hristosbilis.ai Search

Search

Try Merck & Co.validationAnnex 22Discovery

    The AI pipeline

    This page maps the nine parts of an AI system, from the power plant to the model a person uses, and where each part stands now.

    I wrote it for computer system validation (CSV) and quality assurance (QA) leads, and for AI platform and governance leads at regulated pharma companies.

    Notes on where AI is now and where it is going are on Frontier.

    The nine parts

    Power and chips first, then putting a model to work, then how a model gets made.

    Metro line map of the AI pipeline. The line starts at a power plant and runs through nine numbered stations in three groups, along the top row and back along the bottom (top to bottom on phones). Power and chips: 1 Energy, 2 Chips. Putting it to work: 3 Connecting the model to your work, 4 Verifying the output, 5 Putting the model to work, 6 How people actually use it. Making a model: 7 Where the data comes from, 8 Who's training models for pharma, 9 What the model actually is. The line ends at station 9. Below the line, a band labeled Forces outside the pipeline lists Money, Talent, Rules and law, and Security.

    The nine parts of the AI pipeline

    Where each part of the AI pipeline stands now

    The series

    AI pipeline

    The nine parts of the AI pipeline

    A map of the nine parts of the AI pipeline, grouped as power and chips, putting a model to work, and making a model, plus the forces outside it that shape every part. The first piece in the series.

    AI pipeline

    Energy, where it stands now

    Electricity for AI as of mid-2026: how much data centers use, why the total rises while a simple query uses less, and why a grid connection is the slow step.

    AI pipeline

    Where each part of the AI pipeline stands now

    A short account of the nine parts of the AI pipeline, in order. Energy and chips have their own pieces. The other seven are marked coming soon, and a last piece will look at bottlenecks across the chain.

    AI pipeline

    Chips and hardware, where it stands now

    AI chips and hardware in 2026 for pharma and biotech teams: what GPUs, memory, packaging, chip makers and export rules mean for your AI compute, budgets, vendor risk and change control, and why memory is the slow step.