In Response to Satya Nadella, CEO of Microsoft

You Pay for AI Twice.
Keep the Second Payment.

Nadella named the trap. To make a model useful you feed it the knowledge that makes you unique, and it leaks back one correction at a time. His fix is a trust boundary you own. Here is where that boundary lives in your stack, and what governs the meaning inside it.

Alexander Braun · July 2026 · 9 min read
Key Insights
In this article
"You essentially pay for intelligence twice, once with money, and again with something even more valuable: the proprietary knowledge you must reveal to make that intelligence useful."
Satya Nadella, CEO of Microsoft. July 2026
2×
Times an enterprise pays for AI: once in cash, and once in the proprietary knowledge it reveals to make the model useful.
Satya Nadella, Microsoft · July 2026
60%
Share of organizations forecast to abandon AI projects through 2026 without AI-ready, governed data.
Gartner · Feb 2025
Jan ’27
When OpenAI ends new self-serve fine-tuning job creation. The learning infrastructure you rent can be narrowed on the vendor's timeline, not yours.
OpenAI policy change · 2026
Zero
Amount of your actual business data LazyFox ever sends to a model provider. Data is indexed once and abstracted with probabilistic methods, so the model never sees your real data, not even at indexing.
LazyFox architecture
Executive Summary
Key Finding
In the AI age the exposure flips from seller to buyer. To get a useful answer you reveal proprietary knowledge, and the model learns from the exhaust: the prompts people write, the tools agents call, and above all the corrections experts make when it is wrong. That institutional know-how leaks imperceptibly, trace by trace, eval by eval, and by default it is retained by the provider, not the enterprise.
Root Cause
Learning flows one way. Providers hold the learning infrastructure, reserve the right to learn from usage, and impose restrictive terms on distillation. The asymmetry compounds, so economic value converges toward the owners of the learning infrastructure rather than the creators of the knowledge. Nothing in the enterprise stack captures the governed definitions and evals that would keep that value inside the firm.
Market Recommendation
Build a hard trust boundary inside the tenant where data, traces, evals, adapted weights, and memory accumulate together, and claim the right to use model outputs to tune your own models. Nadella frames it as five tests: Control, Capability, Choice, Cost, and Compound. Combine the first four and you get a continuous learning loop that belongs to the firm, not the vendor. "What you create should belong to you."
How LazyFox Delivers on This
Your Actual Data Never Reaches the Model
LazyFox never sends your real business data to a model provider, not even once. It indexes your stack a single time and abstracts everything with probabilistic methods, so the model works from governed code, never from your raw context. The exhaust has nothing to leak because your data never leaves the boundary.
The Veteran Stays When the Generalist Leaves
Context lives in code the company owns, across the structural, logical, and contextual layers, not in a provider’s weights. Swap the model and the institutional memory remains. This is the concrete answer to Nadella’s company-veteran test: the generalist changes, the veteran stays.
Evals and Definitions as Owned Assets
Private metric definitions, versioned business rules, and semantic drift detection live in the logical and contextual layers. What "good" means is defined, governed, and retained inside the tenant, and when a definition diverges across systems, LazyFox flags the conflict before it reaches an output.
Wrong Answers Caught Before They Ship
An agent can retrieve the right documents and still report the wrong number by applying a stale or conflicting definition. LazyFox reconciles meaning across all connected systems at runtime and surfaces the conflict before the answer reaches a board deck or an auditor.
Read the full analysis below
The Article

Arrow’s Paradox, Inverted

The classic problem exposed the seller. AI moves the exposure to the buyer, and it compounds every time you use what you bought.

In 1962 Kenneth Arrow described a problem at the heart of any market for information. Its value to the buyer is not known until the seller reveals it, but once revealed the buyer has it for free. The seller is the one at risk. Patents and NDAs exist to let an inventor show an idea without giving it away.

Nadella’s argument is that AI flips the vulnerability. Now the buyer is exposed. To make a model useful on your work you have to feed it the proprietary knowledge that makes your work valuable, and the better you want it to perform, the more of that knowledge you hand over. You pay once in cash and again in the context you reveal. Over time the asymmetry widens: the provider learns more about you with every session, while you learn almost nothing about what it takes in return.

