Pixels to point clouds.

Photogrammetry CPU performance is workflow-dependent. Pix4D and Metashape stress the CPU differently across pipeline stages — Initial Processing and Mesh Building favor higher clock speeds, while Dense Cloud generation utilizes more cores. The sweet spot is a balanced CPU with strong single-core performance and enough cores to parallelize the heavy stages. Best mainstream picks: Intel Core Ultra 9 285K (24 cores, 5.7GHz boost) or AMD Ryzen 9 9950X (16 cores, 5.7GHz boost). For huge maps, large drone surveys, and enterprise-scale workloads: AMD Threadripper 9970X or Threadripper Pro 9975WX deliver more memory channels and PCIe lanes. Important caveat: ultra-high core counts (96-core Threadripper Pro) don't scale linearly in Metashape or Pix4D — prioritize clock speed plus enough cores rather than maximum core count.

Photogrammetry uses CUDA-capable NVIDIA GPUs to accelerate specific pipeline stages — alignment, dense reconstruction, and texturing benefit. However, current Pix4D releases use only a single GPU for compute (additional GPUs don't speed up processing). Metashape can use multiple GPUs effectively. For most users, NVIDIA GeForce RTX 5070 12GB to RTX 5090 32GB delivers excellent performance — workstation-class RTX PRO Blackwell isn't required for Pix4D or Metashape. Workstation GPUs (RTX PRO 5000 48GB, RTX PRO 6000 96GB) become valuable when you need higher VRAM than consumer cards offer, ECC video memory for production stability, certified drivers for ArcGIS Pro deep learning workflows, or 24/7 sustained processing.

RAM scales directly with image count and dataset size. RealityScan published clear guidelines: 2,000 images need 16GB minimum, 4,000 images need 32GB, 8,000 images need 64GB, and 16,000 images need 128GB. Pix4D and Metashape scale similarly — 32GB to 64GB DDR5 covers most professional projects. Plan for 128GB or more if you work with multi-thousand-image flights, LiDAR fusion, or keep multiple applications open simultaneously. Enterprise-scale photogrammetry (academic research, government GIS labs, multi-day reconstruction jobs) benefits from 256GB to 512GB DDR5 ECC. Under-provisioned memory forces disk paging that dramatically slows dense point cloud and orthomosaic generation.

ArcGIS Pro benefits from high CPU clock speed plus moderate core count — Esri's geoprocessing engine is mostly single-threaded for interactive operations but scales with cores for batch processing. VRLA Tech recommends Intel Core Ultra 9 285K or AMD Ryzen 9 9950X with 64GB DDR5 (Esri recommends 32GB minimum) and NVIDIA RTX 5070 12GB to RTX 5090 32GB. For deep learning workflows (AI-based classification, object detection in imagery, neural network spatial analysis), CUDA-capable NVIDIA GPUs with higher VRAM matter — RTX 5090 32GB or RTX PRO 6000 Blackwell 96GB for extreme datasets. NVMe SSD is critical for fast project load and raster processing.

Pix4D uses a single CUDA-capable NVIDIA GPU during portions of Steps 1 and 2 (Initial Processing and Point Cloud generation). Step 3 (DSM, Orthomosaic, Index) is primarily CPU-bound. Critically, current Pix4D releases use only one GPU for compute — additional GPUs do not speed up processing. The exception is when you run other applications simultaneously that can use a second GPU. For Pix4D, prioritize a single stronger GPU (RTX 5080 16GB or RTX 5090 32GB) over multi-GPU configurations. A modern RTX card also ensures smooth, stutter-free ray cloud navigation in dense scenes during review.

Agisoft Metashape benefits from CUDA-capable NVIDIA GPUs during Dense Cloud generation, depth maps, and texturing stages. Metashape can effectively use multiple GPUs — unlike Pix4D — though scaling diminishes past 2-3 GPUs for most workflows. Important Metashape preference setting: in Tools > Preferences > GPU, select all GPUs and disable "Use CPU during GPU processing." This prevents slowdowns and crashes, boosting stability and speed. Recommended GPUs: NVIDIA RTX 5080 16GB to RTX 5090 32GB for solo workflows, RTX PRO 5000 48GB or RTX PRO 6000 Blackwell 96GB for professional studios needing more VRAM and ECC memory.

