I want to start by saying that I don’t think a laptop is the best workstation for machine learning. You probably already know this, but a desktop can offer a lot more power, and so can online compute power. A laptop can offer portability though, which is what a lot of people are looking for as they want to use the laptop for other things than just a raw ML work machine.
Despite not offering the highest amount of raw power, there are many gaming laptops that are suited for machine learning. In this guide we will cover what I think are among the best options so that you have a good basis for your research.
You should look for the following specs when doing your research.
|CPU||10th gen Intel i7 with 6 cores||10th gen Intel i7 with 8 cores or better|
|GPU||RTX Series||RTX 2080 Series|
|RAM||32 GB or more||64 GB|
|SSD||1 TB||2 TB|
The specs above are pretty steep, and not all laptops will meet these guidelines. If you are a machine learning beginner then it is okay to go with something on the weaker side. But also, if you don’t have the budget for some of these machines then consider trying out an online platform (AWS, Google Cloud) also as there you pay only for what you use when you use it.
Below are three of the laptops that I think are some of the best options.
Best Laptops for Machine Learning
There are a lot of laptops out there, and not all of them are well suited for machine learning. We have gathered the following list of machine learning laptops that do a good job at meeting the recommended system specs.
LamdaLabs provided this laptop for review. There are not a lot of laptops that are designed specifically for machine learning. The TensorBook from LambdaLabs is one of them, and it shows that it is built for resource intensive work. This TensorBook was provided by LamdaLabs for review.
The laptop is a great option for professionals, researchers, as well as students who are working on their thesis. LambdaLabs has put great care in designing a laptop that is able to run most of the popular AI models.
The laptop comes with the Intel Core i7-10875H CPU. The CPU has 8 cores which is as good as it gets currently on laptops. You will find that a lot of laptops that have a new 10th gen CPU only have 6 cores, and I recommend that you get as many cores as you can.
The laptop also has the Nvidia GeForce 2080 Super Max-Q. This GPU is one of the better GPUs you will find in a laptop, and the 8 GB of GDDR6 VRAM is what allows you to use this laptop to run most of the AI models.
Another great feature of the laptop is that it comes with 64 GB of RAM. Most other laptops will only have 16 GB, perhaps up to 32 GB. Although RAM is quite easy to install yourself, you sometimes have to break a warrant sticker to do so.
There are two 1 TB SSDs on this laptop. I usually recommend that people buy an extra external storage device, but with a total of 2 TB SSD space you should be okay.
If you are looking for a machine learning laptop, one that is provided by a company that specializes in machine learning / deep learning hardware, then the TensorBook is a great option.
See the TensorBook on LamdaLabs’ website.
MSI GS66 Stealth
Another great laptop for machine learning is the MSI GS66 Stealth 10SFS-032. MSI is a good company with a lot of experience in making high end gaming laptop. This laptop also looks really good, with the thin bezels adding to the already premium look.
One of the main selling points of this laptop is that it comes with the new Intel Core i9-10980HK CPU. Not a lot of laptops come with this CPU, even in the premium laptop space. The CPU has 8 cores, which is something you want to look for when evaluating the CPU. The more cores you have the better, and this is one of the better CPUs for artificial intelligence work in a laptop.
There are a few configuration options for this laptop, and the model that we picked for this guide is the one that has the RTX 2080 Super Max-Q. You want to get the best GPU you can as it is very important for machine learning. If you are on a budget then downgrading to one of the other RTX options on the product page should be ok for many, but you should stick with this option if possible.
The laptop comes with 32 GB of RAM out of the box. This is not as much as the 64 GB that is in the TensorBook, but it is still pretty good. You can upgrade the amount of RAM yourself, up to a max of 64 GB, but 2/2 slots are already in use.
There is 1 TB of SSD storage which is okay. You might want to get some external storage with this one so that you have enough to keep all your data.
Overall the MSI GS66 is a solid choice as a machine learning laptop.
See the MSI GS66 Stealth on Amazon.
ASUS ROG Strix Scar 15
The next machine learning laptop is the ASUS ROG Strix Scar 15. This version of the laptop offers specs that are suitable for deep learning. The laptop also looks good which is not always the case with so called gaming laptops.
As with the MSI GS66, this laptop comes with the Intel Core i9-10980HK. This 8 core CPU is a great for machine learning on a laptop. The CPU is also important for machine learning, so you want to make sure you get one with as many cores as possible, where 8 is a very good option currently.
