NVIDIA at CES 2025
Jensen Huang, CEO of NVIDIA, took the stage at CES 2025 to announce and launch the new NVIDIA GeForce RTX 50-series GPUs as part of the RTX Blackwell family of GPUs including the RTX 5090, RTX 5080, RTX 5070Ti, and RTX 5070.
GPUs, or graphics processing units, are an essential component of video games and general-purpose processing. Thanks to the CUDA language, GPUs enable high-performance computing opportunities in engineering, machine learning, AI, and much, much more. Exxact will have the RTX 50-series GPUs in its configurable workstations and solutions for order and pre-order quoting.
NVIDIA delivers the newest generation of NVIDIA RTX Blackwell GPUs to consumers and professionals to achieve even higher levels of performance. We go over the new Blackwell Architecture, tech, and specs of the new RTX 50-series GPUs and highlight some groundbreaking changes.
GeForce RTX Blackwell Technological Advancements
The RTX Blackwell series of GPUs features groundbreaking advancements in CUDA cores, Tensor Cores, and RT Cores. Key features include:
- Up to 92 billion transistors and over 20,000 CUDA Cores
- Updated streaming multiprocessor for better AI processing
- New GDDR7 memory for higher memory bandwidths
- 2-slot width design in the RTX 5090
The next-generation CUDA, Tensor, and RT cores, along with improved streaming multiprocessors enhance AI integration in the RTX Blackwell GPUs. This advancement enables DLSS 4.0, a frame generation technology exclusive to RTX Blackwell that uses AI to generate over 75% of inferred frames. This technology increases visual smoothness and reduces GPU load in gaming, real-time rendering, and 3D modeling. DLSS benefits various industries, from triple-A gaming to architectural design and VR applications, where powerful hardware must fit in compact spaces.
At Exxact, we configure workstations using both consumer and workstation-class GPUs. The RTX 5090's 2-slot design did surprise us. Slot width refers to the GPU's height within PCIe slots, which are typically separated by two units in workstations and servers with multi-GPU setups. The 2-slot form factor has been reserved for professional RTX and data center GPUs.
The NVIDIA RTX 5090 Founder's Edition 2-slot design enables dual or quad GPU configurations—ideal for large-scale rendering and complex engineering simulation workloads. However, board partners like ASUS, GIGABYTE, and ZOTAC designs will not initially feature the 2-slot width form factor.
Model | Host Interface | Memory | Memory Bandwidth | CUDA Cores | Tensor Cores | RT Cores | Single Precision (TFLOPS) | Double Precision (TFLOPS) | TDP | MSRP |
---|---|---|---|---|---|---|---|---|---|---|
NVIDIA RTX 5090 | PCIe 5.0 x16 | 32GB GDDR7 | 1792 GB/s | 21,760 | 680 | 170 | 104.8 | 1.64 | 575W | $1999 |
NVIDIA RTX 5080 | PCIe 5.0 x16 | 16GB GDDR7 | 960 GB/s | 10752 | 336 | 84 | 56.3 | 0.88 | 360W | $999 |
NVIDIA RTX 5070Ti | PCIe 5.0 x16 | 16GB GDDR7 | 896 GB/s | 8960 | 280 | 70 | 43.9 | 0.69 | 300W | $749 |
NVIDIA RTX 5070 | PCIe 5.0 x16 | 12GB GDDR7 | 672 GB/s | 6144 | 192 | 48 | 30.9 | 0.48 | 250W | $579 |
NVIDIA RTX 4090 | PCIe 4.0 x16 | 24GB GDDR6X | 1001 TB/s | 16384 | 512 (Last Gen) | 128 (Last Gen) | 82.6 | 1.29 | 450W | $1499 |
The performance benchmark tests released by NVIDIA state a 2x performance over each GPU’s last-generation counterpart. However, these tests show performance numbers with the RTX 50-series using DLSS 4.0 whereas the RTX 4090 can only utilize DLSS 3.5. While FPS is an indicator of performance, the relative performance numbers are inflated due to a software difference. DLSS 4.0 generated three extra frames versus DLSS 3.5 only assisting in generating one extra. Further performance numbers from 3rd party sources will help shed light on the performance of the new RTX Blackwell GPUs. We have no doubt these cards will be highly competitive and new record-setting numbers.
