Todays announcement is that the price is $249. Follow the link in the description to find out more, you can buy it for that price at Arrow, Sparkfun or Seeed Studios today. It may take a little while for the new price to ripple through the rest of the retailers.
Thank you for the kind words. It's a good time of the year to have an excuse to take some time for yourself and have something to play with! Thanks for watching!
I grabbed one from a partner on Nvidia's site. Looks fun to play with. For 250$ is a great way to get into massing with Language models and to have it dedicated with its own hardware. The key is the 8GB GPU Ram. I believe its 15-25 watts of power
Which partner? I just checked and they were backordered on Seeed, Sparkfun, and Arrow. Also, Sparkfun would not guarantee the price f they were on backorder.
Is that the old price, or the new price? The price cut is ~ 50%. This is the day of the announcement, I would guess it'll take a little while for the changes to go through the supply chain. Thanks for watching!
@@JetsonHacks No idea if its the old or new price, it was the price I got after clicking on the retailer here in Sweden that Nvidia linked. The Jetson Orin AGX 64GB was $1800 when we got it.
Just so I understand. Make a video, product becomes popular and sells out. Erase video. Make another video when they come back in stock. Make sure that no one knows in the interim that the product exists and they could back order so they can get one as soon as they become available. Got it. I can become both a stock clerk and video maker. Thanks for watching!
Месяц назад+7
That's pretty amazing actually. The natural monopoly Nvidia has is inspirational... 🙂
10 years ago, the 32-bit 192 core Jetson TK1 was introduced for a little under $200. Even with the inflation of the last few years, a 64-bit 1024 core Orin Nano is $250. The difficult part is having enough imagination on how to use and take advantage of these new capabilities.
Unfortunately I jumped into this too early with couple of Jetson Nanos just to learn that although it is quite expensive but not supported any more unlike other Maxwell chips. Plus even the very cumbersome original setup just seem to work rarely (and mostly not) in a Turing Pi 2 board.
I want to save up for one of these, but I wonder if I can use it similarly like ChatGPTs API. I want to connect to the Jetson over network, chatting to it, and supplying tools (functions) that LLama can call on my computer. I kind of just want to use it as an AI server.
You can do that. However, if you are just calling it from a PC it may be better to invest in a better graphics card instead to run your LLM locally on the PC. Thanks for watching!
@JetsonHacks after some testing for my use case (an AI powered terminal shell) I think the 8b models and below are just too unpredictable. I see that a graphics card could benefit me in several ways. I just don't know which cards to consider. I got a 5700xt at the moment which doesn't accelerate ai
@@oceanicdrop If you want to use models larger than 8B, such as 70B, you will need a lot more graphic card memory. There are many alternatives if you have a PC. You will need a NVIDIA graphics card. The Ampere architecture (as in the Orin Nano) is 30X0, such as 3090. The next generation up is 40X0, such as 4090. There are many articles which discuss the tradeoffs in price and configuration, but mostly it's how much money you want to spend and how much power draw is acceptable from your PC. Good luck on your project!
There's a bunch of hype, for sure. But the two take aways are that the price of the Orin Nano Dev Kit is cut in half, and there's a new power mode (25W) which enables better performance. The $250 price puts it into a better value proposition for many, whereas the $500 appealed mostly to professionals. The second part, LLMs. It's a lot more nuanced subject than I thought just a year ago. People have become a lot better on training the smaller models 2B-8B, and by adding RAG you can get some very interesting results. No, it's no 70B level, but for many applications it doesn't have to be. On the other hand, there are many classes of AI models suited towards being inferenced on "only 8GB". Many visual models, navigation, speech, and so on. In some sense, to say you use a general purpose computer for one thing, especially through marketing, can be confusing. Thanks for watching!
@@ProaliusVideos I am planning to. There are a couple of paths for upgrading, so I want to understand what is needed. The easiest one is to create a new SD card, but the system needs to be running JetPack 6.0+ first. If you are planning on booting from the NVMe SSD, the SDK manager is easiest. However, both of these require a new installation. For the SD card, this isn't a problem (you just need a new SD card), but for the SSD it feels like there needs to be at least a backup plan. Thanks for asking, and thanks for watching!
I'm not sure I fully understand the question. The Jetson Orin Nano Super Developer Kit is a stand alone Linux computer. If you want to work with it from another computer, you can SSH into it. If you want to prepare a SD cad for the Orin Nano to work from, you can use any computer with an appropriate SD card slot. However, if you want to develop code on a host computer for the Jetson, you must use a PC running Linux. Most people use Ubuntu (which NVIDIA recommends) for their development machines. However you can also develop on the Orin Nano. It is recommended for serious development (such as kernel compiling, device tree wrangling, and so on) that you have a PC available. Hope that answers your question, and thanks for watching!
