+49 6430 9227117
Portal loginPortal
NVIDIA MELLANOX ADAPTER · TPM FOR NICS AND DPUS

NVIDIA Mellanox ConnectX & BlueField Maintenance — vendor-independent adapter service

We service NVIDIA Mellanox network adapters vendor-independent — ConnectX-4 (25/40/100G), ConnectX-5 (25/100G), ConnectX-6 Dx (25/100/200G), ConnectX-7 (200/400G) and BlueField-2/-3 DPUs. With OEM adapters from our own warehouse and SLA up to 24×7×4. Economic sweet spots: ConnectX NIC maintenance pays off from fleets of 50+ adapters (volume consolidation), BlueField DPU justifies TPM from first unit (high-value component, 3,000-15,000 EUR hardware value per DPU). 30 to 60 percent below NVIDIA conditions or distributor maintenance.

Quote in 48 h Calculate savings

Which Mellanox adapter models we service

Mellanox ConnectX has been the market leader for high-performance networking adapters in the data center for over a decade — RDMA, NVMe-over-Fabric, GPUDirect, Ethernet and InfiniBand on the same hardware platform. ConnectX-4 (released 2014) and ConnectX-5 (2016) are widespread in DACH fleets, ConnectX-6 Dx (2019) and ConnectX-7 (2022) are the current generations. BlueField DPU is the programmable hardware platform combining ConnectX functionality with ARM cores for workload offload.

ConnectX-4 / ConnectX-5 · older generations
ConnectX-4 (25/40/100G EN+IB) · ConnectX-5 (25/100G EN+IB)
ConnectX-6 Dx · current mid-range
ConnectX-6 Dx (25/100/200G) · ConnectX-6 VPI
ConnectX-7 · current high-end
ConnectX-7 (200/400G EN+IB)
BlueField DPU · data processing units
BlueField-2 (25/100G with ARM cores) · BlueField-3 (200/400G with extended ARM compute)

Why TPM maintenance for Mellanox adapters

Mellanox adapter maintenance has two economic sweet spots: volume maintenance for ConnectX NICs and high-value maintenance for BlueField DPUs. ConnectX NICs are cheap per unit (300-1,500 EUR hardware value) but deployed in high quantities in AI/HPC cluster builds — 100-1000+ adapters are not uncommon. NVIDIA standard support or distributor maintenance per adapter moves at 50-150 EUR per year. For 200+ adapters, this adds up to five-figure amounts — TPM reduces this to 30-50 percent. BlueField DPU is calculated differently: hardware value 3,000-15,000 EUR, NVIDIA support proportionally more expensive, TPM pays off from first unit. For mixed fleets (NICs + DPUs) we offer multi-class contracts with different pricing logics.

We service Mellanox adapters with OEM components from our own warehouse. ConnectX-4 and ConnectX-5 (older generations, high install density in DACH) are abundantly available. ConnectX-6 Dx and ConnectX-7 (current generations) we continuously build the component pool. BlueField-2 and BlueField-3 included as high-value components in our pool. MLNX_OFED drivers and software stack (RDMA, GPUDirect, NVMe-oF, OVS-Offload, ASAP² for BlueField) continue license-free — all driver functions stay active under TPM. Software subscriptions like NVIDIA AI Enterprise or DOCA SDK are independent of hardware maintenance.

30–60 %
Savings vs. NVIDIA support or distributor
From 50 NICs
Economic sweet spot for ConnectX volume
BlueField from unit 1
DPU as high-value component always TPM-attractive
MLNX_OFED stays
RDMA, GPUDirect, ASAP² license-free active

Generations timeline & TPM coverage

Per hardware generation: vendor phase (slate) and TechCare coverage window (teal) up to ~5 years post-OEM EOSL.

Generation status of Mellanox adapters

ConnectX generations follow a 6-8 year lifecycle. ConnectX-4 approaches EOSL from 2025, ConnectX-5 in mid-cycle, ConnectX-6 Dx and ConnectX-7 are current. BlueField generations are all current.

