George Hotz's tinybox just hit the front page of Hacker News again — and this time, it's shipping. Here's why a petaflop in your closet changes everything about who controls AI.
There's a computer sitting on a shelf at tinygrad.org right now that costs $12,000, fits in a closet, and delivers a petaflop of compute. It's called the tinybox red v2, it's in stock, and it ships within a week.
That sentence should terrify every cloud AI provider on the planet.
The Cloud AI Tax Is Real
Let's do some napkin math.
A single NVIDIA A100 on AWS costs roughly $3.06 per hour on-demand. Running 24/7? That's $2,203/month. A realistic 4x A100 setup runs $8,812/month. Over a year: $105,744 just for compute.
The tinybox red v2 costs $12,000 once. Four 9070 XT GPUs delivering 778 TFLOPS of FP16 with 64GB total GPU RAM. Electricity adds maybe $200-300/month at full load. After roughly two months of continuous use, you've broken even. Everything after that is pure savings.
The green v2 is even starker. Four RTX PRO 6000 Blackwell GPUs with 384GB GPU RAM and 3,086 TFLOPS at $65,000 — about six months of equivalent cloud costs.
Why This Is Different From "Build Your Own"
People have been building DIY deep learning rigs for years. The tinybox is different because it's the whole stack, done right:
- Ships with Ubuntu, tinygrad, and PyTorch pre-installed
- BMC for remote access
- Benchmarked on MLPerf Training 4.0
- Engineered power delivery for sustained GPU loads
- Under 50 dB cooling for home/office use
- OCP 3.0 slots for future expandability
And this line from tiny corp tells you everything about their philosophy:
"We don't have a stupid cloud service, you don't have to create a tiny account to set it up, and we aren't tracking how you use the box."
That's not just a feature. That's a worldview.
The Privacy Argument Nobody's Making Loudly Enough
Every serious AI workload runs through someone else's infrastructure. When you fine-tune on AWS, Amazon can see your training data. When you run inference through OpenAI's API, every prompt passes through their servers.
For healthcare organizations, law firms, financial institutions, defense contractors, journalists, and activists — "trust us" is an insufficient security model.
Self-hosted AI compute isn't just cheaper. It's the only architecture that provides actual data sovereignty.
The Open Source Stack Is Finally Good Enough
The hardware argument only works if the software is there. Today, it is:
Inference: Ollama, llama.cpp, vLLM, LM Studio
Training: HuggingFace transformers with LoRA/QLoRA, Axolotl, Unsloth
Agent frameworks: Open-source frameworks like OpenClaw let you run autonomous AI agents on your own hardware with full control
Models: Llama 3.x, Mistral, DeepSeek, Qwen, Phi, Gemma — the open model ecosystem is rich and getting richer
Two years ago, local AI meant massive quality compromises. Today, a well-configured local setup with Llama 3.3 70B produces results 80-90% as good as frontier proprietary models for most practical tasks.
Who This Is Actually For
Makes sense for:
- ML researchers needing sustained compute without grant money disappearing into AWS
- Pre-Series A AI startups
- Enterprise teams needing data sovereignty
- AI consultancies wanting predictable costs
- Serious hobbyists ($12K red box is genuinely reasonable)
- Anyone running AI agent frameworks locally
Doesn't make sense for:
- Teams needing burst compute (100+ GPUs for days)
- Frontier-scale model training (thousands of GPUs)
- People who just want to use ChatGPT
- Anyone uncomfortable with Linux CLI
The Bigger Picture
Actual democratization of AI requires two things: open models and accessible compute. We have the first. The tinybox delivers the second.
When a graduate student in Nairobi or a startup in São Paulo can buy a machine with the same training capabilities as a well-funded Silicon Valley startup, the geography of AI innovation changes.
The cost of serious AI compute is falling. The software is maturing. The models are open. The cloud AI cartel has been operating without meaningful competition for years. The tinybox is a credible threat to that oligopoly — not because it replaces the cloud entirely, but because it gives people a choice.
Own your compute. Own your data. Own your future.
Originally published on TechPulse Daily
Top comments (0)