Buy the NVIDIA RTX 2000 Ada Generation 16 GB: half-height 70 W card, price by quote

RTX 2000 Ada price by written quote, supplied across the EU

The NVIDIA RTX 2000 Ada Generation is a half-height workstation graphics card with 16 GB of GDDR6 memory with ECC that takes its 70 W from the PCIe slot. We supply it across the EU, to IT departments and system builders alike. This page carries no price figure. Tell us how many cards you need, and a written quote with the price, the lead time and the warranty terms is with you within one business day.

The subject here is the card as a part. Every figure NVIDIA publishes for it is in our RTX 2000 Ada specs. A workstation built around Ada cards is a different order, which our page to buy an RTX Ada workstation describes.

A card that is ordered by the seat count

The usual request for this card names a number of seats. An IT department refreshes its compact CAD workstations, or a system builder fits one card type into a series of small-form-factor machines. Three properties explain why the RTX 2000 Ada is chosen for that.

The first is the format. The card is 2.7 in high and 6.6 in long, takes two slot widths and draws 70 W, which are the figures of the RTX A2000 before it. NVIDIA's blog of 12 February 2024 calls the RTX A2000 12GB the previous generation and puts the new card at the same 70 W. By these figures a seat that held an RTX A2000 has the slot and the power budget for this card, and our comparison of the RTX 2000 Ada and the RTX A2000 goes through what changes.

The second is 16 GB in a card without a power cable. On the software makers' requirement pages read on 10 November 2026, a 16 GB card meets every GPU memory figure we found, including the 16 GB lines of Lumion 2026 for its high-end tier, of DaVinci Resolve Studio 21.2 for its advanced AI tools and of ArcGIS Pro 3.7 for deep learning. That statement concerns memory only. Lumion and Twinmotion also set a benchmark score, and the graphics lines of AutoCAD and Inventor add a "Bandwidth" value that Autodesk does not define.

The third is the certification row. Dassault's SOLIDWORKS 2026 certification list, read 10 November 2026, has rows for the RTX 2000 Ada in desktop workstations, among them Lenovo's ThinkStation P3 Ultra Gen 2 and P3 Tower Gen 2 and HP's Z2 Mini G1i and Z2 SFF G1i. What a row certifies is a computer model together with one card and one driver. Laptop rows appear under the same card name, so the row that counts is the one with your exact workstation model. How such a row is read is explained in our guide to NVIDIA cards for SOLIDWORKS.

A company that runs one validated image on Ada cards has a fourth reason: the same card in the new machines keeps one driver and one configuration. We supply the RTX 2000 Ada across the EU; send us the seat count and the workstation models and we reply with a written quote within one business day.

The figures to hold against the chassis

NVIDIA RTX 2000 Ada Generation datasheet (2999417, January 2024); power cable from NVIDIA's quick start guide for RTX Ada Generation cards; sessions from NVIDIA's Video Encode and Decode GPU Support Matrix of 16 October 2026; MIG User Guide (4 November 2026), vGPU supported-GPU list (2 October 2026), RTX PRO Sync Solutions page (15 October 2026); driver release notes 621.14, 597.30 and 582.99.

NVIDIA states the size in inches only. In metric terms 2.7 by 6.6 in comes to roughly 69 by 168 mm by our conversion, and neither millimetres nor a weight are published. The height of 2.7 in is half-height, the format usually called low profile; the NVIDIA documents we read for this card give the dimensions and do not use that term. Our overview of low-profile NVIDIA RTX cards sets it beside the other 2.7 in boards.

One practical detail concerns the monitors. All four outputs are Mini DisplayPort, and NVIDIA's quick start guide lists one adapter from Mini DisplayPort to DisplayPort in the box. A seat with three or four DisplayPort monitors therefore needs cables or adapters for the remaining outputs, and they are best planned together with the cards.

Slot, power and driver in the workstation

The card needs a half-height position that is two slot widths wide and takes a 6.6 in board. The quick start guide lists a PCI Express x16 slot as the requirement. The datasheet gives the link as PCIe 4.0 x8 and adds in a footnote that the card uses a full-length PCIe x8 interface.

No cable runs to the card. The datasheet puts the total board power at 70 W, and the quick start guide for the Ada cards counts the RTX 2000 among the models that do not use a power cable. Those 70 W lie inside what NVIDIA's power guidelines for workstation products of July 2023 call the typical rating of a PCIe slot, 75 W; that document is older than the card and does not list it. A wattage for the power supply is not published. What can be checked is a sum, the machine's present load plus 70 W, against the power-supply rating its maker states; our article on RTX power connectors and power supplies covers the method.