Model family Released OEM support ends TPM status
ConnectX-4 (25/40/100G) 2014 2025–2026 Recommended
ConnectX-5 (25/100G) 2016 2026–2027 Supported
ConnectX-6 Dx (25/100/200G) 2019 ca. 2028+ Supported
ConnectX-7 (200/400G) 2022 ca. 2030+ Supported
BlueField-2 DPU 2020 ca. 2028+ Supported
BlueField-3 DPU 2023 ca. 2031+ Supported

As of 2026. EOSL data based on official vendor roadmaps and subject to change. Binding case-by-case information available on request.

What we deliver

OEM components

Our warehouse and certified refurbishing sources for DGX and Mellanox.

DGX specialist engineer

German-speaking technicians with NVLink/NVSwitch training, 4-hour response time guaranteed.

Flexible SLA per system

Parts Only, 5×9 NBD or 24×7×4 — freely combinable by location and criticality.

Multi-vendor contract

One contract for DGX, Mellanox and all other vendors. AI cluster stack consolidation.

Risk assessment

Component pool status per model before contract conclusion — honest disclosure.

CUDA & AI software stay

CUDA, AI Enterprise, Base Command Manager independent of hardware maintenance.

Comparison

TechCare vs. NVIDIA Mellanox ConnectX & BlueField Adapter

Concrete SLA tiers, OEM comparison, coverage details and source footnotes — on a print-friendly page.

