The difference between an exploit and a strategy is greatly different if you are talking about AI vs AI, or Human vs AI, or Human vs Human. At this time with the mechanics and AI programming, the AI doesnt react well or behave well, with the tremendous amount of information offered by the huge visibility offered by sensor boats. So, in the 3 case mentioned above:
AI vs AI: It wont build such large boats in the first place, and would certainly confuse current AI programming, making it focus less on more close proximity threats/issues
Human vs Human: As long has both humans both understand the game mechanics involved, I dont see this as an exploit, as both sides can do it, and recognize the importance and need.
AI vs Human: definite exploit. A human will take advantage of the information available and use it to their benefit unimaginably better than the AI would.
Let's look at historical uses of intelligence gathering. No denial that economic espionage gives corporations and business persons a tremendous advantage, that's why there are laws against it. Research espionage the same thing, and again why there are laws to prevent it. Nevertheless, corporations go at great length to prevent espionage. But as there are no laws in love or war, we can prove the impact of military espionage throughout the wars in our history. If it wasn't for the intelligence gathering efforts by the allies which were the highest contributors to the victory of the most pinnacle battles of WW2 (North Atlantic, Kursk, Midway), we would be living in a far different world today.
This is why I do like the dev's approach to the sensor solution, giving them a diminishing return. Small amounts of intelligence should be low cost. Medium amounts, at a reasonable to high cost. Godly amounts of intelligence should not be allowed, or be of a cost that is ridiculously high...which as of today, it is not.
Even once they nerf sensors properly, it would be interesting to have a defense against it. Sensor Jamming, which could be placed on planets. This would make things interesting indeed. Of course, if the AI is not coded well to handle this, then it immediately turns into a Human vs AI exploit for the human.