The GPU in this model is the RTX 2070 Super. It is not as good as the RTX 2080 option, but it still has 8 GB of GDDR6 RAM which is important. The GPU plays a big role in artificial intelligence, and it is one of the limiting factors in laptops, as desktops can have better GPUs.
There is 32 GB of RAM in the laptop and 2 TB of SSD storage. Both the RAM and SSD are quite decent, and this should be okay for most people.
See the ASUS ROG Strix Scar on Amazon.
MSI GE75 Raider
The MSI GE75 Raider is one of the few 17 inch laptops on this list. Some people may find the extra screen real-estate nice. The screen is also a 300Hz full HD option, meaning is is very well suited for gaming. Coupled with the great parts that this laptop has, this makes it one of the best gaming options on this list.
This laptop is another one that comes with the Intel Core i9-10980HK CPU. The 8 cores that it offers really is important for Machine Learning, and it seems that the top gaming laptops are also taking advantage of the power that this newer CPU can offer. The most common option out there is going to be a high end i7 model.
The laptop also comes with the Nvidia GeForce RTX 2080 Super Max-Q which is a great GPU for machine learning on a laptop. The Max-Q variant is not as powerful as the desktop class version, but it still does a good job in a laptop offering 8 GB of GDDR6.
As with many of the other options, the laptop has 32 GB of RAM and a 1 TB SSD. The SSD could be a bit bigger, but other than that this is a solid machine learning laptop.
See the MSI GE75 Raider on Amazon.
Acer Predator Helios 300
The Acer Predator Helios 300 is the first laptop on this list that is more budget friendly. Despite that, it still offers amazing parts that could be used for artificial intelligence. The laptop looks like a gaming laptop that some might find bad, but overall this laptop offer a lot of bang for your buck.
The Acer Predator Helios 300 is one of my favorite mid-level laptops. It comes with a new 10th generation Intel Core i7-10750H. This is a 6 core CPU which is not as good as 8 cores, but still it is a decent option, especially when you consider the price of the laptop. This CPU is also suitable for playing video games at pretty decent settings.
The laptop also comes with the RTX 2060 GPU. This is a great GPU, but the 6 GB of GDDR6 VRAM is not as good as the 8 that some of the other RTX models offer. This GPU should be good enough for most people starting out with machine learning.
There is 16 GB of RAM and a 512 GB SSD. This does not meet our minimum recommendations, but these are both things you can upgrade yourself in the future if you find that you need more when you get there. The CPU and GPU are not as easy to upgrade in a laptop so they are more set in stone.
Overall this is a great laptop that will be good enough for most people starting out.
See the Acer Predator Helios 300 on Amazon.
MSI GL65 Leopard
The MSI GL65 Leopard is another more budget friendly option. It is not cheap by any means, but it does not carry the premium price that that some of the best laptops on this list do. This is a great laptop that also looks really good. It is quite similar to the Helios 300 in terms of specs, with the main impactful difference being the GPU.
The Leopard comes with the Intel Core i7-10750H CPU. This is a good CPU for a high end laptop, and works well for gaming as other resource intensive work. It is a 6 core CPU so it will not be as good for ML as the 8 cores, but the 8 core CPUs mostly only come in ultra premium laptops. This CPU will be good enough for most people.
The GPU in this laptop is the RTX 2070, which is a great option for a laptop. This is one of the better GPUs on the list, making it a worthy option for machine learning with 8 GB of VRAM.
There is 16 GB of RAM and a 512 GB SSD on board, both which could be better. Again, these are parts you could upgrade yourself at some point down the line so it is not so bad.
See the MSI GL65 Leopard on Amazon.
Razer Blade 15
The Razer Blade 15 is another great laptop. This one is not as expensive as some of the more premium options, but still not cheap. The parts are very good, and the laptop looks amazing.
This is another laptop that comes with the Intel Core i7-10750H. It is the 6 core CPU which will work well for many. It is also perfectly suitable for playing video games if you like to relax with a game or two every now and then.
The GPU is the Nvidia GeForce RTX 2070 Max-Q. This is not as powerful as some of the other options, but it still offers the 8 GB of GDDR6 VRAM.
There is 16 GB of RAM and 512 GB of SSD. These are both under the minimum recommended specs, but once again you can upgrade these later if you feel like you need more.
The Razer Blade has been around for a long time and it is a great laptop. The build quality is good, and it is made by a reputable manufacturer.
See the Razer Blade 15 on Amazon.