Exciting, and thanks for posting a video about it! I'm ordering one now. Curious to see how fast it will run YOLO models 🤖. EDIT: Oh just kidding. It's out of stock already 😿
Can someone PLEASE give a "REAL WORLD" concrete example of what this thing can do for me? I'm brand new to this but am thinking about buying one. I google it and most answers are nonsense. Thank U!
If you have to ask, this is probably not for you. You have to decide if this is right for you, no one can tell you for sure. Do more research on Machine Learning and how it is being used in almost every industry. Then, if you are still interested, then look for any other manufacturer that makes a Edge Computer in this price range and power - you won't find any. Good luck.
That's a good question. However it's a little broad to give a good answer to it. It depends on your interests. There are 10s of thousands deployed in professional applications. Many have to do with camera and vision. This ranges from quality control on fast moving production lines to identifying weeds while spraying on a tractor. There are medical appliciances. There are a great many deployed in robotic applications such as drones and autonomous vehicles. There are a large number of makers that use them in their projects. To get started, wander over to the Jetson AI Lab: www.jetson-ai-lab.com/community_articles.html and look through the projects that community members have built. Then go to the Tutorial page where there are dozens of examples to see if there is something that interests you. Thanks for watching!
It's similar in some respects. The CPU cluster uses a similar ARM core (A76 vs A78 for the Jetson), but the Jetson has 2 more cores. Where they differ is that the Jetson has a 1024 CUDA capable GPU. The GPU is programmable, making it much more useful for running AI. Here's a good comparison between the two in an actual application: ruclips.net/video/QHBr8hekCzg/видео.htmlsi=9f8fimEE42h0dBcw Thanks for watching!
Hello, I have a Jetson Nano Developer Kit and when I try to open it I get a black screen and I can't get to the terminal. When I try to format it via SDK Manager it stays at 99.80%.
It should boot to the desktop. You might need to look at the serial debug console to tell you what the issue might be. You should ask for help in the official NVIDIA Jetson Nano forum, where a large group of developers and NVIDIA engineers share their experience:forums.developer.nvidia.com/c/agx-autonomous-machines/jetson-embedded-systems/jetson-nano/76 Thanks for watching!
I solved the problem myself, thank you. I use VirtualBox, then I tried Vmware and it was solved. Please show this message to someone who has this problem
Looks like there were a lot of people interested in Orin Nanos, just as a less expensive price. It'll be interesting to see how the supply gets adjusted. Thanks for watching!
@@buffler1 It looked that way, but I placed a back order for 2 of them with Seeed Studios today. The reply email stated that it would be filled on January 1. Then later in the day, I got an email that they were shipping and will get them this Friday!!!! I then received a DHL email with a tracking number and that I would have to sign for the delivery on Friday. Santa must have intervened, I am sure of it 🧑🎄👍
I don't game, so I'm not a good resource. Here's a link to ETA Prime running games on the Orin Nano: ruclips.net/video/nmZ6fhkFmDY/видео.htmlsi=84ZZ_nynRZMV-4OX&t=474 The performance should be better now with the speed bump and now that the software has had a little time to mature. Thanks for watching!
That's certainly one approach. I suggest that they might be different, superficially similar things. However, if a RPi 5 meets your needs, you should do that. Thanks for watching!
Why do you say that? The new price is announced at 6am PST today, it may take some time to get through the entire supply chain. For example, at Arrow: www.arrow.com/en/products/945-13766-0000-000/nvidia?nvid=em-945-13766-0005-000 or Sparkfun: www.sparkfun.com/products/22098?nvid=em-945-13766-0005-000
also, the Sparkfun site says that they can't guarantee the price when they restock - see their note " Note: If this item is available for backorder it is subject to price changes at any time; additionally, we are unable to guarantee time frame for shipping or availability."
Sparkfun states that they will not guarantee the price when they get inventory. See their note - Note:We do not currently have an estimate of when this product will be back in stock. Note: If this item is available for backorder it is subject to price changes at any time; additionally, we are unable to guarantee time frame for shipping or availability."
Todays announcement is that the price is $249. Follow the link in the description to find out more, you can buy it for that price at Arrow, Sparkfun or Seeed Studios today. It may take a little while for the new price to ripple through the rest of the retailers.
Yea, but not available right now for purchases.
