tinybox ships red v2 with 4x 9070 XT and 64 GB GPU RAM for $12,000

• tiny corp is shipping the tinybox page red v2 as a 4x Radeon 9070 XT box with 778 TFLOPS FP16, 64 GB of total GPU RAM, a 32-core AMD CPU, and Ubuntu 24.04 for $12,000.
• The same product page also lists a higher-end green v2 Blackwell system at $30,000 with 4x RTX PRO 6000 Blackwell GPUs, 3,086 TFLOPS FP16, and 384 GB of GPU RAM, plus an exabox roadmap claim of roughly 1 exaflop.
• The Hacker News thread quickly turned into a debate over whether the bundled tinygrad stack justifies the premium over a DIY rig, colo, or cloud, with one commenter calling out “vanilla Ubuntu 24.04” and 1 Gb/s networking in the base system.
• Discussion in the community highlights also questioned workload fit: one commenter said there is “no way” the red v2 handles a 120B model as implied, while newer thread activity summarized in fresh discussion says the clearer pitch is a vertically integrated, “just works” tinygrad environment rather than raw price-performance.

The shipping announcement is straightforward. According to the tinybox page, red v2 is the entry configuration: four 9070 XT GPUs, 64 GB of aggregate GPU memory, 778 TFLOPS FP16, up to 192 GB of system RAM, and a rack-mountable Ubuntu 24.04 box for $12,000. The same page also positions the line upward, with a $30,000 green v2 Blackwell configuration and an “upcoming exabox” described as roughly 1 exaflop.

What makes this more than another small GPU server launch is the software angle. The thread summary says engineers are treating Tinybox as a test of whether bundled hardware plus tinygrad integration is worth paying for versus assembling comparable parts yourself. That framing got sharper in the newer fresh discussion, which says the main appeal may be a vertically integrated, “just works” setup that avoids CUDA and PyTorch toolkit friction.

The pushback is equally concrete. In the community highlights, one commenter said “there is apparently no reason whatsoever to buy this hardware” given the pricing, while another argued there is “no way” the red v2 is doing meaningful work with a 120B model. The same discussion also questioned the exabox claim that it can “function as a single GPU,” which leaves the most ambitious part of the roadmap looking more like an architectural question than a settled product detail.