Does a Cloud Provider Outage Excuse Your SaaS Company? Force Majeure and SLA Credits

By Prakhar Rai, Founder, My Legal Pal. Attorney, Bar Council of India. LL.B, NLSIU Bangalore. Master of Business Laws.

Quick answer: Not automatically. Whether an outage at your cloud provider counts as force majeure depends on the exact words in your contract, and many clauses do not cover failures of your own suppliers. Even when the clause helps, it usually excuses the breach but not the credits, the refund or the customer’s right to leave. And your cloud provider’s SLA will probably give you far less than you owe your own customers. Check that gap before the outage, not after.

Key takeaways

  • In common law countries the clause does the work. There is no general rule that a supplier’s outage excuses a SaaS vendor.
  • Force majeure and the SLA do different jobs. Force majeure excuses a breach. The SLA sets the credits, and sometimes the only remedy.
  • SLA credits can be owed even when you are not in breach, unless the SLA itself excludes the event.
  • Say in words whether a cloud or hosting failure counts as force majeure. Silence invites a fight on the worst day.
  • Your provider’s SLA is usually a credit against your cloud bill, not a payment that covers what you owe your customers.
  • Customers will accept force majeure relief if it comes with limits: notice, a disaster recovery plan, and a way out if the event drags on.

It is 3am and a whole cloud region is down. Your customers cannot log in. Your status page says “upstream provider issue”. By morning a customer asks one question: “Is this force majeure, or do we get our money back?” The answer sits in two clauses that were probably written months apart by different people, and they do not always agree.

This guide is written from the vendor’s side, and it points to the customer’s view where the two differ. For the basics, see our guides to the force majeure clause and the service level agreement. Here we look at what happens when the two meet in a cloud outage. We use AWS as one example because its terms are public. AWS is a trademark of its owner, and this article is not affiliated with or endorsed by them.

The most common surprise is a founder who copied a force majeure clause from a template, then learned during the first big outage that it never mentioned hosting providers, while the SLA excluded “force majeure events”. Each document pointed at the other and the customer read both in its own favour.

Is a cloud provider outage force majeure for your SaaS company?

Not automatically. A court or an arbitrator will ask three questions, and your clause has to answer yes to all of them.

Question Why it matters What to check in your clause
Is the event inside the clause? Some clauses list events. Others say “beyond reasonable control”. A list that says “natural disaster, war, strike” may not reach a software failure Does it name outages of hosting, cloud or network providers?
Are supplier failures carved out? Many clauses cover events beyond your control but exclude failures of your own subcontractors, unless the supplier was itself hit by a force majeure event Is there a “subcontractors and suppliers” exception?
Did the event cause your failure, and could you have avoided it? Customers argue that picking one region or one provider was within your control, and that a failover plan would have avoided the outage Does the clause refer to a disaster recovery plan or mitigation steps?

The phrase “beyond reasonable control” sounds generous. In a SaaS contract it is the opposite. Architecture is a choice you made, so a customer will say the outage was within your control.

Does the law give you force majeure if your contract is silent?

In common law countries, usually not. The clause is your protection, so it has to be written. In other legal systems the starting point differs.

Country or system Starting point What it means for you
United States The Uniform Commercial Code excuses a seller in a contract for sale when performance becomes impracticable because of an event whose non-occurrence was a basic assumption of the contract, and the seller must give timely notice. It deals with sales, and SaaS is usually treated as a service or licence Do not count on a statute. Rely on your clause, and give notice quickly
United Kingdom In a 2024 decision the UK Supreme Court applied a force majeure clause by its wording and stressed freedom of contract and certainty Spell out the event, the causal link and the steps you must take
Civil law countries, such as France Many civil codes define force majeure in a statute, so a contract may get an excuse rule even if it says nothing, and a clause can change it Check local law for customers in those countries
India The Indian Contract Act, 1872 deals with contracts that become impossible (section 56).  Write the clause in full, as you would anywhere else

If you sell to customers in several countries, write the clause for the strictest reading. A clause that works in a common law court also works where a statute fills the gaps.

Does your SLA cover an outage that starts at your cloud provider?

Yes, unless the SLA says it does not. The customer bought availability from you, not from your cloud provider. Most SLAs credit the customer when uptime falls below a target, whether or not you did anything wrong. That is why force majeure and the SLA need to be read together.

Scenario Force majeure clause SLA Likely result for the customer
A Covers cloud provider failures Excludes force majeure events No breach and no credit. Customers often fight this
B Covers cloud provider failures Does not mention force majeure You are excused from breach, but credits may still be owed, because they do not depend on fault
C Silent on supplier failures Silent Breach. Credits owed, and possibly damages up to your liability cap
D Silent on supplier failures Credits are the sole remedy Credits only

Scenario B is the trap. Many vendors assume force majeure removes everything. It removes the breach. It does not remove a credit that the SLA gives automatically. The fix is to say, in the SLA, which events do not count toward downtime.