@@plamengflo They had hundreds of them this morning. I wonder what happened? :)
Thank you for covering all aspects: construction, hardware, software and apps.
You are welcome, and thanks for watching!
Nice video Jim! It's a wonderful time of the year to get going on new Jetson development!
Thank you for the kind words. It's a good time of the year to have an excuse to take some time for yourself and have something to play with! Thanks for watching!
My dog is really going to love this!
As to whether my wife buys that...
Dogs love Jetsons! Most wives seem less understanding. Thanks for watching!
Can’t wait for more contents from you on this. Thanks!!
You're putting the pressure on! I have some ideas ... Thanks for watching!
I’ve been referring to your valuable contents in navigating my journey playing around this toy. Again, it’s a gold mine for noobies. 😊
@@johnlp77 Thank you for the kind words. @ $250, this becomes a much more interesting device to experiment with.
Thank you for the video! Great content -- thank you and happy holidays!
You are welcome, thank you for the kind words. Happy holidays to you and yours!
I grabbed one from a partner on Nvidia's site. Looks fun to play with. For 250$ is a great way to get into massing with Language models and to have it dedicated with its own hardware. The key is the 8GB GPU Ram. I believe its 15-25 watts of power
Which partner? I just checked and they were backordered on Seeed, Sparkfun, and Arrow. Also, Sparkfun would not guarantee the price f they were on backorder.
It certainly is a great entry into exploring AI for the price. Thanks for watching!
Only $658 USD here in Sweden. :D The price of a RTX4070 Super.
Is that the old price, or the new price? The price cut is ~ 50%. This is the day of the announcement, I would guess it'll take a little while for the changes to go through the supply chain. Thanks for watching!
@@JetsonHacks No idea if its the old or new price, it was the price I got after clicking on the retailer here in Sweden that Nvidia linked. The Jetson Orin AGX 64GB was $1800 when we got it.
@@AndrewTSq That's the old price, they cut the price in half this morning. It may take some time for the new price to ripple through the supply chain.
can mining?
@@hobbytoanything It can, but I don't know if you could make any money with it.
wow, great, Jim!thanks!
You are welcome, and thanks for watching!
Thank you Sir!
You are welcome, and thanks for watching!
Maybe make a video if they come back into stock at
that price. Great overview but it might as well be vaporware.
Just so I understand. Make a video, product becomes popular and sells out. Erase video. Make another video when they come back in stock. Make sure that no one knows in the interim that the product exists and they could back order so they can get one as soon as they become available. Got it. I can become both a stock clerk and video maker. Thanks for watching!
That's pretty amazing actually. The natural monopoly Nvidia has is inspirational... 🙂
It feels like things are getting a little back to normal. Where prices of computers actually become cheaper over time. Thanks for watching!
@@JetsonHacks agreed. The compute per dollar is still quite amazing. What a time to be alive.
10 years ago, the 32-bit 192 core Jetson TK1 was introduced for a little under $200. Even with the inflation of the last few years, a 64-bit 1024 core Orin Nano is $250. The difficult part is having enough imagination on how to use and take advantage of these new capabilities.
@@JetsonHacks Great point.
Unfortunately I jumped into this too early with couple of Jetson Nanos just to learn that although it is quite expensive but not supported any more unlike other Maxwell chips. Plus even the very cumbersome original setup just seem to work rarely (and mostly not) in a Turing Pi 2 board.
That doesn't sound like a fun experience! Thanks for watching!
Super super Jetson ..❤
I love that song! Thanks for watching.
wunderbar video danke shön
You are welcome. Thank you for the kind words, and thanks for watching!
Greate Video! Is there a deadline for this discounted price?
That's the new price. I'm guessing NVIDIA is lowering it to get ready for Jetson Thor next year. Thanks for watching!
I want to save up for one of these, but I wonder if I can use it similarly like ChatGPTs API. I want to connect to the Jetson over network, chatting to it, and supplying tools (functions) that LLama can call on my computer. I kind of just want to use it as an AI server.
You can do that. However, if you are just calling it from a PC it may be better to invest in a better graphics card instead to run your LLM locally on the PC. Thanks for watching!
@JetsonHacks after some testing for my use case (an AI powered terminal shell) I think the 8b models and below are just too unpredictable. I see that a graphics card could benefit me in several ways. I just don't know which cards to consider. I got a 5700xt at the moment which doesn't accelerate ai
@@oceanicdrop If you want to use models larger than 8B, such as 70B, you will need a lot more graphic card memory. There are many alternatives if you have a PC. You will need a NVIDIA graphics card. The Ampere architecture (as in the Orin Nano) is 30X0, such as 3090. The next generation up is 40X0, such as 4090. There are many articles which discuss the tradeoffs in price and configuration, but mostly it's how much money you want to spend and how much power draw is acceptable from your PC. Good luck on your project!
