So the question is where would the VHT-1 store the ammunition for it's 105mm cannon? The place with the most volume available would be the torso* (maybe even the nose/B-mode tailcoat), however due to the battloid mode option for sustained firing (even at reduced ROF) this will require a complex feed mechansim to navigate the shoulder and upper arm to the shield w/n the range of motion of any joints, where if it was a Gladiator Mode only weapon it could feed in from what would be the top of the sheild in B-Mode and we wouldn't see it (unless the B-Mode option is only possible due to a pre-chambered round and it won't reload in B-Mode in actual practice). Now such a feed mechanism might be doable to navigate all those joints, but I would think that it would work best with a smaller sized round (105mm is over 4 inches in diameter, and this ignores the length of the shell/round) and be subject to added complexity.
*Torso break down in Spoiler
The use of the legs or the alternate arm and/or arm shield is even more demanding for the feed mechanisim, and it has to contend with very limited volume given it also has to share the feed mechansim and payload with other systems that we know are present in these locations.
This only really leaves the Right Side Arm Shield (location of the cannon) and/or arm. I estimate the upper arm location (which connects to the shield) is approximatly 1/3 the length of the shield, and just as wide/tall as the shield (which makes sense given transformation). Now not all of that space is available, you will lose some of that to the limb actuators on the shoulder and elbow (plus any additional "muscles" that might be present). For simplicity let's assume that only 1/2 of the available space can be repurposed for ammunition storage though I think that might be generous (as such I split the results later).
I estimate the size of each Arm Shield to be 3meters long (it appears to be ~1/2 the length in transport mode) by at least 0.125m wide and
0.125m high (the bore of the 125mm Beam Cannon that is also housed here without apparent modification). Technically I am ignoring barrel thickness and the armor plating thickness so width and height are off. The armor thickness is of less concern IMHO as that is already pre-allocated space I don't have to factor in again later. The Barrel thickness is of more concern as it reduces the internal volume of that space, though it will increase the available space, plus space is lost as I am ignoring the possibility of gaps between the barrel and inner wall of the shield compartment. It also holds a barrel that is 1.75m long, as determined by the difference in length for the T and G modes of the VHT (the sole reason for the difference in length), this doesn't include any loading mechanisim or firing chamber (which logically has to be big enough for a round).
Based on these numbers I am going to use two different approaches to determine if the required quantity can be made to fit: available volume and available length.
By Available Volume. If we treat the 105mm barrel and the arm shield as both simplified rectangular cubes (which they are not), how many 105mm diameter rounds can you fit in the remaining volume with no consideration for how they are stacked, the number in parathesis is for the upper arm (all fractional numbers rounded down):
-NATO standard length (617mm) can hold x5 (+1 in UA)
-at 1/2 the NATO length can hold x10 (+2 in UA)
-at 1/3 the NATO length can hold x15 (+4 in UA)
-as 105mm long cylinder can hold x30 (+8 in UA)
-as 82.95mm long cylinder can hold x38 (+10 in UA)
-as 66.15mm long cylinder can hold x48 (+13 in UA)
-as sphere it can hold x45 (+12 in UA)
By Available Length. If we subtract out the barrel length from the arm shield, and stack the shells infront of each other in the remaining length, number in parthesis is for upper arm (all numbers rounded down):
-NATO standard length (617mm) can hold x2 (+0 in UA, yes none fit)
-at 1/2 the NATO length can hold x4 (+1 in UA)
-at 1/3 the NATO length can hold x6 (+2 in UA)
-as 105mm long cylinder (or sphere) can hold x11 (+4 in UA)
-as 82.95mm long cylinder can hold x15 (+6 in UA)
-as 66.15mm long cylinder can hold x18 (+7 in UA)
-at 26mm long cylinder can hold all x48 (+19 in UA, @20mm long can hold all x60 from RT.com Infopedia in the armshield alone)
So can we put 48x 105mm diameter shells into the Right Arm Shield/Upper-Arm? The answer is yes, provided the shells are short enough in length. That means one of four scenerios is at play (five if you count it doesn't exist and they should all be beam weapons, six if you want to say an invisible/unseen external magazine drum is at play or seven they have undisclosed "macguffin" technology that can take several forms or eight the ammo is stored in the torso):
-the ammunition is telescopic/caseless (this can reduce the length of the round, but it is highly questionable if it could reduce a cannon shell that much)
-the ammunition is using more powerful propellant materials (like seen in SLAP rounds?)
-the ammunition is some combination of the caseless and more powerful propellant
-the ammunition requires no propellant as the weapon is actually an electro-magnetic mass driver (rail gun) firing a canister that can be configured for a variety of roles.
All of these have some basis in the 2E REPG setting. We know the ASC has caseless ammunition (M-33 pistol), they use SLAP and LEAP projectiles (more energetic propellant?, this also includes the M-33), and they have electro-magnetic mass drivers (Tri-star's 220mm Rail Cannons).
So I guess the question is which seems the more plausible explanation to to how the VHT-1 carries all that ammo:
-rounds apply caseless ammunition and/or SLAP/LEAP technology
-electro-magnetic propulsion for the shells (allowing them to be reduced in size)
-conventional shells stored in the body/nose of the tank that feed in from the top of the shield (as in B-Mode), with the Battloid Mode being rectoned into having only 1x round w/o transforming back to G/T (or give it a complicated feed mechanisim to overcome this)
Weapon Configuration Tangent Option in Spiler