Missing caps? Conflicting pins?

So what obviously happens is that when the motors should turn on, the voltage drops very rapidly. This makes the MKR1010 reset and the motor boards never fully receives the correct signals. Or they get it, but since the board is reset, all pins go low. I can hear a very soft tick, but nothing happens. The voltage is immediately back.

Before I find a good solution for powering the MKR1010 separately, I'd likte to test with a cap, which might soften the voltage drop. You see, I have got the motors running and I see in the oscilloscope the drop and a curve going up back. But obviously at some point (when all four motors turn on) the drop is too big. Could a cap fix this, even though it is not the best solution? What kind and size of cap in that case?