Build
We build faibuddy, our AI advisor for founders. Working on a live AI product grounds our view in product decisions, customer behaviour and the limits of current technology.
faibuddy Thesis
AI is not a software category. It is a stack.
AI companies depend on energy, chips, infrastructure, models and applications. A decision or constraint at one layer can change the opportunity at another. We look for companies within each layer and where those dependencies create a new market.
The AI industrial stack
The technology changes by layer. The work of building a company repeats.
- Applications & Embodied AI
- Models
- Infrastructure
- Chips & Compute
- Energy
The company-building layer
Capability to Product to Market to Enduring business.
Disciplines: Product, Business model, Go to market, Partnerships, Regulation, Governance.
Freedom with AI participates in three ways: Build, Advise and Invest.
The technology changes by layer. The work of building a company repeats.
The layers
Five layers, one system.
Each layer has different technology, capital needs, margins and paths to market. Those differences matter when deciding what kind of company can be built there.
-
Applications & Embodied AI
Products at this layer are not limited to screens. They include software, scientific tools, robots, vehicles and other systems that can sense, decide and act.
-
Models
Models range from proprietary frontier systems to open-source and domain-specific alternatives. Their capability, cost, availability and deployment constraints shape what product teams can build.
-
Infrastructure
Data centres, networking, cooling, orchestration, evaluation, security and deployment determine whether models can be used reliably and affordably at scale.
-
Chips & Compute
Processors, memory, interconnects and access to supply shape performance, cost and who can compete.
-
Energy
Every workload consumes power and produces heat. Generation, transmission, siting and cooling affect where capacity can be built and how quickly it can grow.
Where we work
The company-building layer.
We do not design chips, build power plants or engineer data centres. Our work begins where technical capability has to become a product, earn customer trust and support a durable company. Those questions are similar across the stack, even when the technology is not.
- Which problem is important enough to build a company around?
- Who has it, how do they solve it today and what will they pay for?
- Which technical capability creates a meaningful product advantage?
- Which business model works with the costs, margins and capital required?
- Which dependencies, partnerships or supply constraints determine scale?
- What evidence, safeguards, regulation and governance are required for trust?
- What should remain outside the company’s scope?
How we participate
Build, Advise and Invest
Advise
We advise founders on decisions that shape the company: which problem to pursue, what to build, who will buy, how to earn trust and where to focus.
How advisory worksInvest
We are beginning to back a small number of founders with our own capital when we have conviction and can contribute beyond the investment.
Building gives us direct operating feedback. Advising exposes us to different markets and company stages. Investing extends that perspective over a longer horizon.
Where we invest
Our view spans the whole stack, but our direct investment focus is narrower. We look at infrastructure and tooling, deployment, evaluation, security and reliability, industry-specific and regulated AI, robotics and autonomy software, and products that connect models to real operations.
For semiconductors, energy and capital-intensive infrastructure, we may invest alongside specialists with deeper technical and project-finance expertise. We follow the whole stack and invest directly where we can form an independent view.
Advisory and investing are separate decisions. Working with us does not require or promise an investment.
The five-layer stack builds on Jensen Huang’s framing of AI as infrastructure. The company-building layer reflects our own work with founders. NVIDIA on the five layers of AI