Henrikas Simkus 

If you are buying this for work and Trados is a key software for this purpose then I would recommend you don't buy a MAC at all.  You are asking for problems because you will need to be able to run windows inside a VM on your MAC which adds a layer of complexity that you will be responsible for.

Your old Lenovo is probably mid-range by today's standards - very suitable for office work, translation, light development, and day-to-day productivity.  It may show its age with more demanding or future-facing software (like heavy AI tools, some .NET workloads, or virtualisation at scale), but it’s not obsolete.  If it's still running smoothly, there’s no urgency to replace it unless you’re pushing into high-performance territory.  Buying a MAC will not really help you much if you are expecting significant gains when working with Trados Studio.

I'd advise you to stick with it for a while longer, or buy another Windows machine... unless you really have pressing needs for the sort of benefits a MAC can bring and you are comfortable dealing with the requirements of running Windows within a VM.

But worth listening to any advice that people like Jerzy Czopik and Walter Blaser can offer as they are both knowledgeable when it comes to hardware and one uses Windows, the other a MAC :-)

Henrikas Simkus 

Dear Paul , thank you for the mention. I agree with you, that buying a Mac for Trados Studio is not optimal. While I know quite a number of people uses Mac with a VM running Windows for operating Trados, but I wouldn't do for me. Mac is a beautiful machine, but when you run Windows on it, I fail to see those big advantages. For me it is like buying a Ferrari to move your household. My biggest problem with Mac would be the completely different keyboard. But this is just my opinion. I would register myself as business customer with Lenovo and wait for their newsletter. Recently there was an offer of 40% off on a ThinkPad Carbon X1... I would strongly prefer this machine over any Mac—and the price was about a Macbook Air.

1. Thank you very much for your answers! Now I'm thinking, if I stay with a Windows PC: what are the most important settings for "speeding-up" TRADOS? RAMs? Processor?
2. And another question, what if I have about 40+ translation memories and I can't combine them into one because I have to split them by client. When I use all the memories to a project, and I click on "pre-translate" or "analyse", I can just go to lunch until everything is done with all the translation memories. :) it takes so loooong.. TM's are really filled up over 15 years. I am just thinking, what if they were all in one file, in one translation memory, would it be faster? But then that translation memory would be about 10 GB ( I guess so).
3. And another question about the Mac OS - isn't Trados planning to release a version for MAC in the future? Thank you again for your answers.

1. Thank you very much for your answers! Now I'm thinking, if I stay with a Windows PC: what are the most important settings for "speeding-up" TRADOS? RAMs? Processor?
2. And another question, what if I have about 40+ translation memories and I can't combine them into one because I have to split them by client. When I use all the memories to a project, and I click on "pre-translate" or "analyse", I can just go to lunch until everything is done with all the translation memories. :) it takes so loooong.. TM's are really filled up over 15 years. I am just thinking, what if they were all in one file, in one translation memory, would it be faster? But then that translation memory would be about 10 GB ( I guess so).
3. And another question about the Mac OS - isn't Trados planning to release a version for MAC in the future? Thank you again for your answers.

Henrikas Simkus 

Speed in Trados depends mainly from the file system speed. So to have TS working smooth, store all files on the fastest drive you have. The next important part is the overall performance of the PC. I have a comparison between desktop i7 14700 with 64 GB RAM and m2. SSD and a laptop with low voltage i7 12th generation. The speed difference is not that very huge. Of course laptop is slower, but still workable.

Using so many TMs at a time will never be productive, this is simply too much. Consider using only those TMs, which contain similar topics. And then consider using project TMs, which are simply much smaller and thus faster.

Specs for a fast PC would be from my todays point of view: an i7 of 13th generation or younger, 16 GB of RAM and a fast SSD M2. No files on drive C:\, no use of live synchronisation with any cloud. Locally stored files work best. In my case all projects are stored on a NAS. Write speed around 950 mbps, read 1900 mbps. This is enough for even huge projects with 50,000 words+ in several files. But I do not have 40 TMs there, maybe 10 as maximum. I always generate project TMs per project, so the speed is ok.

Henrikas Simkus 

Henrikas Simkus said:

what are the most important settings for "speeding-up" TRADOS? RAMs? Processor?

Jerzy Czopik is more knowledgeable than I am here, but I would say this:

RAM: Whilst Trados Studio is currently a 32-bit application I'd plan on the basis it will eventually become 64-bit.  A 64-bit Trados would be able to use far more memory than the current ~4 GB limit. This makes 32 GB (or even 64 GB for power users) a smart investment, especially for large projects, multiple TMs, or concurrent workflows.

CPU: With 64-bit support, Trados could potentially leverage multi-threading more effectively. That means a CPU with more cores and higher base speeds (e.g. Ryzen 9 or Intel i9) would offer better scalability for future performance gains.

SSD: Critical in my opinion. A fast NVMe SSD would future-proof even better than a basic SATA SSD.

GPU: Largely irrelevant unless future versions introduce GPU-accelerated AI features or rendering-heavy previews (unlikely in the near term, but not impossible given trends in CAT tool AI features).

Another thing to consider would be how you use the software itself and organising your resources is important, as well as file preparation, and also minimise the number of plugins you have installed at any one time. Plugins take memory so offloading them when you don't need them and adding them back when you do (integrated appstore makes this simple) is a smarter approach to loading as many as you can even when you don't use them regularly.

Henrikas Simkus said:

I have about 40+ translation memories and I can't combine them into one because I have to split them by client.

Why not?  Can't you use Fields and attributes so that the TUs are linked to each client?  Then you only need one TM.  This may be helpful... it's around 12 years old but still relevant for this point:

https://multifarious.filkin.com/2013/07/12/fields_attributes/

Henrikas Simkus said:

But then that translation memory would be about 10 GB ( I guess so).

Seems unlikely... but then I don't know what size TMs you're working with in terms of the number of TUs, and I also don't know what sort of content you translate and how much taggy content is added to yourTMs.

Henrikas Simkus said:

And another question about the Mac OS - isn't Trados planning to release a version for MAC in the future?

Unlikely.

Paul i9 in a laptop is beyond good and bad, as we says in Germany. Simply far too expensive. Even in a desktop it is expensive. So I would indeed prefer the newest i7 over i9, but invest in a good brand. I use Dell and am really satisfied with the performance. When buying a laptop, it is crucial to buy as much memory as you can pay. But when buying a desktop I always take the smallest amount offered and upgrade it on my own. Costs much less. For example my current Dell desktop was delivered with 16 GB of RAM. I bought 64 GB on Amazon for more or less half of the upgrade price by Dell. I also installed a second m2 SSD on the mainboard. All this small improvements are not that expensive when you have a desktop.

As for the memory size: my biggest one has been built over 20 years and is over 1,5 million segments. The size on the HDD is 3.82 GB. The newer one, built over the last 5 years, has roughly 100,000 segments and is as tiny as 250 MB.

Using multiple TMs separated by customer is the easiest way to maintain them, much easier than fields. But I doubt all 40 TMs belong to the same field, so I would simply group them by themes. Using 10 TMs and 5 or even more TBs per projects works with no big hassle.