Razer Blade Pro 17
There is a second Razer laptop on this list and that is the Razer Blade Pro 17. This is a 17 inch model and comes with better parts than the 15 inch one. If you want a larger screen then this is a great option.
This version of the Razer Blade comes with the Intel Core i7-10875H. This is an 8 core CPU and is the same that comes in the TensorBook. This is a great CPU and works well for machine learning on a laptop.
The GPU in this version of the Razer Blade is also better. It comes with the RTX 2080 Super Max-Q. This is also the same that comes in the TensorBook. You would want to get a model of the RTX 2080 GPU for best performance, or at least make sure you get at least 8 GB of GDDR6 VRAM.
There is 16 GB of RAM and a 512 GB SSD, both of which are the biggest downsides with this laptop. As they are not so difficult to upgrade it is not so bad, but for the price it should have 32 GB and a 1 TB SSD.
See the Razer Blade 17 Pro on Amazon.
The only HP laptop on this list is the HP Omen. It is a gaming laptop and it looks like one as well. I am not a fan of how the laptop looks, but it has pretty good parts and is reasonably priced for what it offers.
The laptop comes with the Intel Core i7-10750H. We have seen this CPU many times now and it is clear that it is common in mid-level and high end gaming laptops. There are better options available in the ultra premium space, but for most this 6 core CPU will be good enough.
The GPU in the laptop is the RTX 2070 Super Max-Q. This is a decent GPU and works well for most things that you will be doing, at least in the beginning. A desktop class GPU would be better, and the RTX 2080 version is also better. That being said, this is a solid choice, especially for the price of the laptop.
There is 16 GB of RAM and a 512 GB SSD which is a little on the low side.
See the HP Omen 15 on Amazon.
MSI GF65 Thin
The MSI GF65 Thin is an option for those who are on a budget. There are some big drawbacks with this laptop, in particular with the amount of RAM and the CPU. The RAM you can fix, but the CPU is set.
The laptop has the 9th generation Intel Core i7-9750H. This is the previous model of the 10750H that a lot of laptops on this list has. It was a staple in high end gaming laptops for a long time but has now fallen out of favor in exchange for the new 10th gen. It is still a good CPU and it is a way to bring down the cost of the laptop.
The GPU is the RTX 2060. This is a decent GPU option, but it only has 6 GB of GDDR6 VRAM compared to the 8 GB of the 2070 and 2080.
There is also only 8 GB of RAM, which can be fixed quite easily as only 1 RAM chip is in use. It can take up to 64 GB of RAM which is quite good. There is a 512 GB SSD which should also be supplemented by some external storage.
See the MSI GF65 Thin on Amazon.
ASUS ROG Zephyrus S17
The last laptop on this list is another 17 inch option. This laptop has premium parts and is well suited for machine learning on a laptop. It rivals some of the better options in this list in terms of parts. The only drawback is the funny body which will put off some people.
The laptop comes with the Intel Core i7-10875H CPU. There are a couple of other laptops on this list that have the same CPU. It has 8 cores and works well for machine learning on a laptop.
The laptop comes with the Nvidia GeForce RTX 2080 Super, which is a very good option. This is the same as some of the other laptops on this list that are best suited for ML.
There is also 32 GB of RAM and a 1 TB SSD on board. This is more in line with what we are looking for than the 16 GB/512GB combo we have seen a lot of.
See the ASUS ROG Zephyrus S17 on Amazon.
What to look for in a Machine Learning Laptop
It is important that you do your due diligence when looking for a laptop. A lot of the high-end laptop do fairly similar in terms of performance, so you should also check out their product pages to see if there are any other features of the laptop that might interest you.
When looking for the best laptop for machine learning you should take the following into consideration.
What CPU to get
The CPU is the brain of the computer, and it is quite important for machine learning. The best machine learning laptop will have a new generation of the Intel Core i7. I do not recommend going with anything lower than an i7 for artificial intelligence related tasks. Also, you probably want a CPU with as many cores as you can get. The more cores a CPU has, the more it can do simultaneously. Ryzen is coming out with a lot of great CPUs these days as well so a high end Ryzen CPU would work as well.
As far as the generation of the CPU goes you should get a 10th generation one. There will always be better parts just around the corner, but you could consider renting some online resources until there are more options with a newer CPU. That being said, we will have laptops with GPUs that will do much of the legwork for us in terms of computation. As the new generation comes out, the cost of the 9th generation should be coming down slightly as well.
