own cloud dependencies and data egress.

Jamil Shinawi, AHOY's CEO, framed the acquisition plainly: "Wrk is a force multiplier for our mission. Combining Wrk's execution engine with AHOY's orchestration, operational research and sovereign tooling lets customers move from design to production faster, more reliably, and without vendor lock in."

Mohannad El Barachi, CEO of Wrk Technologies and Chief Strategy Officer at AHOY, pointed to the shared philosophy as the basis for the deal: "Our plans to join AHOY are largely tied to a shared philosophy: decentralized, sovereign first AI built for real systems. Together we accelerate a future where intelligence is distributed, locally governed, and operationally reliable."

The Argument Being Made

Underneath the financing announcements and the headcount numbers is something worth taking seriously as an idea. The dominant narrative in AI over the past several years has been one of consolidation: larger models, more compute, more cloud, controlled by a smaller number of increasingly powerful companies. That narrative has produced genuine wonders. It has also produced a world in which most frontier AI capability is accessible to governments, cities, and critical infrastructure operators only on terms set by a handful of American corporations.

AHOY's thesis, the one its entire architecture was designed around, is that a specific and important category of AI cannot operate on those terms. The AI that runs airports and water systems and emergency response networks needs to run where the operations are, governed by the people responsible for those operations, with full auditability and no dependency on a data centre that could be switched off, subpoenaed, or simply made unavailable.

Shinawi captured the company's positioning with unusual clarity: "The market is separating into systems that can only describe and systems that can reliably act, and Real World AI lives on the second side of that line. Today's announcements are how that thesis becomes a global rollout. Montreal anchors our North American base. The partnership with Mila compounds the research. Canada's first Physical AI Lab opens that research to the broader community. And the Perception Suite puts the tooling directly in the hands of operators, governments, and enterprises. Each step extends sovereign infrastructure into a new region, a new layer of the stack, and a new set of builders who need to deploy it."

He added something that reads less like marketing and more like a philosophical position: "AHOY is the only company in the world that has done the research and is shipping the tools that let customers build their own sovereign version of Real World AI. That is what democratizing the next phase of AI actually looks like."

Canada, in 2026, is asking the right questions about who controls its digital infrastructure. AHOY's move to Montreal is a bet that the country is ready to act on the answers, and that the world is watching to see what that looks like in practice.