iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
At its March 18, 2025 keynote, NVIDIA CEO Jensen Huang presented a broad vision for AI infrastructure, desktop AI development, robotics and automotive technology. Major announcements included the Blackwell Ultra data-center platform, DGX Spark and DGX Station desktop systems, the GR00T N1 robotics model, and a collaboration with General Motors. The announcements describe NVIDIA’s plans and reported results—not independent benchmark findings or proof that future products and partnerships have since shipped.
What NVIDIA announced at GTC 2025
GTC ran March 17–21, 2025, and the keynote on March 18 ranged across data-center computing, AI software, networking, robotics and vehicles. Rather than unveiling a single consumer product, NVIDIA outlined systems and software aimed at building and operating AI infrastructure, developing AI models, and applying AI to robots and cars.
| Announcement | What it is | Intended use |
|---|---|---|
| Blackwell Ultra | Data-center platform, including GB300 NVL72 and HGX B300 NVL16 | AI factories and demanding reasoning, agentic AI and physical AI workloads |
| DGX Spark and DGX Station | Desktop Grace Blackwell AI systems | Model prototyping, fine-tuning and inference for developers, researchers and data scientists |
| Isaac GR00T N1 and related tools | Humanoid-robot foundation model, simulation and synthetic-data technology | Training and developing general-purpose robots |
| NVIDIA–GM collaboration | Work spanning vehicles, factories and robotics | Manufacturing simulation, future advanced driver-assistance systems and in-cabin safety experiences |
These are different product classes, not direct substitutes. A rack-scale data-center system, a desktop development computer, and a robotics model cannot be meaningfully placed in one performance ranking.
Free tools Windows power users keep installed
One-click scans. No signup required.
What is Blackwell Ultra?
Blackwell Ultra is NVIDIA’s announced next evolution of its Blackwell AI factory platform. The March 18 announcement named two systems: the rack-scale GB300 NVL72 and HGX B300 NVL16. It was a data-center platform announcement, not a launch of a consumer graphics card or desktop chip.
#1 Best Overall
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
NVIDIA said the GB300 NVL72 would deliver 1.5 times the AI performance of the GB200 NVL72. That is NVIDIA’s stated comparison; the announcement is not an independent benchmark. NVIDIA also described a 50-fold increase in Blackwell’s “revenue opportunity” for AI factories compared with factories built with Hopper. This is a revenue-opportunity claim, not a performance improvement or realized revenue figure.
At the keynote, Huang said, “AI has made a giant leap — reasoning and agentic AI demand orders of magnitude more computing performance.” NVIDIA positioned Blackwell Ultra for those workloads, as well as physical AI. The company said partner products were expected in the second half of 2025; that was a dated expectation, not confirmation of subsequent availability.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5070 Ti
- Integrated with 16GB GDDR7 256bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
What are DGX Spark and DGX Station?
DGX Spark and DGX Station are Grace Blackwell computers in desktop form factors, intended for AI development rather than ordinary home computing. NVIDIA presented them as systems for prototyping, fine-tuning and inference. DGX Spark uses NVIDIA’s GB10 Grace Blackwell Superchip.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minute- DGX Spark: NVIDIA announced up to 1,000 trillion operations per second of AI compute. The company said reservations opened March 18, 2025, and named ASUS, Dell, HP and Lenovo among system builders.
- DGX Station: NVIDIA announced 784GB of coherent memory space and said it expected manufacturing-partner systems later in 2025.
Both figures are announced specifications, not independent benchmark results. The keynote-era reservation and availability statements do not establish current retail availability or a current listing. Huang described the rationale this way: “AI has transformed every layer of the computing stack. It stands to reason a new class of computers would emerge — designed for AI-native developers and to run AI-native applications.”
Rank #3
- Powered by the NVIDIA Blackwell architecture and DLSS 4. System Requirements: Minimum 850W PSU with 16-pin 12V-2x6 (12VHPWR) connector required. Verify before purchasing.
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability. Compatibility: 348mm (13.7") length, 3.6 slots, 4.3 lbs. Confirm case clearance and slot spacing. GPU bracket included.
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
What is NVIDIA Isaac GR00T N1?
Isaac GR00T N1 is an announced open and customizable foundation model for generalized humanoid reasoning and skills. NVIDIA described a dual-system design: a faster model for action and a slower model for deliberate reasoning and planning. It is a robotics development technology, not a robot offered as a finished consumer product.
NVIDIA also announced synthetic-data and simulation tools to support robot development. One is Newton, an open-source physics engine being developed with Google DeepMind and Disney Research. NVIDIA said its tools generated 780,000 synthetic trajectories—described as equivalent to 6,500 hours of human demonstrations—in 11 hours. It also reported a 40% improvement in GR00T N1 performance when synthetic data was combined with real data, compared with using real data alone. These are NVIDIA-reported results under the conditions described by the company, not independent evaluations.
Rank #4
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5060
- Integrated with 8GB GDDR7 128bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
Huang summarized the ambition by saying, “The age of generalist robotics is here.” That is a statement about NVIDIA’s vision, not evidence that general-purpose humanoid robots are already broadly deployed.
What did NVIDIA and GM announce about cars?
NVIDIA and General Motors announced a collaboration spanning next-generation vehicles, factories and robots. GM would use NVIDIA Omniverse and Cosmos for manufacturing-model training, and NVIDIA DRIVE AGX hardware for future advanced driver-assistance systems and in-cabin safety experiences.
Best Value
- Powered by the NVIDIA Blackwell architecture and DLSS 4 OC mode: 2640MHz/Default mode: 2610MHz (Boost Clock)
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.125-slot design with massive fin array optimized for airflow from three Axial-tech fans
- Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads
The announcement did not identify a specific production vehicle or launch date, and it did not announce a retail self-driving car. The distinction matters: advanced driver assistance and in-cabin safety features are not the same claim as a vehicle that can drive itself in all conditions.
Huang also said, “We build technology that almost every single self-driving car company uses.” This is his keynote statement, as quoted in NVIDIA’s recap; it is not an independently measured market-share statistic.
What else was part of the keynote?
The keynote and NVIDIA’s event materials also covered AI inference software, the NVIDIA Dynamo software stack, an Omniverse digital-twin blueprint for planning AI factories, Spectrum-X and Quantum-X photonics networking, and the Llama Nemotron model family. Together, these announcements reinforced the event’s focus on the infrastructure and software surrounding AI systems, not just chips or standalone applications.
Across the keynote, performance comparisons, demonstration results, partner plans and future availability statements came from NVIDIA. They should be read with that scope: they describe what the company announced or reported on March 18, 2025, rather than independently verified results or confirmation of later outcomes.
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.