Are SLA credits the customer’s only remedy?

Only if the contract says so, and enterprise customers push back hard. A “sole and exclusive remedy” clause is common in vendor paper. A customer who has lost a day of work for a credit worth a few percent of one month’s fees will not accept it for long outages.

The usual compromise is a sole remedy for ordinary misses, with a carve-out for repeated or serious failures:

Position Sample wording
Vendor favourable “Service Credits are the Customer’s sole and exclusive remedy for any failure to meet the Availability Target.”
Balanced “Service Credits are the Customer’s sole remedy for a failure to meet the Availability Target, except that the Customer may terminate this agreement on notice if Availability falls below [the target] in any [three] months in a period of [twelve] months.”
Customer favourable “Service Credits are a price adjustment and do not limit any other right or remedy of the Customer, including damages and termination.”

Whatever you choose, write the claim procedure. Say how the customer claims, within how many days, and what evidence is needed. Without a procedure, every outage becomes an argument about when downtime started.

Will your cloud provider cover what you owe your customers?

Usually not by much. Public cloud terms are written to limit what the provider pays. AWS is one example, and the figures below are from its public terms as we read them on 20 November 2026. Terms change, so check yours.

  • Credits are a percentage of your bill. In the AWS compute SLA, credits are 10%, 30% or 100% depending on how far uptime falls, and they are calculated from the monthly bill for the affected region or instance. They are applied as credits and cannot be refunded as cash.
  • The SLA excludes events beyond its control. It does not apply to outages caused by factors outside AWS’s reasonable control, including any force majeure event.
  • The SLA is the sole remedy. The SLA says it sets out your sole and exclusive remedies.
  • You must claim in time. The request must be received by the end of the second billing cycle after the incident.
  • Damages are largely excluded. In the AWS customer agreement, neither side is liable for unavailability of the services, though that does not limit service credits, and each side’s liability is capped at what you paid for the services that gave rise to it in the previous 12 months.
  • Force majeure applies, but not to payment. The agreement excuses delay or failure caused by events beyond a party’s reasonable control, except for payment obligations.

Now compare that with what you promised your own customers:

Item What you promised What you can recover upstream The gap
Credit amount A percentage of the customer’s fee A percentage of your cloud bill for that region Your bill is a small part of the customer’s fee, so the credit may not cover yours
Damages Possibly up to your cap Largely excluded for unavailability You carry the loss
Claim window Your customer’s deadline A fixed upstream deadline, such as two billing cycles If customers claim late, you may not be able to
Evidence Your own uptime records Your provider’s records and root cause report You need both to argue force majeure

Providers also publish post-event summaries for major outages. AWS lists one for an October 2025 disruption in its Northern Virginia region. If you rely on force majeure, that kind of document is your evidence. Add a clause that requires you to share it with customers on request, because customers will ask. Every provider has its own terms, so read yours rather than assuming it matches this one.

What should the force majeure clause say about upstream failures?

It should say, in words, whether a cloud or hosting failure counts. Silence is the worst choice, because both sides will read it in their own favour.

Position Sample wording
Vendor favourable “Force Majeure Event means any event beyond a party’s reasonable control, including a failure or outage of third-party hosting, cloud, network or utility services.”
Balanced “A failure of a third-party provider is a Force Majeure Event only if it is itself caused by an event that would be a Force Majeure Event if it affected the Supplier directly, and the Supplier has followed its business continuity plan.”
Customer favourable “A failure of the Supplier’s subcontractors, hosting providers or cloud providers is not a Force Majeure Event.”

Most customers who negotiate will also ask for these exclusions:

  • Payment obligations are not excused.
  • A security incident caused by a failure to follow the agreed security measures is not force majeure. Your data processing agreement covers breach duties separately.
  • The relief stops when the event stops. You must restart the service as soon as you reasonably can.

Do you need to be multi-region to rely on force majeure?

Not by law, but customers will ask, and your wording decides. Many clauses require you to use “reasonable endeavours” to avoid or reduce the effect of the event. That phrase leaves a lot open. Must you fail over to another region? Must you have tested it?

In a 2024 shipping case the UK Supreme Court held that, without express words, a reasonable endeavours proviso did not require a party to accept an offer of performance different from the contract. It reasoned from freedom of contract and the value of certainty. That case was about payment in a different currency, not cloud services. The drafting lesson is the same: do not leave it to a court to decide what steps you owed. Say what they are.

Commitment Sample wording What to watch
Disaster recovery plan “The Supplier will maintain a documented disaster recovery plan and test it at least once every [twelve] months.” Only promise testing you will actually do
Recovery targets “The Supplier will restore the Service within [x] hours of a failure (recovery time objective) with no more than [y] hours of data loss (recovery point objective).” Do not promise targets stricter than your provider’s own
Architecture “The Service is hosted in [two] availability zones within a region.” A description of the architecture becomes a promise. Keep it accurate

If you promise multi-region failover and your tests do not prove it works, you have turned an excused outage into a breach. Promise what you can prove.

