Panorama Set to Cloud - Processing Performance

We have panorama sets we’ve been doing well with that can have up to the max (100 I believe) hot spots that allows the customer to walkthrough their home on a controlled basis. We are using Lumen for this and would like to run 8k but end up doing 6k due to processing time. We can’t really go under 6k because the result isn’t really good unless you are using a small cellphone only to view.

We purchased a new workstation earlier this year with the following specs;
Case:Phanteks Black Steel / Tempered Glass ATX Mid Tower Computer Case Case Fans:High-performance silent 120mm fans with Anti-Vibration fan mounts

CPU / Processor
Intel Core i9-14900k Raptor Lake 3.2 GHz (6.0 GHz Turbo Boost) 125W 36MB L3 (24 Cores / 32 Threads) CPU Cooling:360mm AIO Liquid CPU Liquid Cooler - Closed Loop CPU Thermal Compound:Antec Nano Diamond Thermal Compound Formula 7 w/ Diamond particles

Motherboard
ASUS ProArt Z790-Creator ATX Content Creator Motherboard, 2x Thunderbolt 4, 1x 10G LAN, 4x M.2

Memory
64GB (2 x 32GB) 288-Pin DDR5 5200 (PC5 41600) Desktop Memory

Power
1350W - 80+ GOLD Certified PCIe Gen.5 12VHPWR Connector Power Supply

Operating System: Pre Install Microsoft Windows 10 Pro 64-bit

GPU
1x NVIDIA GeForce RTX 4090 24GB GDDR6X Graphics Card
Drive
1TB M.2 NVMe PCIe 4.0 x4 SSD - read/write 5000/4800 MBps SATA-SAS RAID

I was hoping to get faster turn arounds but just isn’t happening. We previously used ray tracer but again, wanted to run with Lumen. We need to get another workstation due to the number of orders we have but I like to solve the performance of producing these. I’m fine with running sets that take 8-10 hours but when we are around 14-28 hours @ 6k, its just too long

I’m not against get another workstation with a new 48GB graphics card but like to know if I’m missing something hardware wise that would accelerate the production. Would adding a secondary 24GB graphics card produce better results, from my supplier, they didn’t believe twinmotion performance would be increased with secondary GPUs.