Based on the photo thay you posted your board seems to have an unmarked CH340 usb-to-serial converter (the IC closest to the USB port).
You can roll back tthe driver using the instructions in UNO stopped working after having worked for years - #24 by ptillisch and see if it solves the com-state problem.
yes ,i also have many problems with this, the windows update driver in background without did nothing, many of my clients did same think , yesturday working and today not, and problem was 100% update driver that dont know old ch340/341
Your system / device manager sees the hardware.
Is the separate power supply wired correctly? I can imagine the wrong voltage on the USB port to interfere with communications.
The 328A chip does not seem to be a genuine Atmel part either. Does not necessarily mean that it won't work.
Good. Nothing should be connected to the Arduino for now except the USB cable.
Ok, so it is at least detecting that a hardware device has been connected.
Yes, I do think that we need to know this ^^^ next.
I would also ask, if it is creating a COM port, then what is the description of this device in Device Manager so we know what its identifying itself as?
The newer driver doesn't work for the counterfeit CH340
To expand on what @Deva_Rishi said, it isn't necessarily a matter of age. We have seen many newcomers report encountering the "cannot set com-state for ..." error when uploading to sketchy cheap derivative boards with the unlabeled SOIC-16 ICs that identify as WCH CH340. Most of those users had surely just purchased the board. Of course, that doesn't tell us the age of the chip for certain (and the lack of labeling means we don't have a date code).
Conversely, during the time when I was only a hobbyist circa 2014-2018, I liked to gamble on buying the cheapest of the cheap parts from random Chinese sellers on eBay and AliExpress (I have since kicked that habit). This means I have a large collection of old CH340 chips in my parts stash, purchased from just the type of seller who have been sending out the boards with the bad chip. I have searched through my entire collection and I don't have a single unlabeled chip. All of mine can be uploaded to without any problems with the latest driver. This leads me to the conclusion that the bad chips must not have been on the market prior to 2018. Ironically, the fact I was lucky enough to avoid getting any of the bad chips has actually been unfortunate for me, because being able to reproduce the fault personally would allow me to more effectively support the affected users.
It does seem that the rate of reports of this error have dropped significantly over the last months. Maybe that means a limited stock of the bad chips has been exhausted, or perhaps only that the driver rollback workaround is now much more widely known (e.g., significantly present in LLM training datasets).
I haven't but UNO boards are not the boards i buy a lot.
I have not had to do this for a CH340 yet, but have drawn the conclusion from prolific PL2303 counterfeit chips.
Prolific provided the driver, but added a chip id test to their newer driver to prevent it from working with counterfeit chips that were manufactured by others. On W8.1 only the installation of drivers was automatically the latest, but with W10 updating drivers automatically it became more of an issue.
Remember FTDI gate?
I just remember fiddling with those cheap uploading tools until i started using my real (genuine) UNO and all my issues were solved after that.
The fact I was affected proves that the people I was buying from at that time were definitely not above selling shady chips.
I actually still have an adapter with a "bricked" counterfeit FT232R chip stashed away in my parts hoard. They are easy enough to recover, but I like to keep it that way as a souvenir, and in case the need ever arises to reproduce the conditions that occur when trying to use a chip in that state (unlikely in 2026, but it did make sense in the immediate aftermath of FTDI gate).
it is not bricked, i kept the older driver that can be re-installed, i never use those again, but haven't thrown them out.
I have thought of the 'if i die now scenario' , and who will go through all of those parts i have and what will happen to them. No one ever asks, more like, well if you die before me you can leave me that guitar in particular !
I think you missed the point. The updated FTDI driver actually bricked the FTDI chip.
FTDI later changed the behaviour and the chip started spitting out a text in the line of "this is a counterfeit chip" in between the expected output.
The "brick" (I should have written "soft brick" since indeed it is easily restored using FTDI's own FT_PROG tool) was accomplished by a release of the official FTDI windows driver changing the USB PID of the counterfeit chips from 0x6001 to 0x0000. This caused the USB device to no longer be recognized by the driver.
Oh no, oh i got mine later, that driver wasn't ever installed on my laptop. The counterfeit chips never worked anymore with the more recent driver. Mine never got bricked though.
This whole part of the fdti gate was before my time.
The counterfeit chips that have the correct 0x6001 PID work perfectly with the latest driver versions. I have several of the adapters with these chips and use them all the time.
I never actually experienced the "FTDI gate v2" problem @sterretje mentioned:
I believe it was only one specific version of the driver and after that they went back to normal support without any any-counterfeit measures in order to avoid further negative community relations.
Like those of us who didn't buy anything with a FTDI chip "just in case" since there was no easy way for the average user to tell.
I went to things with CH340 interfaces. I figured it was so cheap nobody would bother to copy it. Looks like lately that is no longer true. 
the 16U2 versions work like a charm.
Yes I have been bitten by this as well. Several years ago I purchased a relatvely expensive adapter for my Astronomy kit which contained a PL2303 chip. One day, it suddenly stopped working and I never got it working again on Windows, although it still works on Linux. I was later to discover that apparently Prolific had done something similar to FTDI. It was possible, with some difficulty, to roll back to the previous driver but Windows Update would inevitably update it again sometime later. The UK supplier had evidently used USB-to-serial breakout boards in their product that had cloned chips. Of course, by the time the driver update had happened, I had no recourse back to the vendor. I still have the board in a box somewhere. Since the product came from a known astronomy supplier and was quite expensive, I had no reason to suppose and certainly much less to expect that it would contain a fake chip. Perhaps the vendor themselves was not aware that the chips were fake? Although I expect that at some point they would have discovered the problem? Their website is still up and running although it doesn't appear to have been updated in years.
On a more recent note, I had one of those NANO boards with the unbranded UART chip that pretends to be an FTDI part fail on me. Upon further investigation, the ID number was just a line of zeroes. I have no idea whether it came like that or whether something caused it to loose its ID, but the FTDI tool for restoring the chip ID would not restore it.
Prolific has trouble with their USB drivers. I had to use this driver for years. EVERY time Windows updates, the Prolific driver was disabled. Perhaps Prolific does not have/purchase the correct certificates? Anyway, I had to install the OLDEST driver to make the driver work because installing the newest failed... I suspect the oldest driver had the "nuget" that was once allowed, but now is not, and all the newer drivers' "nuget" is blocked from being installed.
I also had to change the Prolific driver settings to "run as admin" and be "windows 7" compatible. (the device was an e-bike USB CANBUS programmer)