I don’t understand all the hype. What can do with 8gb unified memory? Of course if you are going to write CUDA code, but running LLMs?
There's a bunch of hype, for sure. But the two take aways are that the price of the Orin Nano Dev Kit is cut in half, and there's a new power mode (25W) which enables better performance. The $250 price puts it into a better value proposition for many, whereas the $500 appealed mostly to professionals.
The second part, LLMs. It's a lot more nuanced subject than I thought just a year ago. People have become a lot better on training the smaller models 2B-8B, and by adding RAG you can get some very interesting results. No, it's no 70B level, but for many applications it doesn't have to be.
On the other hand, there are many classes of AI models suited towards being inferenced on "only 8GB". Many visual models, navigation, speech, and so on. In some sense, to say you use a general purpose computer for one thing, especially through marketing, can be confusing. Thanks for watching!
Excellent! Thanks, can you connect 2 (or more ) of these and share their compute resources? they act as one device?
There are various strategies to aggregate their compute power, such as clustering. Thanks for watching!
Can Windows 11 64bit arm version run on the Jetson Orin Nano Super Developer Kit?
Not at the moment, it's strictly Linux. Thanks for watching!
can you make tutorial of upgrading existing Jetson orin nano 8gb dev kit which has already latest 6.1 jetpack and running to Super mode please?
Are you running from a SD card or from a SSD?
@@JetsonHacks SSD, had issues using SSD installation with SDK manager
@@ramkumarkoppu Thanks, this is useful info.
Will you make the upgrade tutorial?
@@ProaliusVideos I am planning to. There are a couple of paths for upgrading, so I want to understand what is needed. The easiest one is to create a new SD card, but the system needs to be running JetPack 6.0+ first. If you are planning on booting from the NVMe SSD, the SDK manager is easiest. However, both of these require a new installation. For the SD card, this isn't a problem (you just need a new SD card), but for the SSD it feels like there needs to be at least a backup plan. Thanks for asking, and thanks for watching!
Did you build anything with it?
It's been out a couple of years, I've built many things. Mostly small robots built on R/C cars.
nvidia should consider fpga jetson, lower power, more specific opencl cuda compute
Interesting idea. Thanks for watching!
can this be run with a mac? like macbook pro?
I'm not sure I fully understand the question. The Jetson Orin Nano Super Developer Kit is a stand alone Linux computer. If you want to work with it from another computer, you can SSH into it. If you want to prepare a SD cad for the Orin Nano to work from, you can use any computer with an appropriate SD card slot.
However, if you want to develop code on a host computer for the Jetson, you must use a PC running Linux. Most people use Ubuntu (which NVIDIA recommends) for their development machines. However you can also develop on the Orin Nano.
It is recommended for serious development (such as kernel compiling, device tree wrangling, and so on) that you have a PC available. Hope that answers your question, and thanks for watching!
@ Got it! thank you for this indepth explanation!
@@FarisandFarida You are welcome. Happy Holidays!
Exciting, and thanks for posting a video about it! I'm ordering one now. Curious to see how fast it will run YOLO models 🤖. EDIT: Oh just kidding. It's out of stock already 😿
You are welcome! Looks like it's going to be popular! Fortunately there are factories making them. Thanks for watching!
Can someone PLEASE give a "REAL WORLD" concrete example of what this thing can do for me? I'm brand new to this but am thinking about buying one. I google it and most answers are nonsense. Thank U!
ruclips.net/video/QHBr8hekCzg/видео.htmlsi=ryCfRgMwuN32kcOf
@@JFlogerzi Tk. be watching it now.
If you have to ask, this is probably not for you. You have to decide if this is right for you, no one can tell you for sure. Do more research on Machine Learning and how it is being used in almost every industry. Then, if you are still interested, then look for any other manufacturer that makes a Edge Computer in this price range and power - you won't find any. Good luck.
@JFlogerzi - I think this is an excellent video. I'm a little jealous :) Thanks for recommending it.
That's a good question. However it's a little broad to give a good answer to it. It depends on your interests. There are 10s of thousands deployed in professional applications. Many have to do with camera and vision. This ranges from quality control on fast moving production lines to identifying weeds while spraying on a tractor. There are medical appliciances. There are a great many deployed in robotic applications such as drones and autonomous vehicles. There are a large number of makers that use them in their projects. To get started, wander over to the Jetson AI Lab: www.jetson-ai-lab.com/community_articles.html and look through the projects that community members have built. Then go to the Tutorial page where there are dozens of examples to see if there is something that interests you. Thanks for watching!
