I'd like to add a slightly different take on this, based on recent testing, and experiments done in the past. Both mesh and airflow tiles behave thermally like debris. They interact with the tile BELOW them, and exchange heat with that tile. So they are not perfect insulators with things below them. It is true that they do not exchange heat with tiles to the side or above, but they do exchange heat with tiles below. This is like debris, not buildings.  The exchange rate is the same rate as the same mass of mined matter sitting as debris. This is true in vacuum, gas, and liquid environments. Heat transfer to surrounding medium, and to the tile below, matches the heat transfer of debris. To me, this suggests the code treats the 100kg metal from the mesh/airflow tile as if it were debris, for heat calculations. It's subtly different than buildings, which can help with planning builds later on. Consider the images below.  In the first image, in a vacuum, the debris and mesh tile transfer heat at exactly the same rate to the tile below. The solid tile of metal transfers heat much faster, while the mesh tile maintains perfect temp.  In the second image, hydrogen gas enters the picture. The two setups transfer heat to the tile below at almost the same rate. I say almost because my initial temps of two two metals were 1738.9K and 1739K. Some very interesting turbulent up/down changes of gas temp on both sides that cause the two sides to cycle between which transfers heat faster. The cyclic behavior suggests some numerical scheme is at play (high deltaT clamping because of the multiply by 25 gas/solid interacion). However, the average behavior suggested the two really do transfer heat at the same rate.  In the third image, the liquid water stabilized the turbulent temp changes, and the two pictures transfer heat at the exact same rate.  In summary, think of mesh and airflow tiles as debris for heat transfer computations. Whether this is intended behavior or not, that's another question.   One fun side effect of this fact is that you can rapidly alter the temperature of an insulated tile by building a mesh or airflow tile ABOVE it. I ran into this issue quite a bit with the old steam turbine, and for a while I wondered why some of my insulated ceramic tiles would stay at 20C, while others would reach >150C. The issue was always either (1) metal debris above the hot insulated tile or (2) a mesh or airflow tile above the hot insulated tile. This fact can be used to both heat and cool insulated tiles rapidly. Coupled with running  debris through an insulated tile on conveyors, you can watch insulated ceramic behave like regular obsidian (hyperbole).