All figures as of early September 2026. I own Shopify and Constellation Software. This is analysis, not personalized advice. This article is an AI-generated article using the transcript of the video below. The video is the source of truth.
In 2026, the market went through Canadian tech with a label maker. AI winner. AI victim. Winner. Victim. Then it repriced all six of the biggest names in the country based on those labels, and five of the six now trade below their 52-week highs. Some of those labels are right. Some of them are badly wrong, and that gap is where the opportunity sits.
Before you assume you need US tech stocks to build wealth, look at the last decade. Two of these six companies have beaten the Nasdaq over ten years, and not by a little. A third matched the index almost exactly. You did not need Silicon Valley for world-class returns.
So today I’m ranking the six giants of Canadian tech from S tier to E tier: Shopify (SHOP), Constellation Software (CSU), Celestica (CLS), CGI (GIB.A), Descartes Systems (DSG) and OpenText (OTEX). One per tier, worst to first. I own two of these six personally, and I’ll tell you exactly which two.
Tech 101: three rules for judging these six
Rule one: for most of this board, the P/E ratio lies. Shopify’s reported earnings are polluted by gains on its investments. Constellation’s are distorted by acquisition accounting, and even the company tells you to ignore them. Descartes carries amortization from 25 years of deals. And Celestica is a cyclical manufacturer, and cyclical earnings always look cheapest right at the top of the cycle. For those four, use free cash flow. P/E is really only usable for CGI.
Rule two: separate organic growth from acquired growth. Three of these six buy a lot of their growth. Nothing wrong with that, but you have to check the health reading underneath: what does the existing business grow on its own? For one company below, that number is the entire debate.
Rule three: compare each stock to its own history. Several of these are the cheapest they have been in a decade against their own past. Cheap versus your own history is only a bargain if the business is not broken. Figuring out which ones are broken is basically this whole article.
E tier: OpenText (OTEX), deep value or value trap?
OpenText is one of Canada’s oldest software companies, born at the University of Waterloo in 1991, built through 80 plus acquisitions into a giant of boring but essential software: the systems banks, hospitals and governments use to store, secure and move documents and data. It was even one of the only Canadian tech stocks with a real dividend growth streak.
The dividend did not save anyone. The last mega deal, Micro Focus, was one of the worst acquisitions in the company’s history at the worst possible time, and it left roughly US$5 billion of net debt against a market cap of about US$5.7 billion. They owe almost as much as the entire company is worth.
Here is the tempting part. OpenText converts revenue to cash beautifully: roughly US$1.1 billion of free cash flow a year on 36% EBITDA margins. That is a 14% free cash flow yield, a 4.7% dividend that was just raised, a buyback authorized for 10% of the float, and a stock at about 6 times forward earnings while software peers trade above 20. Add it up and you are being paid close to 10% a year to wait.
Now the trap side. The biggest segment shrinks about 2% a year. Guidance calls for revenue down again in 2027, and the market never rewards a shrinking top line with a bigger multiple. The roll-up is literally running in reverse: Vertica and the mainframe unit have been sold to pay down debt. And of everything on this list, legacy document management is the software I think AI disrupts most directly. Sticky revenue is not the same as growing revenue. If you cannot raise prices and cannot add customers, high switching costs are just a lock on a leaking boat.
A payout is not a return. E tier.
D tier: CGI (GIB.A), high quality meets a real bear case
CGI used to be what I called the blue chip of Canadian tech. About 94,000 consultants build and run the IT systems behind everyday life: provincial health records, government tax platforms, bank back offices. Over half the revenue comes from managed services contracts that run five to ten years, clients renew more than 95% of the time, and the backlog sits at C$31.8 billion, nearly two full years of revenue already under contract.
It is also the cheapest it has been in a decade: about 12.6 times trailing earnings versus a ten year average around 19.5, with a free cash flow yield near 12%. Management clearly thinks the stock is a gift, because relentless buybacks turned 14% profit growth into 22% earnings per share growth last quarter.
