Cloud adoption has reached a major tipping point: spending on public cloud now exceeds spending on traditional on-premises IT. Yet despite these massive investments, dissatisfaction remains widespread. Nearly 95% of organizations report regrets about the first contract they signed with a hyperscaler.
Why is that? Cost management has become the number one cloud challenge for organizations, even surpassing security. Businesses are moving away from a blind cloud-only mindset in favour of a cloud-smart strategy. In 2026, the cost-effectiveness of your infrastructure will no longer depend solely on the price of a virtual machine, but on your ability to manage five cost factors that are still too often underestimated.
As long as organizations continue to rely on foreign technology giants, they will remain exposed to unpredictable cloud costs. And in a context shaped by trade disputes and political tensions, that situation is unlikely to improve.
Hyperscalers — Microsoft Azure, Amazon AWS, and Google Cloud — remain American companies, even when they operate through Canadian subsidiaries. As a result, their pricing is set in US dollars. Even when invoices are ultimately issued in Canadian dollars, the amount is still calculated on the basis of USD pricing.
The consequence is straightforward: even if your usage remains perfectly stable from one month to the next, your bill can fluctuate significantly based solely on CAD/USD exchange rate movements. In a geopolitically unstable environment like 2026, this variable is anything but minor. For organizations whose cloud spending reaches into the hundreds of thousands of dollars, a change of only a few cents in the exchange rate can translate into thousands of dollars in unbudgeted costs.
The prudent approach is to build an explicit contingency into your budget to absorb those fluctuations — or better yet, choose a provider that bills directly in Canadian dollars.
Egress fees are additional charges billed by hyperscalers when data is moved out of their cloud environment: to the internet, to another provider in a multi-cloud environment, or even between two regions within the same cloud. These fees add up quickly.
At first glance, a few cents per gigabyte may seem negligible. The reality is very different. More than one-third of organizations report cloud budget overruns of 20% to 40%. NASA offers a particularly telling example: after signing a US$65 million annual agreement with AWS to host scientific data, an internal audit found that unanticipated egress fees could add another US$30 million per year — nearly half the value of the original contract.
The model is designed to keep customers locked in: data goes in for free, but getting it back out comes at a cost. That is what our experts call the lobster trap. In 2026, as hybrid and multi-cloud architectures become more common, exposure to these charges is higher than ever.
One important nuance: although the EU Data Act of January 2024 forced AWS, Azure, and Google to eliminate egress fees in the case of a definitive migration to another provider, data transfer fees in the normal course of operations still apply. And the much-publicized free allowances, often 100 GB per month, amount to a token discount of roughly $7 — barely enough to back up the data from a smartphone.
The resurgence of protectionist policies in the United States in 2025 and 2026 has introduced a new cost factor for organizations that rely on American hyperscalers. Tariffs imposed on technology equipment ripple through the supply chains of major cloud providers: servers, processors, and network appliances all become more expensive to build and operate. Those higher costs will inevitably be passed on to customers.
Organizations should expect pricing revisions from major US-based cloud providers in the coming months. For Canadian organizations, this is especially concerning: they are already paying in a foreign currency, and they are also exposed to price increases driven by political decisions entirely outside their control.
Beyond exchange rates, egress fees, and tariff impacts, there is a whole constellation of hidden charges that make cloud bills difficult to forecast: API fees, premium support tiers, cancellation fees, and variable charges tied to read operations.
A single terabyte of object storage with a major hyperscaler can carry additional fees of US$0.60 per 10,000 read operations, on top of the base storage rate. For self-service analytics tools such as Power BI, those operations are nearly impossible to predict in advance. Month after month, financial predictability is seriously undermined.
In a context of heightened foreign interference risks, data sovereignty is no longer a theoretical concept. It is a business imperative. Decision-makers should nonetheless be wary of false promises: an American hyperscaler that guarantees your data remains physically in Canada is not, by definition, offering a sovereign cloud.
A truly sovereign cloud requires that data be hosted locally, operated by Canadian citizens, controlled by a local legal entity, and protected from foreign laws, including the US CLOUD Act and FISA 702. This is a fundamental distinction that decision-makers in the public sector, healthcare, and financial services can no longer afford to ignore.
Failing to meet obligations around the protection of sensitive and confidential data could eventually lead to legal action, along with the associated costs and penalties, not to mention a loss of trust among citizens, customers, and partners.
In 2026, the success of your cloud transformation will depend on your ability to forecast your costs.
In the face of growing complexity, the best alternative is to work with a Canadian provider that offers genuine contractual transparency and a sovereign solution with no hidden fees and no technological lock-in.
Le cloud souverain Cirrus de Micrologic a été conçu précisément pour éliminer les facteurs qui rendent les coûts imprévisibles pour les organisations qui passent à l’infonuagique.
| Hyperscalers | Cirrus Sovereign Cloud | |
| Exchange Rate | Variable exchange rates that affect costs | Billing in Canadian dollars, with no exchange rate fluctuations |
| Egress Fees | Egress fees that penalize data mobility and data exit | No egress fees or API fees, allowing data to move freely |
| Cross-Border Tariffs | Exposure to price increases and the effects of cross-border policy decisions | A more stable local model, less exposed to foreign commercial and political decisions |
| Hidden Fees | Hidden fees that are difficult to forecast: API calls, support, read operations, add-ons | Simple, transparent, predictable billing with no hidden fees; resources can be scaled up or down without penalty |
| Data Sovereignty | Data residency without true legal sovereignty; potential legal risks | True sovereignty: data hosted, operated, and protected in Canada, under Canadian jurisdiction |