Service Level Agreement (SLA)

This Service Level Agreement (this “SLA”) is a policy governing the use of the LeaderGPU GPU server rental services (the “Service”) and applies to Customers with an active paid account in Good Standing. This SLA forms part of the agreement between Customer and LeaderGPU (the “Provider”, “we”, “us”, “our”).

This SLA describes availability targets for Covered Services and the Service Credits available if LeaderGPU fails to meet those targets. Service Credits are Customer’s sole and exclusive remedy for any failure to meet a service commitment described in this SLA, unless otherwise required by mandatory applicable law or an executed Order Form states otherwise.

1. Scope

This SLA applies only to the following production components of the Service in the single LeaderGPU Region (Rotterdam, the Netherlands) (the “Region”), collectively referred to as the “Covered Services”:

• Compute: GPU virtual machines and dedicated GPU servers equipped with datacenter or professional-class GPU accelerators listed in the Covered GPU Models table below. Compute availability is measured as external network connectivity to the resource, as defined in Section 3. This SLA measures Compute availability solely by external network connectivity, as defined in Section 3. Local storage attached to a Compute resource is not a separately measured SLA metric; storage availability, durability, and data integrity are not guaranteed under this SLA.
• Customer Portal and API: the Customer Portal and API are provided for convenience and are not Covered Services under this SLA. LeaderGPU will use commercially reasonable efforts to maintain portal and API availability but does not guarantee specific uptime targets or offer Service Credits for portal or API unavailability.

Covered GPU Models

This SLA applies exclusively to servers equipped with the following datacenter and professional-class GPU accelerators:

LeaderGPU may update the Covered GPU Models list from time to time. New GPU models will be added to or removed from the list with at least 30 days’ prior notice via the Customer Portal or email.

GPU models not listed in the Covered GPU Models table are not Covered Services under this SLA.

This SLA does not apply to beta/preview features, free services, third-party software, or Customer-managed configurations inside the operating system. Spot, preemptible, or interruptible instances (if offered) are excluded from this SLA entirely and carry no uptime commitments, high availability guarantees, or Service Credit eligibility. Due to their interruptible nature, such instances may be reclaimed, interrupted, or terminated at any time by LeaderGPU, with or without prior notice, at LeaderGPU’s sole discretion.

2. Region Model (Single Region)

LeaderGPU operates the Service in a single Region. There is no multi-Region or multi-zone failover under this SLA. Customers are responsible for architecting appropriate redundancy, backups, and disaster recovery for their workloads.