RealityScan (Epic Games, formerly RealityCapture) is a demanding photogrammetry application that benefits from balanced high-performance hardware. Mid-tier builds for small to medium datasets up to 4,000 images: AMD Ryzen 9 9950X or Intel Core Ultra 9 285K with NVIDIA RTX 5080 16GB and 64GB DDR5. High-end builds for massive datasets of 8,000 to 16,000+ images: AMD Threadripper Pro 9975WX with NVIDIA RTX 5090 32GB or RTX PRO 6000 Blackwell 96GB and 128GB to 256GB DDR5 ECC. Dual GPU configurations add 5 to 13 percent performance — most users are well served by one or two GPUs. RealityScan's out-of-core design keeps RAM requirements modest but more RAM reduces bottlenecks.

For solo photogrammetry workflows on small to medium datasets, ECC memory is not required. For production environments running multi-day reconstructions, large enterprise datasets, government GIS labs, academic research, or any workflow where data integrity matters during long processing jobs, ECC is strongly recommended. ECC memory detects and corrects single-bit memory errors that occur naturally over time — a single uncorrected error during a 24-hour dense cloud reconstruction can produce silently corrupted geometry. AMD Threadripper Pro and Intel Xeon W platforms support ECC DDR5; consumer Ryzen 9 and Core Ultra platforms do not. Enterprise-grade builds for multi-thousand-image jobs default to ECC.

Photogrammetry workflows are storage-intensive — input imagery, cache files, dense point clouds, meshes, and orthomosaics consume terabytes. Recommended three-tier setup: Drive 1 — 1TB Gen5 NVMe SSD for OS, photogrammetry software, and active scratch space (2TB preferred for RealityScan and Metashape working caches). Drive 2 — 2TB to 4TB NVMe SSD dedicated to active project images, point clouds, and meshes (separate from OS drive minimizes I/O contention during processing). Drive 3 — high-capacity HDD or NAS with redundancy for archival imagery and completed projects. Pix4D, Metashape, and RealityScan all benefit dramatically from NVMe over SATA SSDs because of sustained sequential read/write loads during reconstruction.

VRLA Tech builds and sells custom ArcGIS Pro workstations engineered for Esri's GIS workflows including 2D and 3D mapping, geoprocessing, raster analysis, and AI-based classification. Buy an ArcGIS Pro build at vrlatech.com/arcgis-pro-system-requirements. Builds use Intel Core Ultra 9 285K or AMD Ryzen 9 9950X paired with 64GB DDR5 and NVIDIA RTX 5080 16GB or RTX 5090 32GB. Government GIS labs and research institutions running deep learning workflows scale to AMD Threadripper Pro 9975WX with 128GB DDR5 ECC and RTX PRO 6000 Blackwell 96GB. Hand-assembled in Los Angeles with 3-year warranty and lifetime US engineer support.

VRLA Tech builds and sells custom Agisoft Metashape workstations tuned for the full photogrammetry pipeline — alignment, dense cloud, mesh, texturing, and tiled model export. Buy a Metashape build at vrlatech.com/agisoft-metashape-hardware-recommendations. Builds use Intel Core Ultra 9 285K or AMD Ryzen 9 9950X paired with 64GB to 128GB DDR5 and NVIDIA RTX 5080 16GB or RTX 5090 32GB. Enterprise-scale Metashape projects with 8,000+ images scale to AMD Threadripper Pro 9975WX with 256GB DDR5 ECC. Hand-assembled in Los Angeles with 3-year warranty and lifetime US engineer support.