How long can force majeure last before the customer can leave?

The clause should set a limit, because customers will insist on one. Without a limit, you can stay “excused” for months while the customer pays for a service it cannot use.

A short checklist:

  1. Notice. You tell the customer promptly, and give updates while the event continues.
  2. Fee relief. The customer does not pay for periods when the service is unavailable, and gets a refund of prepaid fees for the unused period.
  3. Exit. If the event lasts more than [30, 60 or 90] days, either party can end the agreement on written notice.
  4. Data. During the event and after exit, the customer can export its data, as set out in the exit clause. Our guide to SaaS IP clauses covers what should survive the end of the contract.
  5. Report. You give a short written report on the cause and what you changed, within [x] business days.

Mistakes that cost SaaS companies money

Mistake Why it hurts Fix
Force majeure clause lists “internet or cloud failure” with no limits Customers read it as a licence to fail Tie it to a disaster recovery plan and to causation
SLA excludes force majeure, but the clause is silent on suppliers Arguments on every outage about which document wins State the treatment of upstream failures in both
Credits promised to customers are larger than you can recover upstream You pay the gap Match credits to what you recover, and cap total credits
Sole remedy clause with no exit for repeated failure Enterprise buyers will not sign Add a termination right after repeated misses
No claim procedure Disputes over when downtime began and who must prove it Set the window, the evidence and the method
Promising multi-region hosting you have not tested A force majeure event becomes a breach Promise only what you have tested
Marketing page promises “99.99% uptime” but the contract says something else A customer can hold you to the sales promise Use one number everywhere

A checklist before the next outage

  1. Does your force majeure clause say whether cloud, hosting and network failures count?
  2. Does it carve out supplier failures, or include them with conditions?
  3. Does the SLA say which events do not count as downtime, and does it match the clause?
  4. Do you know whether credits are owed without a breach?
  5. Are credits the sole remedy, and is there an exit for repeated failure?
  6. Is there a claim procedure with a window, evidence and a method?
  7. Have you compared your credits with what your provider gives you?
  8. Do you know your provider’s claim deadline, and can you meet it?
  9. Does your liability cap cover the gap? Our post on the limitation of liability clause explains how caps work.
  10. Is there a time limit on force majeure and a right to exit?
  11. Does your disaster recovery promise match what you have tested?
  12. Do your terms, status page and sales material say the same thing?

Get these two clauses to work together

If you are drafting or reviewing a SaaS agreement and want the force majeure clause, the SLA and the liability cap to line up, our team can help, and we can review your agreement for your jurisdiction. See our SaaS agreement drafting service, speak to our technology lawyers for tech startups, SaaS and AI platforms, or start from our SaaS agreement template.

Quick answers

Is a cloud provider outage force majeure? Not automatically. It depends on the wording of your clause. Many clauses exclude failures of your own suppliers unless the supplier was itself hit by a force majeure event.

Can a SaaS company avoid SLA credits because its cloud provider was down? Only if the SLA excludes that event. Credits often do not depend on fault, so force majeure may excuse a breach and still leave the credit owed.

Are SLA credits the customer’s only remedy for downtime? Only if the contract says so. Customers often negotiate a carve-out that lets them terminate after repeated failures.

Does force majeure excuse a SaaS vendor from refunds? Not by default. Say in the contract whether the customer gets a fee reduction or a refund of prepaid fees when the service is unavailable.

Does my cloud provider’s SLA cover what I owe my customers? Rarely in full. In the AWS example, credits are a percentage of your own bill, they are the sole remedy, and liability for unavailability is largely excluded.

Do I need multi-region hosting to rely on force majeure? Not by law. But your clause may require reasonable steps to avoid the event, and customers may ask for a tested disaster recovery plan. Promise only what you have tested.

Can a customer terminate after a long outage? Only if the contract gives that right. A good clause sets a limit, such as 30, 60 or 90 days, after which either side can end the agreement.

Does force majeure excuse payment? Often not. Many force majeure clauses, including the AWS one, exclude payment obligations from relief.

Is force majeure implied if the contract is silent? Usually not in common law countries. Some civil law systems define it in a statute, so check local law for customers there.

A note on who we are

My Legal Pal is an international team of contract lawyers. This article is general information, not legal advice on your facts. We can review your agreement for your jurisdiction.


About the author. Prakhar Rai is an advocate enrolled with the Bar Council of India and the founder of My Legal Pal. He holds an LL.B from NLSIU Bangalore and a Master of Business Laws, and works on contract drafting, review and disputes for founders and businesses across borders. Full profile

Leave a Reply

Your email address will not be published. Required fields are marked *

Are you human? Please solve:Captcha