The reason it is cheap has two letters. The market believes AI is coming for IT consultants. And here is where I differ from the “every wave feeds the consultants” crowd: cloud, ERP and the internet were all supposed to kill IT services and fed them instead, and CGI’s AI pipeline has in fact doubled to about $10 billion. But consulting bills human hours, and if AI compresses hours across the industry, contracts simply renew smaller. No backlog protects you from that. Growth was stalling before AI showed up, and this time I think the bears are at least partly right.
Great company, fair price, missing catalyst. D tier.
C tier: Celestica (CLS), the best returns and the most ways to lose them
Celestica is the best performing large cap stock on the TSX over the last three years, up about 1,255%. It is now one of the 30 largest companies in Canada at roughly C$50 billion. And I put it in C tier. Let me explain before you close the tab.
Celestica is the AI arms dealer. It builds the physical machines of the boom: the networking switches, custom servers and GPU racks inside AI data centres. About 81% of revenue is that cloud and connectivity business, and the important part is that 41% of revenue is now hardware Celestica designs itself for the hyperscalers, not just assembles. Designers earn roughly double the margins of assemblers and are much harder to fire. 2026 revenue guidance has been raised twice to US$20.5 billion, about 65% growth, with component lead times over 52 weeks and orders booked essentially into 2028. There is even a custom rack program tied to OpenAI.
So why C tier? History and concentration.
This exact company has been here before. Celestica was a dot-com superstar that lost roughly 97% of its value when that boom stopped faster than anyone thought possible, then spent 20 years in the wilderness as a forgotten low margin manufacturer. The best teacher about Celestica’s risk is Celestica. Today, three customers make up 63% of revenue, operating margins are only about 8%, and the company is issuing roughly 10% more shares to fund a capex bill that jumps from US$1 billion to US$1.5 billion next year. If the AI buildout even pauses, 60 to 80% downside would not be outrageous. If it continues, this is probably the highest returning stock on this list. Highest ceiling, lowest floor, same stock. C tier.
B tier: Descartes Systems (DSG), the tollbooth that profits from chaos
Descartes is the quietest company on this board and the second one with a near death story. By 2004 the stock traded under $2 and the company was nearly finished. It rebuilt around one idea, a network for global trade logistics, and that brush with death is why it runs with zero debt, as policy, and roughly US$400 million of cash to this day.
The business is a tollbooth on global trade paperwork. Every international shipment generates customs filings, security screenings, tariff calculations and routing decisions, and Descartes’ network processes them for recurring fees. About 93% of revenue is recurring, and EBITDA margins just hit a record 46%.
Here is the beautiful, slightly dark part: chaos is revenue. The tariff and duty database is now the single biggest revenue contributor because the rules change weekly. US trucking volumes fell about 4% over the past year, a genuine freight recession, and Descartes’ organic growth accelerated anyway, four straight quarters: 4%, 7%, 8%, 9.1%. The complexity boost is beating the volume drag. Management just spent about US$220 million in two weeks on two acquisitions while targets are cheap, and openly talks about buying distressed peers if the downturn deepens.
The catch is that Descartes needs the world messy but not broken. A full trade recession kills volumes; a genuine lasting trade peace fades the complexity tailwind. And at about 27 times earnings it is only cheap against its own history, roughly 19 times EV/EBITDA versus a ten year median around 32. The highest quality boring company in Canadian tech, on sale because it is boring. B tier.
A tier: Constellation Software (CSU), the machine versus two tests
Constellation is the greatest compounding machine in Canadian stock market history. Since 1995 it has collected more than 1,100 tiny software companies serving niches too small for big tech to bother with: funeral homes, marinas, city transit systems. Buy at a fair price, almost never sell, send the cash to headquarters, buy a hundred more. Returns on invested capital have held in the high teens even as the capital base exploded, with barely any debt and about $2 billion a year in free cash flow.
The stock just went through the deepest drawdown in its history, about 34% below its high, on two fears at once. Founder Mark Leonard stepped down for health reasons and did not seek re-election to the board, so the culture faces its first full cycle without him in the building. And the market worries AI makes niche software cheap to rebuild. The honest bear case is real: organic growth has slipped from 6% to 4% to 2% over three quarters, and they paid about 4 times revenue for a recent deal versus a historical 1 to 1.5 times. Discipline drift is how serial acquirers die.
