This is the machine I work from every day. It runs 12 Hyper-V VMs, 37 VirtualBox VMs and 12 Docker containers — not all at once, but the headroom is there when I need it.

CPU: Intel Core i7-8750H

Six cores, twelve threads, 2.2GHz base with a 4.1GHz single-core boost. The 8750H is an 8th-gen Coffee Lake H chip from 2018 and it still handles everything I throw at it. Compiling, running VMs, AI inference — the twelve threads keep latency low when several things compete for resources at once.

RAM: 64GB DDR4 2667MHz

Two 32GB sticks running dual-channel at 2667MHz. Most laptops of this generation shipped with 16GB. Upgrading to 64GB is the single best thing I did to this machine. Hyper-V allocates RAM dynamically, so spinning up a 2016 Server VM no longer tanks everything else running.

If you want to do the same upgrade, 2x 32GB laptop SODIMMs at 2667MHz are what you need:

32GB SODIMM DDR4 2666/2667 Kit on Amazon

Storage: 2x Samsung 970 EVO Plus 2TB NVMe

Two drives, 4TB total, both fast. The 970 EVO Plus uses Samsung's TLC NAND with an SLC write cache that sustains sequential writes above 3GB/s before the cache fills. For VM disk images, that matters. VMs boot off NVMe at a noticeably different speed than SATA.

One drive holds Windows, my active projects and the Hyper-V VM store. The second holds VirtualBox images, bulk data and overflow. I also have a 9.3TB external drive on USB 3.0 for cold storage and backups.

Samsung 970 EVO Plus 2TB NVMe on Amazon

GPU: NVIDIA GTX 1060 Max-Q 6GB + Intel UHD 630

The GTX 1060 Max-Q runs at lower power than the full 1060 to stay within laptop thermals. It handles CUDA tasks for local AI inference and renders the desktop while the integrated Intel UHD 630 handles display output when the Nvidia chip powers down.

6GB VRAM runs quantized 7B models in Ollama at reasonable speed. Not fast enough to replace a desktop GPU for AI work, but it runs.

Network: Intel Wireless-AC 9260 160MHz + Tailscale

The Intel 9260 supports 160MHz channel width on Wi-Fi 5, which pushes real-world throughput past 800Mbps on a compatible router. Tailscale runs as a mesh VPN overlay so I reach this machine from anywhere without port forwarding.

OS: Windows 11 Pro

Windows 11 Pro for Hyper-V and WSL2. All Docker containers run through Docker Desktop backed by WSL2. VirtualBox runs alongside Hyper-V using the Hyper-V backend option in VirtualBox 7+.

The Full Picture

A six-year-old laptop running 64GB RAM and 4TB of NVMe storage, hosting a full enterprise lab environment. I've seen people spend more on a new machine than these upgrades cost.

Component Spec
CPU Intel i7-8750H, 6C/12T, 4.1GHz boost
RAM 64GB DDR4 2667MHz
Storage 2x Samsung 970 EVO Plus 2TB NVMe
GPU GTX 1060 Max-Q 6GB + Intel UHD 630
OS Windows 11 Pro
Network Intel AC 9260 + Tailscale

As an Amazon Associate, IT Lad earns from qualifying purchases made through links on this page. Read the full affiliate disclosure.

Previous Post Next Post