Open Battle-Card

FAQ on Mellanox adapter maintenance

Which Mellanox adapter models do you service?
Complete adapter family: ConnectX-4 (25/40/100G Ethernet and InfiniBand), ConnectX-5 (25/100G EN+IB), ConnectX-6 Dx (25/100/200G), ConnectX-6 VPI (Virtual Protocol Interconnect with dual EN/IB), ConnectX-7 (200/400G EN+IB) plus BlueField-2 DPU (25/100G with ARM cores) and BlueField-3 DPU (200/400G with extended ARM compute platform). Including all PCIe form factors (standard, OCP 2.0, OCP 3.0) and optics modules for respective speeds.
From how many adapters does TPM pay off for Mellanox?
Rule of thumb: from 50 ConnectX NICs (same or mixed generation) volume TPM becomes economically attractive — contract administration, SLA reporting and onsite engineer readiness justify minimum contract value. For 100+ NICs savings are significant (five-figure amounts per year). For BlueField DPUs TPM pays off from first unit because hardware value per DPU is 3,000-15,000 EUR and NVIDIA support proportionally more expensive. Recommendation: consolidate mixed fleets (e.g. 200 ConnectX-5 + 20 BlueField-3) in one multi-class contract.
What does TPM cost for Mellanox adapters concretely?
ConnectX NIC maintenance: NVIDIA support or distributor typically 50-150 EUR per adapter per year (depending on generation and SLA), with TechCare 25-70 EUR per adapter. For 200 ConnectX-5 adapters: 10,000-30,000 EUR NVIDIA support, 5,000-14,000 EUR TechCare. BlueField DPU: NVIDIA premium support typically 800-2,500 EUR per DPU (depending on generation and SLA), with TechCare 350-1,000 EUR per DPU. For 20 BlueField-3 DPUs: 16,000-50,000 EUR NVIDIA, 7,000-20,000 EUR TechCare. Mixed fleet 200 NICs + 20 DPUs: 26,000-80,000 EUR NVIDIA total, 12,000-34,000 EUR TechCare total.
Do MLNX_OFED, RDMA, GPUDirect and DOCA continue to work without an NVIDIA contract?
Yes. MLNX_OFED drivers and software stack continue license-free on the adapters — RDMA (RoCE/InfiniBand), GPUDirect (for AI clusters), GPUDirect Storage, NVMe-over-Fabric, ASAP² (for BlueField), Open vSwitch offload and all standard driver functions stay functionally active. MLNX_OFED updates to newer versions are freely available from NVIDIA — TPM customers can download them without restriction. DOCA SDK (for BlueField programming) and NVIDIA AI Enterprise are software subscriptions, independent of hardware maintenance — customers actively using them keep them with NVIDIA. Firmware updates on adapters require an active NVIDIA contract but are usually uncritical with stable configurations.
Which SLA levels do you recommend for Mellanox adapters?
Strongly depends on cluster architecture: ConnectX NICs in productive AI/HPC cluster nodes with redundant multi-path setups (standard in modern AI build): 5×9 NBD economically appropriate — single NIC failure absorbed by multi-path. BlueField DPUs as central storage offload or security inspection components: 24×7×4 mandatory due to service impact on failure. For single NIC setups (rare in modern AI builds) or critical database storage targets: 24×7×4 mandatory. For fleet migration during cluster refresh: Parts Only sufficient for out-phase adapters.
Can you service older ConnectX-3 or pre-Mellanox-acquisition fleets?
ConnectX-3 (released 2012) and ConnectX-3 Pro (2014) are completely EOSL and practically no longer productive in the market. If fleets are present, we can check coverage on request — typically refresh is the only sensible option here. ConnectX-2 and older are no longer part of our pool. Mellanox acquisition by NVIDIA was 2020 — all adapter models from ConnectX-4 are today under NVIDIA branding and serviced by us regularly, regardless of original manufacturer branding (Mellanox before 2020, NVIDIA since 2020).
Can we get optics for Mellanox adapters?
Yes. We supply optics for all speeds (25/40/50/100/200/400G) as certified third-party optics or original NVIDIA/Mellanox optics — both options technically compatible. SFP28/QSFP+/QSFP28/QSFP56/QSFP-DD form factors. DAC and AOC cables as well. Third-party optics are typically 50-70 percent cheaper than original, technically identical — Mellanox adapters accept them without restriction. For InfiniBand setups (HPC) we often recommend original NVIDIA optics for maximum compatibility with Quantum switch latency optimizations.
Can we have Mellanox adapters and switching in the same contract?
Yes. Multi-class NVIDIA contracts are exactly for AI cluster stacks — DGX (compute) plus ConnectX NICs (networking adapters) plus Spectrum or Quantum switches (networking backbone) plus BlueField DPUs (programmable networking) — all in one contract, one point of contact, one SLA report set. Plus all other vendors (Supermicro AI servers, Dell PowerEdge GPU nodes, storage vendors) in the same construct.
Service performance

Real actuals Q1 2026 — straight from our ITIL ticketing.

99,2 %
Tickets resolved within agreed response time
2,4 h
Avg. first response on 4h SLA tier
88 %
First-time fix on initial dispatch
97 %
Spare part on site within 4 h, DACH depots

Save on NVIDIA Mellanox ConnectX & BlueField Adapter without risk

Send us your inventory list — we deliver a binding fixed-price quote within 48 hours. No migration, no risk, clear comparison to OEM maintenance.

Request quote Download whitepaper
Related topics
Core service
Hardware maintenance in detail
How we set up TPM: contract, SLA tiers up to 24×7×4, engineer pool, spare parts strategy — the service behind the vendor hubs.
View service →
Compliance
Audit documentation in the contract pack
DORA, NIS2, KRITIS, BAIT, TISAX, ISO 27001 — what the auditor checks, what TechCare delivers automatically with the contract.
View audit topics →
Practice
Real references
Three templates from European customers 2024–25 — how the migration from OEM to TPM ran in reality.
View cases →
More from NVIDIA

Other NVIDIA models and service

All NVIDIA models

Overview of all NVIDIA maintenance pages with SLA, spare-parts strategy and vendor-specific notes.

Open NVIDIA hub →
NVIDIADgx Legacy
View more →
NVIDIADgx
View more →
NVIDIAMellanox Switching
View more →
Service options

Contract, SLA tiers, engineer pool, spare parts — the service behind vendor maintenance in detail.

View hardware maintenance →