What GPU should you get
The best laptop for machine learning will definitely have a powerful GPU. A GPU is great for parallel computations, something that is incredibly useful in machine learning. Machine learning requires a GPU to perform well.
As far as which GPU you should get, you should get the best one that fits in your budget. The GPU is one of the most important parts of a laptop for doing machine learning. Most gaming laptops already come with GPUs which is great. For a laptop you should be looking for a high end Nvidia GPU. The RTX 2080 is a good option.
The most important part is that the laptop you get actually has a GPU.
How much RAM do you need
RAM is another important component for machine learning. You need not only a large amount of RAM, but also the fastest kind of RAM you can get your hands on. RAM is important for holding information readily available in memory. In machine learning you often have to load a lot of data into memory at once in order to do calculations on it. The more times you have to go to secondary memory (ssd / hdd) to retrieve data, the slower it will be.
The best laptop for machine learning will have at least 32 GB of DDR4 RAM. The good thing is that most laptops can handle 32 GB, so even if you get one that only has 16 GB you can buy more RAM and upgrade it yourself. Most laptops have two slots for RAM chips, and in a 16 GB model they usually have two 8 GB chips. The important thing to keep in mind is that you can not upgrade from DDR3 to DDR4 RAM. The chips have different sizes and will not fit in each others slot.
Is Mackbook good for machine learning?
I don’t recommend Macbooks for machine learning because you need a GPU. If you want to use online compute power then it does not matter which laptop you have, and in that case a Macbook is fine.
How much storage (SSD) do you need
Many data sets for machine learning can reach hundreds of gigabytes. Even if you only keep them on the computer for a limited time while you work on your algorithms, they can quickly fill up your entire disk. Because of this, I recommend that you get at least a 1 TB hard drive. An SSD is better than hard drive, and many laptops will have the SSD option instead of the hard drive. Also, some laptops have a combination of a hard drive and an SSD, which is a nice compromise between size and price. The smallest storage device on your laptop should not be less than 256 GB, and the total should not be less than 1 TB. You might also want to pick up an external hard drive that you can use to store your data sets long term. It is not feasible to keep that many on a laptop as they will fill it up very quickly.
You can often upgrade the SSD in a laptop yourself. If you want to go for a laptop that does not have a lot of storage, do some research into if you can upgrade it easily or not.
Alternatives to a Laptop for Machine Learning
You should not get a laptop if you want the absolute best performance.
The following are alternatives to a machine learning laptop that you should consider.
Desktop for Machine Learning
As laptops are not the best workstation for machine learning, you should consider the alternatives. The first one is a desktop computer. A desktop can be more powerful and can give you more bang for your buck if you build it yourself. The big drawback of course is that you can’t carry it around like a laptop. If your situation allows for it, a desktop is a better alternative than a laptop for machine learning. If you are looking for something that you can bring back and forth to places, then a laptop might be the way to go, but you still have one more alternative if that is the case.
Online Platforms for Machine Learning
Renting online compute power is not very expensive these days, at least not in the short term. If you are using it a lot then the bills will stack up and buying either a laptop or desktop would be cheaper. You have both Amazon AWS and Google offering machines that you can connect to remotely do to your work. You can probably get more powerful machines there than what you would be able to acquire yourself. A big advantage is that you can leave the machine running without interrupting anything. If you use a laptop or a desktop it will kinda be locked up for the duration of your learning if you want to give it all the resources.
Check out the two options below if you are interested in trying an online platform instead of a laptop:
Final Thoughts on Machine Learning Laptops
You should be able to find the best laptop for machine learning if you follow the pointers from this guide. With the advent of gaming laptops there are powerful alternatives that should be able to handle much of what you can throw at it. You should keep the pointers from this guide in mind when you do your research. If you are on a budget then you are better off renting space online, as low-end laptops are not well suited for machine learning. That being said, you can still use them.
You should also consider the option of building a desktop machine, as they can be more powerful and more up-gradable in the future. A laptop does offer a lot of advantages as well, but the main advantage is that it is portable. If you are a student then a laptop is a good option as you can bring it to campus as well.
We hope you found this resource helpful, and that it helps you with your research into what is the best option for you. The best laptop for machine learning should have a great CPU, great GPU, and a lot of RAM. You also need a lot of storage.
I am a Software Engineer who enjoy writing about things that interest me. Here on OamaTech I write about programming as well as other tech related topics. Visit my author page to learn more about me and what I am working on right now.