Final Remarks and Key Takeaways
With the announcement of GeForce RTX 50-series GPUs, we eagerly anticipate the upcoming professional RTX GPUs. These will target professional workloads like engineering simulation, AI training and deployment, and more. Both the GeForce RTX 5090 and the entire RTX Blackwell family demonstrate exceptional performance for productivity tasks.
Here are the key takeaways from the new GeForce RTX 50-series launch:
- Refer to third-party sources for independent performance benchmarks
- NVIDIA's provided performance benchmarks show FPS numbers with frame generation enabled, indicating expected performance relative to previous generation cards
- The RTX 5080, 5070Ti, and 5070 offer excellent performance per dollar and can handle most everyday computing tasks
- All Founder's Edition RTX 50-series GPUs (except the 5070Ti, which won't have a Founder's Edition) feature a dual-slot design fitting most cases. Third-party models will maintain thicker heatsink designs
- The RTX 5090's TDP has increased significantly from 450W to 575W—check your power supply before upgrading. Other models remain within 50W of their predecessors
For questions about release dates or to get a free quote on a workstation featuring the new RTX Blackwell GPUs, contact Exxact today. Our configurator now includes the new RTX 50-series GPUs for pre-order and quoting. We custom configure and deliver solutions specifically designed to accelerate your innovation.
Accelerate Training with an Exxact Multi-GPU Workstation
With the latest CPUs and most powerful GPUs available, accelerate your deep learning and AI project optimized to your deployment, budget, and desired performance!
Configure NowNVIDIA Launches RTX Blackwell 50-Series GPUs - RTX 5090, RTX 5080, RTX 5070Ti, RTX 5070
NVIDIA at CES 2025
Jensen Huang, CEO of NVIDIA, took the stage at CES 2025 to announce and launch the new NVIDIA GeForce RTX 50-series GPUs as part of the RTX Blackwell family of GPUs including the RTX 5090, RTX 5080, RTX 5070Ti, and RTX 5070.
GPUs, or graphics processing units, are an essential component of video games and general-purpose processing. Thanks to the CUDA language, GPUs enable high-performance computing opportunities in engineering, machine learning, AI, and much, much more. Exxact will have the RTX 50-series GPUs in its configurable workstations and solutions for order and pre-order quoting.
NVIDIA delivers the newest generation of NVIDIA RTX Blackwell GPUs to consumers and professionals to achieve even higher levels of performance. We go over the new Blackwell Architecture, tech, and specs of the new RTX 50-series GPUs and highlight some groundbreaking changes.
GeForce RTX Blackwell Technological Advancements
The RTX Blackwell series of GPUs features groundbreaking advancements in CUDA cores, Tensor Cores, and RT Cores. Key features include:
- Up to 92 billion transistors and over 20,000 CUDA Cores
- Updated streaming multiprocessor for better AI processing
- New GDDR7 memory for higher memory bandwidths
- 2-slot width design in the RTX 5090
The next-generation CUDA, Tensor, and RT cores, along with improved streaming multiprocessors enhance AI integration in the RTX Blackwell GPUs. This advancement enables DLSS 4.0, a frame generation technology exclusive to RTX Blackwell that uses AI to generate over 75% of inferred frames. This technology increases visual smoothness and reduces GPU load in gaming, real-time rendering, and 3D modeling. DLSS benefits various industries, from triple-A gaming to architectural design and VR applications, where powerful hardware must fit in compact spaces.
