Nvidia RTX 40 Series Index
A sourced specification index of every cataloged Nvidia RTX 40-series variant, including memory, throughput, and power.
Why the RTX 40 family needs an index
The RTX 40 series arrived as a broad family rather than a uniform step. Products sharing the label occupied very different power, memory, and system classes, and later additions extended the generation after its first launch wave. A list ordered by date makes that expansion visible. It also prevents a flagship specification from becoming shorthand for every card built from the generation’s architecture.
Under the checked source-section rule, this page contains 21 variants GPU variants launched from 2022 through 2024. It counts variants: distinct catalog configurations remain distinct even when their names share a familiar root. Nvidia accounts for 21 variants; AMD accounts for 0 variants and Intel 0 variants because neither belongs inside an Nvidia series set. Partner-board editions are not independently enumerated unless the source catalog treats them as a variant.
Specifications describe constraints, not verdicts
The table aligns memory capacity and type, published FP32 throughput, TDP, process node, and transistor count where available. Together they show how the architecture was adapted across the range. They do not reproduce application performance. Clock behavior, memory bandwidth, cache, drivers, cooling, and workload shape realized results, while theoretical throughput remains an upper-bound specification. TDP is likewise a design class and should not be read as a guaranteed meter reading.
The generational view is valuable because segmentation decisions become legible. Similar silicon can appear with resources disabled, a different memory interface, or a lower power target. Mobile implementations can pursue another balance entirely. Those choices explain why buying by family name alone is unreliable and why a complete variant table is more informative than a ranked shortlist. This page makes no claims about present retail availability or value. It answers the historical catalog question, keeps missing fields missing, and supplies revision provenance so the membership and specifications can be checked against the source.
Catalog variants
| Vendor | Variant | Year | VRAM | Memory | FP32 GFLOPS | TDP | Connector evidence | Node | Transistors M |
|---|---|---|---|---|---|---|---|---|---|
| nvidia | GeForce RTX 4090 | 2022 | 24,576 MB | GDDR6X | 73,236 | 450 W | not recorded by source | — | 76,300 |
| nvidia | GeForce RTX 4080 | 2022 | 16,384 MB | GDDR6X | 42,900 | 320 W | not recorded by source | — | 45,900 |
| nvidia | GeForce RTX 4070 Ti | 2023 | 12,288 MB | GDDR6X | 35,482 | 285 W | not recorded by source | — | 35,800 |
| nvidia | GeForce RTX 4080 Mobile/ Laptop | 2023 | 12,288 MB | GDDR6 | — | — | not recorded by source | — | 35,800 |
| nvidia | GeForce RTX 4090 Mobile/ Laptop | 2023 | 16,384 MB | GDDR6 | — | — | not recorded by source | — | 45,900 |
| nvidia | GeForce RTX 4050 Mobile/ Laptop | 2023 | 6,144 MB | GDDR6 | — | — | not recorded by source | — | 18,900 |
| nvidia | GeForce RTX 4060 Mobile/ Laptop | 2023 | 8,192 MB | GDDR6 | — | — | not recorded by source | — | 18,900 |
| nvidia | GeForce RTX 4070 Mobile/ Laptop | 2023 | 8,192 MB | GDDR6 | — | — | not recorded by source | — | 22,900 |
| nvidia | GeForce RTX 4070 | 2023 | 12,288 MB | GDDR6X | 22,610 | 200 W | not recorded by source | — | 35,800 |
| nvidia | GeForce RTX 4060 Ti | 2023 | 8,192 MB | GDDR6 | 20,106 | 160 W | not recorded by source | — | 22,900 |
| nvidia | GeForce RTX 4060 | 2023 | 8,192 MB | GDDR6 | 11,244 | 115 W | not recorded by source | — | 18,900 |
| nvidia | GeForce RTX 4060 Ti 16 GB | 2023 | 16,384 MB | GDDR6 | 20,106 | 165 W | not recorded by source | — | 22,900 |
| nvidia | GeForce RTX 4090 D | 2023 | 24,576 MB | GDDR6X | 66,540 | 425 W | not recorded by source | — | 76,300 |
| nvidia | GeForce RTX 4070 Super | 2024 | 12,288 MB | GDDR6X | 28,385 | 220 W | not recorded by source | — | 35,800 |
| nvidia | GeForce RTX 4070 Ti Super | 2024 | 16,384 MB | GDDR6X | 39,537 | 285 W | not recorded by source | — | 45,900 |
| nvidia | GeForce RTX 4080 Super | 2024 | 16,384 MB | GDDR6X | 47,002 | 320 W | not recorded by source | — | 45,900 |
| nvidia | GeForce RTX 4070 | 2024 | 12,288 MB | GDDR6X | 22,610 | 200 W | not recorded by source | — | 45,900 |
| nvidia | GeForce RTX 4060 | 2024 | 8,192 MB | GDDR6 | 11,244 | 115 W | not recorded by source | — | 22,900 |
| nvidia | GeForce RTX 4060 Ti | 2024 | 8,192 MB | GDDR6 | 20,106 | 160 W | not recorded by source | — | 35,800 |
| nvidia | GeForce RTX 4070 Ti Super | 2024 | 16,384 MB | GDDR6X | 39,537 | 285 W | not recorded by source | — | 76,300 |
| nvidia | GeForce RTX 4070 GDDR6 | 2024 | 12,288 MB | GDDR6 | 22,610 | 200 W | not recorded by source | — | 35,800 |