How to Find GPU and Data Center Capacity: A Buyer's Playbook
Everyone wants GPUs and megawatts, and almost nobody can find them quickly. A practical process for turning a vague need into a shortlist of real, available options.
Blog
Notes on AI infrastructure.
How to find, size and buy data center and GPU capacity, plus the architecture, memory and economics of running AI in production.
Colocation, GPU Cloud or Your Own Build: How to Choose
Three ways to get AI capacity, three very different commitments. How to weigh control, speed, cost and risk before you pick a path.
How Many Megawatts Do You Need? Sizing a GPU Deployment
Power is the unit capacity is sold in. A simple method for turning a GPU count into a megawatt figure you can take to providers, with a worked example.
Liquid Cooling or Air Cooling for AI Racks: What Changes at High Density
GPU racks pack far more power into the same footprint than traditional servers. Where air cooling runs out, what liquid cooling involves, and how to check a facility's claims.
InfiniBand or Ethernet for AI Clusters: Choosing the Network
The network decides how well many GPUs behave as one system. How to think about InfiniBand and Ethernet, and what to ask a provider about fabric design.
20 Questions to Ask a Data Center Before You Sign
A checklist for diligence on any data center or colocation deal, grouped by power, cooling, network, site and commercial terms.
Why AI Data Center Capacity Is So Hard to Find
Demand for power and GPUs has outrun supply in many markets. The forces behind it, and why searches and deals so often stall.
What a Data Center Broker Actually Does
Sourcing, qualifying, negotiating and managing a deal to signature. What a data center advisor does for a buyer, what to expect, and what to ask about how they are paid.
How to Read a Data Center Listing: MW, Density and "Available"
Listings use terms that sound precise but mean different things. What megawatts, rack density, status and availability really tell you, and what to verify yourself.
Powered Land and Shell Space: How Operators Can Reach AI Buyers
AI buyers search for specific power, density and timing. How data center operators and developers can present capacity, including powered land, so the right buyers find it.
Building AI Inference Silicon in the Age of Hyperscale GPUs
Why start a semiconductor company for AI inference now, when GPU supply dominates the conversation? Because the customers who need something different aren't hyperscalers.
From Training to Serving: How AI Infrastructure Priorities Shift
Once a model ships, the infrastructure conversation changes completely. What mattered during training rarely matters most once you're serving real traffic.
Private AI and On-Premise LLMs: Rethinking Enterprise Infrastructure
More enterprises want AI running inside their own infrastructure, not just behind someone else's API. That shift changes what accelerator architecture needs to optimize for.
Why Silicon Alone Doesn't Win the Inference Race
The best accelerator in the world underperforms without software that knows how to use it. Why hardware-software co-design, not silicon alone, determines real-world performance.
Tokens per Dollar: Why Inference Economics Beat Peak FLOPS
Peak theoretical performance makes for a good spec sheet. Tokens per dollar is what actually determines whether an AI product is viable to run in production.
Tokens per Watt: The Efficiency Metric AI Infrastructure Can't Ignore
Power isn't a line item you deal with later. In AI inference, it's often the constraint that decides how much capacity you can actually deploy.
What Makes Transformer Inference Different from Training
Training and inference both run on transformers, but they stress compute in almost opposite ways. Conflating the two is how infrastructure decisions go wrong.
INT8 and INT4: How Lower Precision Unlocks Inference Efficiency
Lower-precision math sounds like a compromise until you look at what it actually costs in accuracy versus what it buys back in speed, power, and memory.
The Memory Wall: Why Bandwidth Decides Inference Performance
Compute gets the marketing slides, but in production inference, memory bandwidth is usually the real ceiling. Here's why the memory wall matters more than peak FLOPS.
Why Purpose-Built Beats General-Purpose for AI Inference
General-purpose GPUs were built to do everything reasonably well. AI inference rewards architecture that does one thing exceptionally well. Here's why that distinction matters.