On the software side the card runs on the NVIDIA RTX Enterprise Production Branch driver. On Windows that means, as of 10 November 2026, release 621.14 in branch R620, 597.30 in R595 or 582.99 in R580; the Linux releases of the same three branches list the card as well. Windows 11 is covered from 22H2 to 25H2 and Windows 10 in version 22H2. For Windows 10 the release notes add a limit: Release 620 will be the last branch to support it. For a fleet that still runs Windows 10, the driver branch is therefore part of the purchase decision.

If the cards go into machines you already run, name the workstation models and the operating system, and the written quote covers the cards for exactly those machines.

The RTX 2000 Ada with 8 GB is a laptop GPU

NVIDIA uses the name for more than one product, and listings do not always keep them apart.

NVIDIA RTX 2000 Ada Generation datasheet (2999417, January 2024); NVIDIA's "Compare Professional Laptop GPUs" page, read 10 November 2026.

Only the first row is a PCIe card. The laptop GPU, announced on 21 March 2023 and so before the desktop card, is a component of mobile workstations and cannot be ordered as a board. NVIDIA also lists an embedded GPU of the same name with 8 GB. A request that says "RTX 2000 Ada Generation, desktop card, 16 GB" leaves no room for a mix-up.

A second source of confusion is the RTX 2000E Ada. It is a PCIe card as well, with the same 16 GB, but it is specified at 50 W of total board power and occupies one slot. The letter E in an order line changes the part.

What the RTX 2000 Ada price depends on

We answer a price request in writing within one business day. The quote gives the price for the number of cards you asked for together with the warranty terms; quantities and lead time are confirmed in writing with the quote. The card is available to order.

Three things move the price: the quantity, the delivery address and the date of the order. List prices and MSRP figures are not repeated on this page. If the card is to arrive inside a workstation or a server, it becomes one line of a quote for that system; our page on workstations and servers built to order shows what such a system contains.

We can quote at once when a request gives the number of cards, the product (the 16 GB desktop card), the machines they are meant for and the delivery address together with the company to be invoiced. For a roll-out in stages, add the dates of the stages, because the date of an order is one of the three factors.

How an order for a fleet of seats runs

The purchase is made on one European contract, and the invoice is issued in Austria. Cards travel to any address in the EU; where the destination lies outside it, our logistics partner takes over. If a card arrives dead it is replaced, and later warranty cases are handled through us under the manufacturer warranty. The first consultation is free.

We supply larger quantities as well. IT departments and system builders name the total and the machines, and the quote covers the whole quantity. How such a request is put together is the subject of our page on NVIDIA RTX cards in volume for system builders.

RTX PRO 2000 Blackwell, RTX 2000E Ada or RTX A1000 instead

Three cards we supply share the 2.7 in height with the RTX 2000 Ada, and each is the better order in one situation.

NVIDIA datasheets: RTX A1000 (April 2024), RTX 2000E Ada Generation (4328932, October 2025), RTX 2000 Ada Generation (2999417, January 2024), RTX PRO 2000 Blackwell (5322351, June 2026).

The NVIDIA RTX PRO 2000 Blackwell fits the same slot and also works without a power cable. It has 4,352 CUDA cores, a PCIe 5.0 x8 link, Mini DisplayPort 2.1b outputs and Tensor Cores of the fifth generation with FP4 support. Puget Systems' professional GPU roundup of 18 December 2025, run on a Ryzen 9 9950X3D, measured it 44 per cent faster than the RTX 2000 Ada in V-Ray GPU and 31 per cent faster in Blender Cycles. For new seats without a tie to Ada it is the card to quote first: buy the RTX PRO 2000 Blackwell, RTX PRO 2000 Blackwell specs, and the existing comparison of the RTX PRO 2000 and the RTX 2000 Ada.

The NVIDIA RTX 2000E Ada Generation keeps the 16 GB and the 2,816 CUDA cores in a single slot at 50 W. NVIDIA's datasheet gives it 8.9 TFLOPS in single precision against 12.0 for the RTX 2000 Ada. It is the order for chassis with one free slot: buy the RTX 2000E Ada, RTX 2000E Ada specs, and the direct comparison RTX 2000 Ada vs RTX 2000E Ada.

The NVIDIA RTX A1000 is the step down: 8 GB, a single slot and 50 W. NVIDIA states no ECC for it, and its support matrix gives it at most 8 NVENC sessions and no AV1 encode. Details are on the page to buy the RTX A1000 and in the RTX A1000 specs.

Above 16 GB the half-height format continues with the RTX PRO 4000 Blackwell SFF Edition, a 70 W card with 24 GB that our comparison of the 70 W cards covers. If the choice between the Ada card and the Blackwell card is still open, write to us with the applications and the number of seats, and we quote both.

FAQ