The title basically is the question.
Well, I know why Apple added chip to their cables and changed the connector for good measure. That's of course not a corporate greed but a loving care about their customers... and making sure it would cost them a fortune to leave.
But nowhere in the USB specification it requires anything but the wire. Of course high speed introduces strict limits to electrical characteristics etc. But still... it is JUST A WIRE! Billions of cables in the world work just fine without anything else.
However when it comes to magnetic cables, which are basically regular cables with connector attached to wires by some pins and magnets, they all have chips inside. The end result is that it is practically impossible to find cable which works with all USB devices in home. Each cable comes with a list of incompatible devices almost as long as compatible. Fast charging that used to be enabled by simple pull-up in micro connector now requires "improved smart chip" and does not work with half of the mobile phones.
The question is - why? Is there any reason I don't see? Unlike Apple products these cables are dirt-cheap, and if there is no increased cost then there should not be a reason for manufacturers to make all this mess out of perfectly simple and extremely useful idea.
Regarding suggestions to tear down some cable - yes, tearing down and analyzing them with DSO may provide an answer to compatibility problems. But the whole point of posting a question here was the hope that there could be an electrical engineer who knows an answer.
Also, according to (very limited) marketing info, some manufacturers include identification chip into plug portion, designed to trick Apple devices into thinking they connected to authentic cable and avoid annoying pop-ups. Legal issues aside, this has nothing to do with the question, since those cables still have more chips in the connector portion. Funny thing though, is that most devices on "incompatible" lists are Android micro-B / type-C devices.
Anyway, here are the theories presented so far:
By @pjc50: To simulate the order of connection required by USB spec and guaranteed mechanically in regular cables. This is top contender at the moment, because it is applicable to all 3 types of cables on the market (non-reversible, electronically reversible and mechanically reversible).
By @Ali_Chen: To provide additional ESD protection highly important here due to exposed pins. This is another strong argument for having chips in the cable, also applicable to all types of cables.
By @dim: To avoid dangling pins in dual-row models, which might affect high-speed transmission. Not sure if this is critical for USB 2.0 speeds though.
By @dim: To reverse outputs in single-row models. Yes, this certainly is the case. Note that many cables reverse power only, leaving you without data. Also note, that simple reversing should not result in selective compatibility issues.
By @tom: To control the LED. Now, this is certainly a reason to have some circuitry inside. But if the price is loss of primary function with half of the devices then this is dubious reason, to say the least.
Is seems there is no one here with insider knowledge of what is really going on in those connectors.
So, I am willing to accept @pjc50 and @agent-l suggestions that those chips facilitate an order of connection required by USB specification, as something that is supposed to be inside from engineering point of view.
At the same time, considering how cheap those cables are, I'd like to point out @ali-chen suggestion that they are simple TVS as something that is actually there.
As a side note, very interesting idea came out of @ali-chen comments - that there are more reasons to have chips inside plugs than inside cables. I know that iPhone plugs have chips in them to simulate genuine Apple cable. But since those plugs expose USB interface pins to all the ESD around, having TVS in all plugs regardless of connector type strikes me as another must have.
Also for electrically-reversible cables the simplest solution to swap power lines would be to put diode bridge in the plug, instead of trying to somehow guess the insertion direction and manage power in the connector.
So, until someone tears down actual cable and tells us the truth I propose to consider this question answered.