3. Definitions

• Availability Check: An automated probe performed by LeaderGPU’s monitoring infrastructure to determine whether a Covered Service resource is reachable at the applicable Demarcation Point. Each Availability Check tests for a successful response to an ICMP Echo (ping) request and/or a TCP connection to the resource. The Covered Service is considered reachable if at least one of the tested protocols returns a successful response. LeaderGPU publishes the IP addresses used by its monitoring probes; Customer is responsible for ensuring that these addresses are not blocked by host-based firewalls or network access controls. Blocking monitoring probe traffic does not constitute Unavailability.
• Monitoring Interval: A consecutive fifteen (15) minute period. Each calendar month is divided into consecutive, non-overlapping Monitoring Intervals beginning at 00:00 UTC on the first day of the month. LeaderGPU performs one Availability Check per Monitoring Interval.
• Failed Interval: A Monitoring Interval in which the Availability Check determines that the Covered Service is Unavailable (as defined below). If the Availability Check within a Monitoring Interval fails, the entire 15-minute Interval is counted as Downtime.
• Monthly Uptime Percentage (“MUP”):
  MUP = ((N − F) / N) × 100
  where N = total number of Monitoring Intervals in the calendar month and F = number of Failed Intervals for the applicable resource. For a 30-day month, N = 2,880.
• Unavailability:
  • For Compute: loss of external network connectivity to the instance/server caused by LeaderGPU, as measured at the Demarcation Point. The following are NOT Unavailability: local storage failures, data loss, or degraded disk performance that do not result in loss of external network connectivity; inability to boot or loss of connectivity caused by Customer’s modifications to system configuration (including but not limited to fstab, bootloader/GRUB, kernel parameters, network interface configuration, or disk mount settings); GPU hardware faults, performance degradation, or errors detectable only via OS-level tools (e.g., nvidia-smi, DCGM) and not resulting in loss of external connectivity; GPU performance variability within manufacturer specifications; thermal throttling; correctable or uncorrectable ECC memory errors; CUDA/driver/firmware compatibility issues arising from Customer-installed software; training convergence failures; numerical precision issues; application-level performance degradation; and performance differences between GPU models or generations.
• Demarcation Point:
  • For Dedicated Servers, availability is measured at the LeaderGPU network port delivering connectivity to the server (top-of-rack switch port).
  • For Virtual Machines, availability is measured at the virtual network interface provided by LeaderGPU.
  • Availability does not guarantee reachability to application ports that Customer may block (e.g., via host-based firewall).
• Service Credit: A monetary credit applied to the Customer’s prepaid balance or future invoices, as described in Section 5.
• Month: A calendar month (e.g., 1–28/29/30/31 of the applicable month). Where a Customer’s Service is active for only part of a calendar month, calculations are based on the portion of the month during which the Service was active.
• Good Standing: A Customer account with no outstanding Proforma Invoices or Invoices more than 30 days past due and no active violations of the Acceptable Use Policy or Terms of Service. Good Standing is evaluated at the time of the Downtime incident and at the time the Service Credit request is submitted.
• Eligible Charges: The net charges actually invoiced to Customer (or deducted from Customer’s prepaid balance) for the affected resource during the applicable calendar month, including pro-rated amounts for prepaid weekly or monthly plans. Eligible Charges exclude: (a) taxes, duties, and government-imposed surcharges; (b) one-time setup fees and migration charges; (c) professional services and consulting fees; (d) third-party pass-through charges; (e) promotional credits, sign-up bonuses, referral credits, and any other credits not representing direct payment by Customer; (f) credits or balances provided under a free trial, beta program, or evaluation agreement; and (g) previously issued Service Credits. For prepaid committed-use plans, Eligible Charges shall be the pro-rated monthly equivalent.
• Order Form: A written ordering document or online order executed by Customer and LeaderGPU that references this SLA and specifies the Covered Services, pricing, and any service-specific terms. In the event of a conflict between this SLA and an executed Order Form, the Order Form shall prevail to the extent of the conflict.

Measurement and Source of Truth

LeaderGPU performs one Availability Check per Monitoring Interval from LeaderGPU’s monitoring infrastructure within or adjacent to the data-center network. Each Availability Check tests external network connectivity to the resource at the applicable Demarcation Point using the protocols described in the Availability Check definition above. If the Availability Check fails, the entire Monitoring Interval is recorded as a Failed Interval.

Monthly Uptime Percentage and Failed Intervals are calculated based on LeaderGPU monitoring, telemetry, and logs. LeaderGPU records are the source of truth for SLA calculations. Customer-provided evidence may be used to corroborate the claim. If Customer disagrees with LeaderGPU’s determination, Customer may submit supporting evidence within 14 days of receiving the credit determination. LeaderGPU will review the evidence in good faith and provide a written response within 30 days. If the dispute remains unresolved, it may be escalated under the dispute resolution process in Section 12.

For the avoidance of doubt, N (total Monitoring Intervals) includes all Monitoring Intervals in the applicable month during which the Covered Service was active, including Monitoring Intervals that fall within Scheduled or Emergency Maintenance periods that are excluded under Section 8. Such excluded intervals are not counted as Failed Intervals.

4. Service Commitments

LeaderGPU will use commercially reasonable efforts to make the Covered Services available with the following Monthly Uptime Percentage targets, measured per calendar month and per affected resource. These targets apply only to resources equipped with Covered GPU Models as defined in Section 1. If the Monthly Uptime Percentage for a Covered Service falls below the applicable target, Customer may be eligible for Service Credits as described in Section 5, subject to the exclusions in Section 7.

5. Service Credits

If LeaderGPU does not meet a service commitment in Section 4, Customer may be eligible to receive Service Credits as described below, subject to the exclusions in Section 7 and the request process in Section 6.

5.1 Service Credits Upon Request Only

Service Credits are provided only if Customer submits a Service Credit request in accordance with the process and deadlines described in Section 6. Failure to comply with Section 6 forfeits Customer’s right to receive a Service Credit.

5.2 Credit Type

Service Credits are applied as a monetary deduction on Customer’s next invoice for the affected Covered Service, calculated as a percentage of the Eligible Charges (as defined in Section 3) for the affected resource for the applicable month. Service Credits are not refunds and are not convertible to cash, except where required by law.