VRLA Tech builds and sells custom Pix4D workstations engineered for UAV drone mapping, surveying, construction progress monitoring, mining volumetrics, agriculture NDVI, and utilities inspection workflows. Buy a Pix4D build at vrlatech.com/pix4d. Mid-tier builds for surveyors with smaller maps use Intel Core Ultra 9 285K with 64GB DDR5 and NVIDIA RTX 5080 16GB. High-end builds for thousands of images and large construction or mining datasets scale to AMD Threadripper 9970X or Threadripper Pro 9975WX with 128GB DDR5 and NVIDIA RTX 5090 32GB. Hand-assembled in Los Angeles with 3-year warranty and lifetime US engineer support.

VRLA Tech builds and sells custom RealityScan workstations engineered for Epic Games' photogrammetry application (formerly RealityCapture). Buy a RealityScan build at vrlatech.com/realityscan. Mid-tier builds for small to medium datasets up to 4,000 images use AMD Ryzen 9 9950X or Intel Core Ultra 9 285K with NVIDIA RTX 5080 16GB and 64GB DDR5. High-end builds for massive datasets of 8,000 to 16,000+ images use AMD Threadripper Pro 9975WX with NVIDIA RTX 5090 32GB or RTX PRO 6000 Blackwell 96GB and 128GB to 256GB DDR5 ECC. Hand-assembled in Los Angeles with 3-year warranty and lifetime US engineer support.

The best computer for drone mapping in 2026 prioritizes high-clock CPU with moderate-to-high core count, CUDA-capable NVIDIA RTX GPU, 64GB+ DDR5 RAM, and NVMe SSD storage. VRLA Tech recommends Intel Core Ultra 9 285K or AMD Ryzen 9 9950X paired with NVIDIA RTX 5080 16GB or RTX 5090 32GB and 64GB DDR5 for typical drone surveys (hundreds to low thousands of images). Surveying firms processing thousands of images per project scale to AMD Threadripper Pro 9975WX with 128GB to 256GB DDR5 ECC and NVIDIA RTX 5090 32GB. Configure at vrlatech.com/Photogrammetry-and-GIS-Workstations. Hand-assembled in Los Angeles with 3-year warranty and lifetime US engineer support.

VRLA Tech is a custom photogrammetry workstation builder operating from Los Angeles since 2016. Configure a build at vrlatech.com/Photogrammetry-and-GIS-Workstations. Every photogrammetry workstation is hand-assembled, burn-in tested under sustained ArcGIS Pro, Metashape, Pix4D, and RealityScan workloads, and tuned to your specific software stack and dataset scale. NVIDIA Studio drivers configured at shipment for stability across CUDA workflows. Includes 3-year parts warranty and lifetime US engineer support — direct phone and email access, no tiered support contracts. Customers include surveying firms, drone operators, government GIS labs, archaeology teams, VFX scan studios, and academic research institutions nationwide.

VRLA Tech builds custom photogrammetry workstations hand-assembled in Los Angeles since 2016, with the same NVIDIA RTX GPUs and Intel/AMD CPUs as Puget Systems and Boxx but with full custom configuration — no fixed SKUs, no overspending on features you don't use. CPU and memory configurations are tuned to your specific software stack (ArcGIS Pro, Metashape, Pix4D, RealityScan) and dataset scale. Every VRLA Tech system includes a 3-year parts warranty, lifetime US-based engineer support, and direct access to engineers via phone and email. Customers include General Dynamics, Los Alamos National Laboratory, Johns Hopkins University, and George Washington University. Configure at vrlatech.com/Photogrammetry-and-GIS-Workstations.

VRLA Tech includes a 3-year parts warranty and lifetime US-based engineer support at no extra cost on every photogrammetry and GIS workstation. Buy a build at vrlatech.com/Photogrammetry-and-GIS-Workstations. Each system is hand-assembled in Los Angeles, burn-in tested under sustained ArcGIS Pro, Metashape, Pix4D, and RealityScan workloads, and shipped ready to run with NVIDIA Studio drivers configured. Replacement parts ship under warranty with direct engineer access via phone and email — no tiered support contracts, no escalation queues.