But look at what the machine actually did during the panic.
It spent US$1.4 billion on acquisitions in the first half of 2026 alone, the biggest first half in company history, already ahead of all of 2025. Management reports zero AI related customer attrition across 1,100 plus businesses, and AI is cutting their own development costs. Nobody is going to vibe code the city’s bus routes; the moat is switching costs and 20 year customer relationships, not the difficulty of writing code. At roughly 24 times free cash flow versus 30 plus historically, for a machine still compounding cash flow around 17% a year, every previous Constellation drawdown was a buying opportunity. This is the largest one ever. A tier, and I own it.
S tier: Shopify (SHOP), the AI shopping rail
The S tier came down to a genuine 55/45 decision between Shopify and Constellation. Shopify got the nod, and here is why.
The business is the operating system of online selling: subscriptions are about 22% of revenue, and merchant solutions, the other 78%, takes a slice of the payments, shipping and lending flowing through its stores. Revenue grew 34% last quarter, accelerating, on flat headcount, so operating income jumped 68%. That operating leverage is why I ignore the scary P/E here. Free cash flow margins are 18% and climbing, with about US$5 billion in net cash.
The double engine is the chart above. Gross merchandise volume, every dollar sold on Shopify, has grown from US$15 billion to a run rate around US$464 billion, and the share of it processed by Shopify Payments has climbed from 39% to 68%. The pile grows, and Shopify’s slice of the pile grows.
Then there is the AI shopping bet. Shopify co-wrote the Universal Commerce Protocol with Google, with Walmart, Target and Etsy on board, so its merchants’ products flow into AI assistants automatically, and AI driven orders tripled year over year. Long, specific requests that would die on page nine of a marketplace search get answered by an AI with a Shopify merchant’s product. Those are sales that simply did not happen before.
The honest bear case: roughly 14 times sales and about 70 times forward earnings is priced for perfection, and this is a consumer spending stock with a 2.6 beta that halves regularly. Down more than 80% in 2022, down about 45% into this spring. You need the stomach for that. I own it, I am not selling, and I expect growth to accelerate from here. S tier.
The final board
| Tier | Company | Ticker | The one liner |
|---|---|---|---|
| S | Shopify | SHOP | The AI shopping rail. Priced for perfection, growing into it |
| A | Constellation Software | CSU | The greatest compounder in Canadian history, at its biggest discount ever |
| B | Descartes Systems | DSG | The zero debt tollbooth that gets paid when trade gets messy |
| C | Celestica | CLS | The best three year return on the TSX, and the most ways to lose it |
| D | CGI | GIB.A | High quality at a decade low price, but the AI bears have a point |
| E | OpenText | OTEX | Paid 10% to wait on a shrinking business. A payout is not a return |
FAQ
What is the best Canadian tech stock to buy right now?
For my own money, the two I own are Shopify and Constellation Software, and the S versus A call between them was 55/45. If you want growth and can stomach huge swings, Shopify. If you want a compounder at its biggest discount in a decade, Constellation.
Which Canadian tech stock pays the best dividend?
OpenText, with a 4.7% yield and close to a 10% total shareholder yield. But read the E tier section before you chase it: the payout sits on a shrinking business, and yield is not return.
Is Celestica still a buy after going up 1,255%?
It is one of the cheapest AI infrastructure stocks anywhere at about 21 times forward earnings, and also the riskiest name on this list: three customers are 63% of revenue and the whole valuation depends on the AI buildout continuing. Size the position like it can fall 60% plus, because it can.
Do Canadian tech stocks beat US tech stocks?
Two of these six beat the Nasdaq over the past decade outright, and Constellation matched it. The average Canadian tech stock does not, but you do not buy averages, you buy companies.
What is the safest stock on this list?
Descartes. Zero debt as policy, about US$400 million in cash, 93% recurring revenue, and a business model that gets paid more when the world gets messier.
—
Want my current top Canadian stock picks with the full research behind them? Grab the free report here: [LINK]. And if you prefer video, the full ranking with every chart is on YouTube: [LINK].