5.3 Credit Schedule

If the Monthly Uptime Percentage for a Covered Service falls below the applicable target in Section 4, the Service Credit is determined by the following tables:

Service Credits are calculated per affected resource per calendar month. Only the highest applicable credit percentage applies for each resource in a given month.

5.4 Maximum Credit

The maximum Service Credit for a single affected resource for a given month shall not exceed the Eligible Charges for that resource for that month. The maximum aggregate Service Credit across all affected resources for a given month shall not exceed 100% of the total Eligible Charges for all Covered Services for that month.

5.5 Credit Application

Approved Service Credits will be applied as a deduction on Customer’s next invoice for the affected Covered Service following the month in which the credit request is confirmed by LeaderGPU. Service Credits that have not been applied at the time of termination of Customer’s account are forfeited, except where prohibited by applicable law.

5.6 No Stacking of Credits

Credits may not be stacked or combined for the same period of Downtime. A single period of Downtime generates only one Service Credit per affected resource, regardless of the number of contributing causes.

6. How to Request a Service Credit

6.1 Incident Reporting

Customer should report Downtime incidents to LeaderGPU support as soon as practicable, and no later than 72 hours after the start of the incident. Incidents reported after 72 hours remain eligible for Service Credits, but LeaderGPU will rely solely on its own monitoring data to validate the claim, and Customer-provided evidence submitted after 72 hours will not be considered.

6.2 Credit Request

To receive a Service Credit, Customer must submit a formal credit request within 30 days after the end of the month in which the SLA failure occurred. The credit request must include:

1. Customer account identifier.
2. Affected resource IDs (VM instance ID, dedicated server ID).
3. Dates, times, and timezone of each claimed Downtime incident.
4. A brief description of the incident and its observed impact on the Covered Service.

Customer may also provide additional supporting evidence (e.g., support ticket number, monitoring logs, traceroutes, API error responses) to corroborate the claim.

6.3 Forfeiture

Failure to submit a credit request within the deadline specified in Section 6.2, or failure to provide the required information listed above, will forfeit Customer’s right to receive a Service Credit for that incident.

6.4 Source of Truth

LeaderGPU monitoring and logs are the source of truth for all SLA calculations. Customer-provided evidence may be used to corroborate a claim but does not override LeaderGPU records.

7. Exclusions

Downtime does not include any Unavailability, suspension, or performance issues caused directly or indirectly by:

• Scheduled Maintenance or Emergency Maintenance (Section 8).
• Force majeure events (overmacht) that prevent LeaderGPU from delivering the Covered Service. For the purposes of this SLA, the following events constitute force majeure, provided the event is beyond LeaderGPU’s reasonable control and could not reasonably have been avoided or overcome:
  • natural disasters (earthquake, flood, hurricane, volcanic eruption) directly and physically affecting LeaderGPU data-center facilities;
  • war, armed conflict, terrorism, civil insurrection, or sabotage;
  • government sanctions, embargoes, or export-control restrictions that make performance legally impossible;
  • government orders or regulatory measures that directly and mandatorily prohibit the delivery of the Covered Service;
  • national or regional power-grid failures that overwhelm LeaderGPU’s on-site backup-power systems (UPS and generators);
  • widespread Internet backbone disruptions affecting multiple independent transit providers and Internet exchange points beyond LeaderGPU’s network;
  • cyberattacks of unprecedented scale targeting critical infrastructure broadly, provided LeaderGPU maintained commercially reasonable security measures (including timely patching of known vulnerabilities) at the time of the attack.

  The following are expressly not force majeure: LeaderGPU’s internal staffing shortages, illness, or internal labor disputes; localized power outages within LeaderGPU facilities that on-site backup systems are designed to handle; hardware or software failures within LeaderGPU’s infrastructure (which are governed by this SLA); financial difficulties or insolvency of LeaderGPU; and failures of LeaderGPU’s subcontractors or suppliers except to the extent caused by an event listed above.

  Force majeure does not excuse liability for intent (opzet) or conscious recklessness (bewuste roekeloosheid) of LeaderGPU’s management (bedrijfsleiding), as described in Section 10.2.

  A party invoking force majeure must: (a) notify the other party in writing within 5 business days of becoming aware of the event, specifying the nature, expected duration, and affected obligations; (b) take all commercially reasonable steps to mitigate the impact and resume performance; and (c) provide periodic updates at least every 5 business days until the event is resolved. Failure to notify does not extinguish the force majeure defense but may affect the period for which the defence is available.