Is this similar to a Raspberry Pi?
It's similar in some respects. The CPU cluster uses a similar ARM core (A76 vs A78 for the Jetson), but the Jetson has 2 more cores. Where they differ is that the Jetson has a 1024 CUDA capable GPU. The GPU is programmable, making it much more useful for running AI. Here's a good comparison between the two in an actual application: ruclips.net/video/QHBr8hekCzg/видео.htmlsi=9f8fimEE42h0dBcw
Thanks for watching!
Hello, I have a Jetson Nano Developer Kit and when I try to open it I get a black screen and I can't get to the terminal. When I try to format it via SDK Manager it stays at 99.80%.
It should boot to the desktop. You might need to look at the serial debug console to tell you what the issue might be. You should ask for help in the official NVIDIA Jetson Nano forum, where a large group of developers and NVIDIA engineers share their experience:forums.developer.nvidia.com/c/agx-autonomous-machines/jetson-embedded-systems/jetson-nano/76 Thanks for watching!
I solved the problem myself, thank you. I use VirtualBox, then I tried Vmware and it was solved. Please show this message to someone who has this problem
@@sporkline6272 I'm glad you were able to get it to work. Thanks for sharing!
OUT OF STOCK EVERYWHERE!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Looks like there were a lot of people interested in Orin Nanos, just as a less expensive price. It'll be interesting to see how the supply gets adjusted. Thanks for watching!
@@buffler1 It looked that way, but I placed a back order for 2 of them with Seeed Studios today. The reply email stated that it would be filled on January 1. Then later in the day, I got an email that they were shipping and will get them this Friday!!!! I then received a DHL email with a tracking number and that I would have to sign for the delivery on Friday. Santa must have intervened, I am sure of it 🧑🎄👍
Does it play crysis?
Good one! Thanks for watching!
so cool
Thanks for watching!
JetsonHacks is the doggo on the left
Could be. Thanks for watching!
Dude you' funny!
Thank you! And thanks for watching!
Try to buy one! OUT OF STOCK IMMEDIATELY< everywhere. Nuts.
Yes, there's big demand at this price point. The inventory should start building up soon. Thanks for watching!
Could you PLEASE try gaming on it?
not what its used for.
I don't game, so I'm not a good resource. Here's a link to ETA Prime running games on the Orin Nano: ruclips.net/video/nmZ6fhkFmDY/видео.htmlsi=84ZZ_nynRZMV-4OX&t=474 The performance should be better now with the speed bump and now that the software has had a little time to mature. Thanks for watching!
Just bought the Orina month ago. It doesnt even have GPU video encoding. Man I got scammed hard.
You should return it if you don't feel it meets your needs.
I like fast a cheap !!
Me too! Thanks for watching!
jetson nano does not recognize sd card
From your description, it sounds like there's an issue with recognizing the SD card. Maybe somethings wrong?
@@JetsonHacks yes, how can i fix it
@@MehmetKürt-v4m What model Jetson is it? How did you prepare the SD card? Is it an Orin Nano?
Just get a raspberry pi 5 for 50 bucks..
That's certainly one approach. I suggest that they might be different, superficially similar things. However, if a RPi 5 meets your needs, you should do that. Thanks for watching!
Reported for pricing misinformation.
Why do you say that? The new price is announced at 6am PST today, it may take some time to get through the entire supply chain. For example, at Arrow: www.arrow.com/en/products/945-13766-0000-000/nvidia?nvid=em-945-13766-0005-000 or Sparkfun: www.sparkfun.com/products/22098?nvid=em-945-13766-0005-000
She’s the kind of person people smile at but roll their eyes as soon as they turn their backs.
Don’t listen to these toxic comments.
Great vid btw!
@@JetsonHacks Every one of the sites is sold out and on backorder. Seeed, Sparkfun, and Arrow. I love the price, but a little misleading.
also, the Sparkfun site says that they can't guarantee the price when they restock - see their note " Note: If this item is available for backorder it is subject to price changes at any time; additionally, we are unable to guarantee time frame for shipping or availability."
Sparkfun states that they will not guarantee the price when they get inventory. See their note -
Note:We do not currently have an estimate of when this product will be back in stock.
Note: If this item is available for backorder it is subject to price changes at any time; additionally, we are unable to guarantee time frame for shipping or availability."