- Nvidia Jetson Agx Thor debuts with Blackwell GPU, 128 GB memory and 1 TB storage
- Early reviews describe a skilled platform that offers serious performance improvements over Jetson Orin
- Reviewers agree that it will appeal to developers who need powerful hardware for projects
Nvidia recently launched the Jetson Agx Thor Developer Kit, a $ 3,499 platform designed for robotics and edge AI development – and it has had a warm initial reception from reviewers.
In the heart, the Jetson T5000 module is built on the Blackwell architecture, combining a GPU with 2,560 CUDA cores, 96 tensus kernels and a 14 core arm neo -sover CPU.
It is paired with 128 GB LPDDR5X memory that offers over 270 GB per second of bandwidth and 1 TB onboard storage. Connection options include USB C, USB A, HDMI 2.1, WI FI 6E, Bluetooth, Gigabit Ethernet and a 100GBE port.
“GOBS of Horse Power”
The first reviews of the kit are now in, and they suggest that Nvidia has built an impressive opportunity for developers despite its higher price compared to Jetson Orin.
Hothardware’s Testing showed that Jetson Agx Thor was a strong artist, even with limited comparisons. Nvidia’s arm64 containers ran smoothly, but test against other Blackwell hardware was not possible and the older ORIN kit could not complete workloads.
However, the call in capacity was clear, with Orin closer to an RTX 3050 and Thor approached RTX 5070 levels.
Large language models worked well in testing. Seam Hothardware Points out, “LLMs are an area where Jetson is distinguished and it is necessary as humanoid robots are expected to mix language with visual input.”
The review concluded that the kit has “Gobs of Horse Power” for Robotics and AI projects, and noticed, “If you want to run very large AI models in a friendly multi-tasking environment using Nvidia’s software stack, Jetson Agx Thor Developer-Kit Nvidia continues to refine and update its softwarstak with additional edge- “
Servethehome’s Review found performance came close to matching Nvidia’s claims, including 149.1 tokens per day. Second on Llama 3.1 8B against the expected 150.8.
CPU Multi-threaded performance placed it near an AMD Ryzen AI 7 350 or Mac Mini M4, which was considered sufficient in view of its GPU focus.
In Benchmark -Test, Thor, as expected, consisted of consistently Orin across any model. Gains on smaller workloads such as QWEN 2.5-VL 7B and Llama 3.1 8B were modest, with Thor coming about 1.3 times faster.
Deepseek-R1 7B showed a greater improvement about 1.5 times the speed. The most dramatic difference came with Qwen 3 32B inferences, where Thor almost reached five times the performance of Orin, highlighting its strength as they ran larger and more demanding models.
While the current pull can challenge battery systems, Servethehome Completed Thor offers the calculation and memory needed for advanced robotics. It also managed to identify the included 1TB SSD as a WD/SANDISK SN5000S.
Both reviews described Jetson Agx Thor as a skilled step forward for Edge AI and Robotics projects and praised his mix of computing power, memory capacity and developer tools, while noting that software updates will be necessary to unlock all its Poent.
Seam Servethehome Put it, the new set “will sell like hotcakes. If you build advanced next generation robotics, this is the platform you want to do it on.”