• Customer actions or inactions, including those of Customer’s authorized users, agents, contractors, and suppliers, including Customer-managed firewalling or network configuration within the operating system (ufw/iptables/nftables), routing changes, VPN/tunnel configuration, or port blocking.
• Customer-side connectivity or equipment failures, including DNS resolution issues, domain name expiration, IP transit provider failures, bandwidth constraints, or internet connection quality at Customer’s location.
• Third-party software or images not provided and managed by LeaderGPU.
• Denial-of-service attacks or security incidents that affect service availability despite commercially reasonable mitigation efforts by LeaderGPU.
• Customer-requested downtime or planned shutdowns.
• Account restrictions, suspensions, or terminations (including for violation of acceptable use policies), or any period during which Customer has overdue payments as defined in the Good Standing requirements of Section 3.
• Documented resource quota limits, platform-standard throttling, or rate-limiting as published in LeaderGPU documentation, including throttling or resource restriction applied in response to suspected abusive behavior.
• Initial provisioning, deployment, or migration of instances (SLA coverage begins when the resource reaches “running” state for the first time).
• Downtime caused by Customer’s use of the Covered Services in a manner inconsistent with their documented features and functionality, including: (a) failure to respond to resource health notifications sent by LeaderGPU within 24 hours of notification; (b) failure to implement security patches to Customer-managed operating systems after LeaderGPU has provided notice of a critical vulnerability with a severity rating of “Critical” or “High” (as classified by the applicable vendor’s security advisory) and a remediation period of not less than 72 hours, or a remediation period of not less than 30 days for vulnerabilities rated “Medium” or “Low”; (c) use of configurations or resource allocations that exceed documented operational limits published in LeaderGPU documentation at the time of the incident; or (d) failure to comply with a specific, technically reasonable remediation recommendation provided in writing by LeaderGPU technical support in response to a Customer-reported issue, where Customer acknowledged receipt of the recommendation and where the failure to comply is the direct and proximate cause of the Downtime.
• Unauthorized access to the Service by means of Customer’s passwords, API keys, access tokens, or equipment, or otherwise resulting from Customer’s failure to follow appropriate security practices.
• Customer-provided faulty input, instructions, or API requests (for example, requests to access resources that do not exist or attempts to perform operations that are not supported).
• Defects or bugs in hardware, firmware, or software for which no commercially reasonable known solution exists at the time of the incident.
• Customer’s use of Service features, GPU models, or software versions that have been deprecated or reached end-of-life, where LeaderGPU has provided at least 90 days’ prior notice of deprecation.
• Features or services designated as beta, preview, early access, or pre-general availability.

If a force majeure event (as described above) renders the Covered Services substantially unavailable for more than 30 consecutive days, either party may terminate the affected Covered Service upon written notice. In such case, Customer shall receive a pro-rata refund of any prepaid unused fees for the terminated Service.

8. Maintenance

LeaderGPU may perform maintenance on the Service, whether planned or unplanned. Maintenance that has been communicated to Customer in advance does not count as Downtime for the purposes of this SLA.

8.1 Scheduled Maintenance

LeaderGPU will provide notice of Scheduled Maintenance through the Customer Portal, email, or status page no less than forty-eight (48) hours in advance. LeaderGPU will use commercially reasonable efforts to schedule maintenance during periods of lower usage.

8.2 Emergency Maintenance

Emergency Maintenance is maintenance required to address security vulnerabilities, imminent stability risks, or legal compliance obligations. LeaderGPU will provide notice of Emergency Maintenance as soon as reasonably practicable, and in any event no less than two (2) hours in advance where technically feasible, via the status page and email to the account’s primary contact. Emergency Maintenance communicated in advance (even with less than 48 hours’ notice) does not count as Downtime.

8.3 Uncommunicated Maintenance

Maintenance that is not communicated to the affected Customer before the start of the maintenance window is not excluded from Downtime calculations and may be counted toward the Monthly Uptime Percentage.

The Service is provided on bare-metal and GPU-passthrough infrastructure. Live migration is not supported. Any maintenance event requiring host-level access will result in a stop and restart of the affected server. Such restarts are handled as maintenance under this Section 8 and do not constitute Unavailability under Section 3.

