RAM/GPU Surcharge Print

  • 0

======================================================================
HARDWARE COMPONENT SURCHARGE POLICY
======================================================================

Last Updated: October 30, 2026

As of September 1, 2026, new orders at 42U Networks L.L.C. ("42U Networks")
are subject to targeted hardware component surcharges: a 30% memory
surcharge on standard compute tiers, and an additional 50% GPU surcharge on
GPU-accelerated instances and dedicated GPU hardware.

This document outlines what these surcharges apply to, why they exist, how
they are calculated, and our roadmap for phasing them out as supply chains
stabilize.


----------------------------------------------------------------------
1. THE SHORT VERSION
----------------------------------------------------------------------

- Memory Surcharge: New compute orders carry a 30% memory surcharge.
- GPU Surcharge: New orders featuring dedicated or shared GPU acceleration
(compute, AI/ML inference/training, rendering) carry a 50% GPU surcharge.
- Affected Services: Virtual Private Servers (VPS), dedicated bare metal
servers, game hosting, and web hosting.
- Unaffected Services: Customer-owned Colocation and unmetered IP Transit
carry zero surcharges.
- Legacy Exemption: Services active prior to September 1, 2026 are grandfathered
at their original price points and are exempt from recurring surcharges for
the lifespan of the continuous service. Upgrades enroll into the current
surcharge schedule.
- Transparent Line Items: Base plan pricing remains unchanged. Surcharges are
itemized separately on invoices so they can scale down or be removed entirely
as upstream hardware procurement costs normalize.


----------------------------------------------------------------------
2. WHY THE SURCHARGES EXIST
----------------------------------------------------------------------

MEMORY MARKET CONDITIONS:
Server-grade memory pricing has experienced unprecedented inflation. For
context, a standard 128GB DDR5 ECC server kit that cost roughly $400 in
mid-2025 has escalated to over $1,900 from identical wholesale distributors.
This surge is driven by global hyperscale AI datacenter buildouts, which have
monopolized production capacity for High-Bandwidth Memory (HBM) and enterprise
DRAM. Similar constraints have hit enterprise NAND flash storage. Industry
fabricators project standard DRAM supply tightness to persist well into 2027.

GPU MARKET CONDITIONS (50% SURCHARGE):
High-performance server GPUs face even steeper supply bottlenecks, severe
foundry packaging backlogs (such as CoWoS constraints), inflated secondary-market
premiums, and intensified datacenter electrical and thermal engineering costs.
Securing enterprise accelerators and reliable GPU host chassis requires
massive capital outlays that dwarf traditional CPU compute nodes.

We absorbed these differences for as long as possible by drawing down
pre-purchased inventory throughout early 2026. However, scaling new nodes and
deploying new GPU deployments requires acquiring components at current market
rates.


----------------------------------------------------------------------
3. WHY A SURCHARGE INSTEAD OF A PERMANENT PRICE INCREASE?
----------------------------------------------------------------------

We view current memory and GPU procurement costs as cyclical rather than
permanent economic baselines. While we do not expect component pricing to
fully return to pre-2025 levels, locking permanent base price increases into
our catalog would arbitrarily penalize customers long after manufacturing
yields recover.

A surcharge is a transparent, flexible line item:
- It can be lowered or eliminated immediately as wholesale costs decline.
- It prevents baking temporary hardware spikes into permanent recurring base
tariffs.
- When supply chains stabilize and long-term component pricing settles, base
plans will be adjusted to reflect real costs, and the surcharges will be
retired.


----------------------------------------------------------------------
4. SCOPE: WHAT THE SURCHARGES APPLY TO
----------------------------------------------------------------------

AFFECTED SERVICES:
- Standard VPS & Dedicated Servers: 30% Memory Surcharge applied to base compute.
- GPU VPS & Dedicated GPU Servers: 50% GPU Surcharge applied to GPU-allocated
instances, plus applicable base memory additions.
- Shared Web Hosting & Game Servers: 30% Memory Surcharge applied to plans where
42U Networks supplies node RAM.

EXCLUDED / UNAFFECTED SERVICES:
- Colocation: In colocation arrangements, you supply, rack, and own your
hardware; 42U Networks does not supply RAM or GPUs, so colocation is exempt.
- IP Transit & BGP Ports: Network transport and bandwidth allocations carry
no server memory or GPU component dependencies.


----------------------------------------------------------------------
5. EXISTING SERVICES & UPGRADES
----------------------------------------------------------------------

- Grandfathered Status: If your service package was provisioned and active
before September 1, 2026, your subscription remains locked at its original
rate. Surcharges will not be retroactively applied to those active slots.
- Service Upgrades: If you upgrade a grandfathered service (for example, adding
RAM, expanding storage, or attaching a GPU accelerator), the upgraded
resources will be enrolled into the applicable surcharge schedule at prevailing
rates.
- Renewals for Post-Sept 1 Orders: Any service ordered on or after September 1,
2026 maintains the surcharge on recurring renewal invoices until market
adjustments occur.


----------------------------------------------------------------------
6. SPECIFICATIONS AND ALLOCATIONS REMAIN UNCOMPROMISED
----------------------------------------------------------------------

Rather than implementing transparent surcharges, some hosting providers quietly
degrade service allocations—reducing memory allowances, switching to budget
storage, or oversubscribing virtual CPU and GPU cores.

42U Networks refuses to compromise performance:
- Memory allocations remain dedicated, unthrottled, and full-speed.
- GPU cores, VRAM, and PCIe lanes remain dedicated without noisy-neighbor
oversubscription.
- Dedicated servers receive genuine, factory-grade enterprise hardware as
specified in the order configuration.


----------------------------------------------------------------------
7. FREQUENTLY ASKED QUESTIONS (FAQ)
----------------------------------------------------------------------

Q: How will the surcharge look on my bill?
A: Surcharges are broken out as distinct line items (e.g., "Component Surcharge
- Memory (40%)" and "Component Surcharge - GPU Hardware (50%)") directly
beneath the base plan cost.

Q: If component prices fall, will the surcharge decrease?
A: Yes. Surcharges are designed to move in both directions. As wholesale
component acquisition costs drop, we will adjust the percentages downward
accordingly.

Q: Will 42U Networks notify me before modifying surcharge percentages?
A: Yes. We will provide at least seven (7) days advance notice via email and
portal announcement prior to any rate modifications taking effect on
upcoming renewal invoices.

Q: Can I lock in terms or pay upfront?
A: Yes. If you wish to discuss term commitments, annual contracts, or bulk
hardware reservations, please submit a ticket to our sales and billing
department.


----------------------------------------------------------------------
8. QUESTIONS & SUPPORT
----------------------------------------------------------------------

If you have specific billing inquiries or wish to review configuration options
for upcoming deployments, please reach out to our team:

- Client Support Portal: Open a ticket under the "Billing" department.
- Website: https://42u.networks / https://get42.us
======================================================================


Was this answer helpful?

« Back