At Exxact, we configure workstations using both consumer and workstation-class GPUs. The RTX 5090's 2-slot design did surprise us. Slot width refers to the GPU's height within PCIe slots, which are typically separated by two units in workstations and servers with multi-GPU setups. The 2-slot form factor has been reserved for professional RTX and data center GPUs.
The NVIDIA RTX 5090 Founder's Edition 2-slot design enables dual or quad GPU configurations—ideal for large-scale rendering and complex engineering simulation workloads. However, board partners like ASUS, GIGABYTE, and ZOTAC designs will not initially feature the 2-slot width form factor.
Model | Host Interface | Memory | Memory Bandwidth | CUDA Cores | Tensor Cores | RT Cores | Single Precision (TFLOPS) | Double Precision (TFLOPS) | TDP | MSRP |
---|---|---|---|---|---|---|---|---|---|---|
NVIDIA RTX 5090 | PCIe 5.0 x16 | 32GB GDDR7 | 1792 GB/s | 21,760 | 680 | 170 | 104.8 | 1.64 | 575W | $1999 |
NVIDIA RTX 5080 | PCIe 5.0 x16 | 16GB GDDR7 | 960 GB/s | 10752 | 336 | 84 | 56.3 | 0.88 | 360W | $999 |
NVIDIA RTX 5070Ti | PCIe 5.0 x16 | 16GB GDDR7 | 896 GB/s | 8960 | 280 | 70 | 43.9 | 0.69 | 300W | $749 |
NVIDIA RTX 5070 | PCIe 5.0 x16 | 12GB GDDR7 | 672 GB/s | 6144 | 192 | 48 | 30.9 | 0.48 | 250W | $579 |
NVIDIA RTX 4090 | PCIe 4.0 x16 | 24GB GDDR6X | 1001 TB/s | 16384 | 512 (Last Gen) | 128 (Last Gen) | 82.6 | 1.29 | 450W | $1499 |
The performance benchmark tests released by NVIDIA state a 2x performance over each GPU’s last-generation counterpart. However, these tests show performance numbers with the RTX 50-series using DLSS 4.0 whereas the RTX 4090 can only utilize DLSS 3.5. While FPS is an indicator of performance, the relative performance numbers are inflated due to a software difference. DLSS 4.0 generated three extra frames versus DLSS 3.5 only assisting in generating one extra. Further performance numbers from 3rd party sources will help shed light on the performance of the new RTX Blackwell GPUs. We have no doubt these cards will be highly competitive and new record-setting numbers.
Final Remarks and Key Takeaways
With the announcement of GeForce RTX 50-series GPUs, we eagerly anticipate the upcoming professional RTX GPUs. These will target professional workloads like engineering simulation, AI training and deployment, and more. Both the GeForce RTX 5090 and the entire RTX Blackwell family demonstrate exceptional performance for productivity tasks.
Here are the key takeaways from the new GeForce RTX 50-series launch:
- Refer to third-party sources for independent performance benchmarks
- NVIDIA's provided performance benchmarks show FPS numbers with frame generation enabled, indicating expected performance relative to previous generation cards
- The RTX 5080, 5070Ti, and 5070 offer excellent performance per dollar and can handle most everyday computing tasks
- All Founder's Edition RTX 50-series GPUs (except the 5070Ti, which won't have a Founder's Edition) feature a dual-slot design fitting most cases. Third-party models will maintain thicker heatsink designs
- The RTX 5090's TDP has increased significantly from 450W to 575W—check your power supply before upgrading. Other models remain within 50W of their predecessors
For questions about release dates or to get a free quote on a workstation featuring the new RTX Blackwell GPUs, contact Exxact today. Our configurator now includes the new RTX 50-series GPUs for pre-order and quoting. We custom configure and deliver solutions specifically designed to accelerate your innovation.
Accelerate Training with an Exxact Multi-GPU Workstation
With the latest CPUs and most powerful GPUs available, accelerate your deep learning and AI project optimized to your deployment, budget, and desired performance!
Configure Now