9. Customer Responsibilities

Customer is responsible for:

• maintaining current contact information for incident notifications;
• ensuring that LeaderGPU’s published monitoring probe IP addresses are whitelisted in any host-based firewall or network access control configuration on Customer’s resources, so that Availability Checks can be performed;
• implementing and regularly verifying backups and disaster-recovery procedures appropriate for the workload. Local storage on Compute resources is not replicated by LeaderGPU; Customer bears sole responsibility for data backup, redundancy, and recovery;
• maintaining security and configuration of the operating system and applications inside instances/servers (unless a managed service is purchased);
• reporting incidents within the timeframes specified in Section 6 and cooperating with reasonable troubleshooting requests;
• complying with the LeaderGPU Acceptable Use Policy and Terms of Service;
• ensuring that Customer’s use of the Service complies with all applicable laws and regulations, including data protection and export-control requirements;
• securing and safeguarding all account credentials, API keys, and access tokens;
• monitoring resource utilization and planning capacity needs;
• maintaining all required third-party software licenses for software running on the Service (including NVIDIA, CUDA, and other vendor licenses);
• implementing checkpointing and state-saving mechanisms for long-running GPU workloads to minimize the impact of service interruptions;
• exporting Customer data before termination or expiration of the Service.

10. Limitations of Liability

10.1 Exclusive Remedy

Service Credits described in Section 5 are Customer’s sole and exclusive remedy for any failure by LeaderGPU to meet the service commitments described herein, unless otherwise required by mandatory applicable law or an executed Order Form states otherwise.

10.2 Intent and Conscious Recklessness

Nothing in this SLA excludes or limits liability for damages caused by intent (opzet) or conscious recklessness (bewuste roekeloosheid) on the part of LeaderGPU’s management (bedrijfsleiding).

10.3 General Liability Cap

Except as provided in Section 10.2, LeaderGPU’s total aggregate liability for all claims arising out of or relating to this SLA, regardless of the form of action (whether in contract, tort, or otherwise), shall not exceed the total fees paid by Customer for the affected Covered Service during the twelve (12) months preceding the first event giving rise to the claim. This limitation applies to all causes of action except liability that cannot be excluded or limited under mandatory Dutch law.

10.4 Exclusion of Consequential Damages

In no event shall LeaderGPU be liable for any indirect, incidental, special, consequential, or punitive damages, including but not limited to loss of profits, revenue, data (subject to Section 10.5), business opportunity, goodwill, costs of procuring substitute services, or loss arising from business interruption, regardless of the cause of action or theory of liability, even if LeaderGPU has been advised of the possibility of such damages, and regardless of how such damages might otherwise be classified as direct (directe schade) or indirect (indirecte schade) under applicable law. This exclusion does not apply to liability for intent (opzet) or conscious recklessness (bewuste roekeloosheid) of the management (bedrijfsleiding).

10.5 Data Loss Disclaimer

The Service is a compute platform, not a data storage or backup service. Automated backups are not included unless explicitly purchased as a separate add-on with its own service terms.

LeaderGPU is not liable for loss, corruption, or unauthorized access to Customer data stored on Compute instances or local disks, except to the extent such loss is directly caused by a failure of LeaderGPU infrastructure that constitutes Downtime under this SLA, in which case Customer’s sole remedy is the Service Credit described in Section 5. This disclaimer does not apply to data loss caused by intent (opzet) or conscious recklessness (bewuste roekeloosheid) as described in Section 10.2.

Customer acknowledges that: (a) the Service is not designed as a data archival or disaster recovery solution; (b) Customer is solely responsible for implementing backup, replication, and disaster recovery strategies appropriate for the workload; and (c) Customer should maintain off-site backups of all critical data.

10.6 Risk Allocation

The parties acknowledge that the fees charged for the Service reflect the allocation of risk set forth in this SLA, including the limitations on liability and the exclusive remedy of Service Credits.

11. Changes to This SLA

LeaderGPU may update this SLA from time to time. If LeaderGPU makes a material change, LeaderGPU will provide at least 60 days advance notice (via the Customer Portal, email, or status page) before the change takes effect, and will post an updated version and change the “Last updated” date above. Changes apply prospectively.

