Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesNVIDIA announced a liquid-cooled A100 80GB PCIe accelerator on May 23, 2022, presenting it as the company’s first data-center PCIe GPU with direct-chip liquid cooling. The cards were sampling at announcement time, with general availability expected in summer 2022. NVIDIA later described liquid-cooled H100 options as an early-2023 plan. Those dates and availability statements were launch-era expectations, not confirmation of current stock or every planned shipment.
What NVIDIA announced in May 2022
The announced product was an A100 80GB PCIe GPU using direct-chip liquid cooling. NVIDIA positioned it for artificial intelligence, data analytics and high-performance computing in data centers that need substantial accelerator performance while managing power and heat.
NVIDIA said the A100 cards were sampling on May 23, 2022, and expected to become generally available that summer. The announcement described liquid cooling as a response to demand for more efficient, high-performance data-center designs.
What happened with system support
In its COMPUTEX 2022 recap, NVIDIA said at least a dozen system builders would support liquid-cooled A100 PCIe GPUs, with the first systems expected to ship in the third quarter of 2022. “At least a dozen” was NVIDIA’s stated support expectation at the time, not a verified count of current shipping configurations.
#1 Best Overall
- Standard Memory: 40 GB
- Host Interface: PCI Express 4.0
- Cooler Type: Passive Cooler
- Product Type: Graphics Card
A liquid-cooled accelerator is not a drop-in replacement for any server with an open PCIe slot. The host system must provide the appropriate card power, mechanical clearance, plumbing, cold plate or manifold connections, pumps and heat-rejection capacity. Buyers need the complete server and cooling design specified by an OEM or system integrator.
How the H100 follow-on fit in
NVIDIA said liquid cooling would also be available in an HGX H100 server and as an H100 PCIe card in early 2023. This was a forward-looking product plan presented in 2022, not a current availability guarantee.
Rank #2
- 24GB Video Memory
- Fourth Generation Tensor Cores
- HALF HEIGHT BRACKET ONLY
The H100 is based on NVIDIA’s Hopper architecture, which succeeded the Ampere generation used by the A100. An H100 PCIe card, an H100 SXM module and an HGX H100 system are different configurations with different host, interconnect and cooling requirements.
A100 versus H100: the comparisons that matter
| Comparison | A100 liquid-cooled PCIe | H100 liquid-cooled options |
|---|---|---|
| Architecture | Ampere | Hopper |
| Announced configuration | A100 80GB PCIe accelerator | H100 PCIe card and HGX H100 server options were planned |
| Announcement timing | May 23, 2022; sampling then, summer availability expected | Early 2023 availability was described as a plan |
| System form | PCIe card requiring a compatible liquid-cooled server | PCIe card or HGX system; these are not interchangeable implementations |
| Current price and stock | Not established by the announcement | Not established by the announcement |
Choosing between generations requires more than comparing model names. Evaluate the workload’s memory and performance needs, the server’s PCIe and power support, software compatibility, networking and interconnect requirements, and whether the facility can operate the specified liquid loop.
Rank #3
- Graphics Card Interface: Pci E
What NVIDIA claimed about efficiency
NVIDIA said the launch-era liquid-cooled cards could deliver the same performance with less energy, and described higher performance at the same energy as a future possibility. Those are NVIDIA’s claims. The announcement materials available here do not provide independent, card-specific measurements proving realized energy savings or performance gains for liquid-cooled A100 or H100 PCIe cards.
That distinction matters because cooling efficiency depends on the entire installation: pumps, heat exchangers, facility water or coolant temperature, airflow, workload, power limits and server design. A card-level claim cannot by itself predict a data center’s total power usage.
Rank #4
- Item Package Dimension -14.7L X 8.8W X 3.4H Inches
- Item Package Weight - 2.4 Pounds
- Item Package Quantity - 1
- Product Type - Video Card
Zac Smith, head of edge infrastructure at Equinix, said the company’s lab was receiving its first liquid-cooled GPU and that customers wanted sustainable ways to run AI. He also argued that wattage alone is insufficient and that operators should consider performance relative to carbon impact. These statements express deployment priorities, not independent benchmark results.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Will a liquid-cooled H100 or A100 work in an existing server?
Not necessarily. PCIe describes the host interface, not universal server compatibility. Before ordering, verify all of the following with NVIDIA or the OEM:
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Best Value
- The exact accelerator model, board power and auxiliary power connectors.
- Server qualification for the card’s dimensions, firmware and PCIe generation.
- Required cold plate, manifold, tubing and quick-disconnect hardware.
- Pump, coolant, filtration and heat-rejection specifications.
- Operating temperature, leak-detection and maintenance procedures.
- Validated drivers, virtualization support and the intended AI, analytics or HPC software stack.
An SXM GPU or HGX platform should not be treated as equivalent to a PCIe card simply because it uses the same accelerator family. The complete platform determines what can be installed and operated safely.
What is known—and what is not
The historical record establishes the May 2022 A100 80GB PCIe announcement, NVIDIA’s summer-2022 general-availability expectation, its statement about at least a dozen system builders and its early-2023 H100 liquid-cooling plan. It does not establish present-day inventory, pricing, a definitive OEM compatibility list or independently measured savings for a particular deployment.
Organizations evaluating a purchase should request a current quote and qualification statement directly from NVIDIA or an authorized OEM. The supplier should identify the exact card, server model, cooling-loop configuration, warranty terms and delivery status rather than relying on the 2022 or 2023 forecasts.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →
