See Beyond The Noise.
Research the infrastructure powering artificial intelligence and discover the companies shaping the next decade.
Foreword
Demystifying the buildout.
One of humanities biggest dilemmas is fearing what we do not understand. The majority of people fundamentally do not understand what ai is, how it works, or the infrastructure the allows the technology to function. By demystifying this evolution of human ingenuity we hope to help you understand this technology if you do not already. This is not a sentient life form that will take our jobs and enslave us, it is a program derived from machine learning that will serve as a tool to accomplish tasks given to it by the user. It the most powerful tool we’ve ever had access to that will reshape what humanity has perceived to be possible.
There are many doubts surrounding this technologies efficacy as well as its profitability that are causing tremendous turmoil amongst both people in and out of the know. There is a massive information gap between ai skeptics and the enthusiasts who have adopted this technology early on. Allow us to provide rudimentary knowledge that will allow you to make a more informed decision for ai’s potential usefulness to your daily life or enterprise as well as investment opportunities related to the buildout of ai before its applications are widely realized.
The biggest companies on earth are currently funding the greatest expansion in infrastructure in human history. This is a multi trillion dollar investment into ai development that is well underway and is projected to increase over the next few years. This cap ex is building the foundation for future growth. This is a once in a lifetime investment opportunity and many of the most important companies that will benefit from this technology are at attractive valuations.
The thesis
AI is not in a bubble. Nothing fundamental has changed for semiconductors or the hyperscalers — demand is still far past what supply can provide.
Recent growth has led many to believe we are in a bubble. The surface has barely been scratched. Markets are operating healthily by rotating out of assets that moved too far too fast; momentum is cooling and some names are down 30-50% in a month on rotation, not on fundamentals. If this were truly a bubble, the entire market would fall alongside semiconductors.
Capex will still increase, demand will stay high, and the gains made are real against strong balance sheets. The hyperscalers are now watched for capex converting into revenue — and that conversion fuels the next leg of AI's explosive growth. Many of the most important companies that benefit are still at attractive valuations.
- $7T
- Investment flowing into data centers
- $700B+
- Hyperscaler capex in 2026 alone
- 3x
- Data center demand growth by 2030
- 60%
- More transmission capacity needed by 2030
Jensen Huang's framing
AI as a five layer cake
Bottom to top — energy at the base, applications on top. Select a layer to jump to it.
Layer 1
01
Energy
The grid is the bottleneck
Energy is required for training and inference, for cooling every machine processing every piece of data, and for running the network itself. The U.S. grid has not been meaningfully updated in nearly 60 years while demand is set to climb 2% a year for two decades. Liquid cooling alone consumes 40% of a data center's electricity. Independent estimates call for 60% more transmission capacity by 2030.
Layer 2
02
Compute
Semiconductors are the heart of computing
This layer is what an AI model runs on to perform the mathematical calculations that turn a question into an answer. Creating a semiconductor chip is arguably the most technologically challenging process humanity has created — design, then deposition, photolithography, etching and doping repeated hundreds of times. The hyperscalers are set to spend over $700 billion in 2026 alone, and even more in 2027, with capital flowing to chipmakers in a frenzied race for compute.
Layer 3
03
Infrastructure
Data centers — where compute physically lives
Data centers are the physical space where all our online data resides and where AI training and inference happen. When a company says it is compute constrained, it means it lacks access to data centers and the machines inside them. Demand for data centers is projected to triple by 2030 with $7 trillion of investment behind it. Once those facilities come online, the providers of compute see revenue.
Layer 4
04
Models
Large language models are sophisticated mathematical functions that predict what word should come next for a piece of text, assigning probabilities across every possible next word. A model's behavior depends on its parameters, or weights — hundreds of billions of them, random at the start of training and refined across trillions of strings of text. That process is only possible on the high tech GPUs we have today, which run enormous numbers of operations in parallel.
AI used to read one word at a time until Google developed the transformer, which let a model soak in all the words at once. Language is encoded as numbers, and through an operation called attention, those lists of numbers talk to one another and refine the meanings they encode based on surrounding context — all in parallel. Researchers provide the framework; the generation is entirely up to the model. That is why we call it artificial intelligence.
If you used AI and were unimpressed, you do not have access to what the frontier labs are working with. A recent OpenAI model solved the planar unit distance problem posed by Paul Erdős in 1946. An Anthropic model found mathematical flaws in algorithms used for data encryption. Google's AlphaEvolve helped design their newest TPUs and discovered new mathematical truths.
OpenAI and Anthropic are the consensus leaders, and the primary constraint is no longer capability — it is cost. Amazon, Microsoft and Google provide competent alternatives at a fraction of the price, and Chinese open source models are free. Kimi k3 recently closed the gap between open and closed source. OpenAI reached 700 million customers in under two years, faster than any consumer product in history.
Who builds the models
OpenAI
Private. Alongside Anthropic, provides the most powerful models for both individuals and businesses, capable of advanced reasoning and mathematical calculations. Reached 700 million customers in less than 2 years, faster than any other consumer product in history.
Anthropic
Private. A consensus frontier leader whose models are used to solve the world's most pressing questions — one recently found mathematical flaws in algorithms used for data encryption.
Google DeepMind
Gemini is trained on Google's own TPUs and offered at a leading low price point; AlphaEvolve helped design their newest TPUs and discovered new mathematical truths.
Microsoft AI
A strong lineup including MAI-Image-2.5, MAI-Transcribe-1.5 and MAI-Thinking-1, plus value versions competitive with the world's leading cost-effective models.
Amazon & NVIDIA
As with many other layers, both provide models for individuals and businesses, branching into cyber security and coding models on top of general purpose ones.
Open source
Chinese open source models are free, and Kimi k3 recently closed the gap between open and closed source. Capability is no longer the constraint — cost is.
Layer 5
05
Applications
The next frontier
This is where the technology is applicable to the real world and its profitability is widely realized. We are currently not at this point but this technology is exponentially evolving and super intelligence is already available at certain price points. The potential for economic expansion as a result of this technology is theoretically limitless and the advancements many different industries will experience will be life changing. Robotics will enter the picture towards the end of the decade and automated services will become visible to the public in the coming decades.