If a change materially reduces the service commitments or credit entitlements under this SLA, Customer may terminate the affected Covered Service without early-termination fees by providing written notice to info@leadergpu.com within 30 days after the change notification. Prepaid unused balances will be refunded on a pro-rata basis. This termination right does not apply to changes that (a) increase service commitments or credit entitlements, (b) are required by applicable law, regulation, or court order, or (c) are non-material corrections or clarifications.

If the Monthly Uptime Percentage for a Covered Service falls below its target in any three (3) consecutive calendar months, Customer may terminate the affected Covered Service without early-termination fees by providing written notice within 30 days after the end of the third such month. Prepaid unused balances will be refunded on a pro-rata basis.

The termination rights in this Section are in addition to, and do not limit, either Party’s statutory rights under applicable law, including dissolution (ontbinding) pursuant to Article 6:265 of the Dutch Civil Code.

12. Governing Law and Dispute Resolution

This SLA is governed by the laws of the Netherlands, with the exclusion of the United Nations Convention on Contracts for the International Sale of Goods (CISG).

Any dispute arising from this SLA shall be resolved through the following escalation process:

1. Support. The parties shall first attempt to resolve the dispute through the normal support channel.
2. Management Escalation. If unresolved within 14 days, the dispute shall be escalated to a designated manager at each party for good-faith negotiation.
3. Mediation. If still unresolved within 30 days of escalation, either party may refer the dispute to mediation administered under the rules of the Mediatorsfederatie Nederland (MfN) in Amsterdam. The costs of the mediator shall be shared equally between the parties.
4. Litigation. If mediation fails or is declined within 30 days of referral, either party may submit the dispute to the competent court in Amsterdam, the Netherlands.

Nothing in this section prevents either party from seeking interim or injunctive relief (kort geding) from a competent court where urgency so requires.

13. General Provisions

13.1 Survival

Sections 3 (Definitions), 5 (Service Credits), 6 (How to Request a Service Credit), 7 (Exclusions), 10 (Limitations of Liability), 12 (Governing Law and Dispute Resolution), and this Section 13 shall survive termination or expiration of this SLA and the underlying service agreement.

13.2 Entire SLA

This SLA constitutes the entire agreement between the Parties with respect to service levels for the Covered Services and supersedes all prior or contemporaneous communications regarding service levels, unless an executed Order Form expressly modifies specific provisions.

13.3 Expiry of Claims (Vervaltermijn)

Any right or claim arising out of or in connection with this SLA shall lapse (vervallen) and cease to exist if the other Party has not been notified in writing of the claim within twelve (12) months after the date on which the claimant became aware, or ought reasonably to have become aware, of the facts giving rise to the claim. This provision constitutes a contractual expiry period (vervaltermijn) and not a modification of any statutory limitation period (verjaringstermijn). This expiry period does not apply to claims arising from intent (opzet) or conscious recklessness (bewuste roekeloosheid) on the part of the other Party’s management (bedrijfsleiding).

13.4 Severability

If any provision of this SLA is held by a competent court to be invalid, illegal, or unenforceable, the remaining provisions shall continue in full force and effect. The invalid provision shall be replaced by a valid provision that most closely reflects the original commercial intent of the Parties.

13.5 No Waiver

The failure of either Party to enforce any provision of this SLA shall not constitute a waiver of that Party’s right to enforce that or any other provision in the future.

13.6 Assignment

Customer may not assign or transfer this SLA or any rights hereunder without LeaderGPU’s prior written consent. LeaderGPU may assign this SLA to an affiliate or in connection with a merger, acquisition, or sale of all or substantially all of its assets, provided the assignee assumes all obligations under this SLA.

13.7 No Third-Party Rights

This SLA does not create any right or claim for or on behalf of any third party. No person or entity other than the Parties may derive rights from or enforce any provision of this SLA.

13.8 Headings

Section headings are for convenience only and do not affect the interpretation of this SLA.

14. Contact

For Service Credit requests and technical support: info@leadergpu.com or via the Customer Portal.

For legal notices (including termination, dispute escalation, and formal claims): by email to info@leadergpu.com and by registered mail (aangetekende post) to LeaderTelecom B.V., Zekeringstraat 17 A, 1014 BM Amsterdam, the Netherlands.

Notices relating to termination, dispute escalation, or formal claims must be sent by both email and registered mail. Such notices are deemed received on the third (3rd) business day after posting (within the Netherlands) or the seventh (7th) business day after posting (from outside the Netherlands). Operational notices (support, credit requests) may be sent by email only.