The leak is not a single document walking out the door. It is diffuse. Models learn from what Nadella calls exhaust: the prompts your people write, the tools your agents call, and especially the corrections your experts make when the answer is wrong. Each correction is a small piece of institutional judgment, the kind a competitor could never buy, distilled into a rented system trace by trace and eval by eval.

"In consuming intelligence, you are creating intelligence. And what you create should belong to you."

Satya Nadella, CEO of Microsoft

His prescription is not more data protection. It is a trust boundary: a hard line inside the enterprise across which nothing crosses without consent, not prompts, not traces, not evals, not adapted weights, not memory. Inside that boundary an organization’s data and learning are supposed to accumulate and compound as an owned asset. That is the right idea.

The practical question he leaves open is where the boundary sits in a real stack, and what governs the meaning of the data inside it so an agent gives the same answer your business would. That is an architectural requirement, not a policy. The sections below take his five tests, Control, Capability, Choice, Cost, and Compound, and show what each demands in practice.

The Inversion

The Buyer Is Now Exposed

Every prompt, tool call, and correction is training data. In consuming intelligence, you create it, and by default someone else keeps it.

Think about what a good AI answer requires from you. Not just a question, but the context that makes the question answerable: how your business defines its terms, which system is authoritative for which number, the historical exceptions that only live in your team’s heads. Nadella’s point is that this context is the asset, and the model needs it to be useful. You cannot get the value without revealing the thing that creates the value.

The corrections are the sharpest example. When a senior analyst tells the model "no, revenue here means net of returns for the fiscal quarter ending March 31," that is not a prompt. It is a piece of your operating knowledge, the product of years of judgment, handed to a system in a form it can absorb. Multiply that across six thousand employees and a year of usage, and you have transferred a meaningful part of how your company thinks, without a line item ever appearing on an invoice.

This is why the problem is easy to miss. A single confidential file leaving the building triggers alarms. Ten thousand small corrections, each individually trivial, do not. They leave as ordinary usage, and the aggregate is a map of your judgment. If your knowledge is only ever expressed as corrections to a model you do not control, then your most valuable context has no home you own. It exists as weights and history on someone else’s infrastructure, and the moment terms change, it does not come with you.

Who keeps what you create?

If your corrections train a model you rent, the knowledge ends up on someone else’s infrastructure. If they govern a layer you own, the model is just a processor, swappable on day one.

Model-dependent stack
Enterprise Workflows
Business context, domain logic, metric definitions
AI Model (v1)
Fine-tuned with your business context
Institutional knowledge
⟳ Model deprecated / switched
AI Model (v2)
New model, starts with no company context
Institutional knowledge
LazyFox architecture
Enterprise Workflows
Business context, domain logic, metric definitions
Semantic Governance Layer
Indexed once. Versioned. Owned by the company.
Institutional knowledge
⟳ Model switched freely
Any Model (v1, v2, v3…)
Receives governed queries. Executes. Nothing more.
Without a semantic layer, the model becomes the vessel for institutional knowledge. Switch the model, lose the knowledge. The "company veteran" lives inside someone else’s system.
With LazyFox, your data is indexed once and abstracted before anything reaches a model. The model gets governed queries, never your raw business data. Swap the model freely. The semantic layer persists.
Compound

A Loop That Compounds For You

Nadella’s fifth test is the point of the other four: bring them together and the learning loop hill-climbs inside a boundary the firm owns.

Nadella describes the goal as a "hill climbing machine", a system where every workflow that runs deepens the organization’s knowledge, making the next one better, cheaper, and more accurate. The question is not whether that loop exists. With any real AI usage, it does. The question is who it compounds for.

Most enterprise AI deployments compound in the wrong place. Each query goes to the model, and what it reveals about your data, terminology, and edge cases is retained by the provider, not by you. Your side of the ledger is as naive on day 500 as it was on day one, because nothing durable was kept inside your boundary.

Compounding for the firm requires that each interaction leaves something behind on your side: an output that refines your own understanding of what your data means. That refinement has to live somewhere persistent, versioned, and queryable, in code you control. It has to be structural, not just conversational, or it evaporates into a rented model the moment the session ends.

Semantic richness compounds.
Model costs don’t.

Each system connected, definition governed, and cross-system conflict resolved makes the semantic layer richer, and every subsequent request cheaper and more accurate, without growing model costs.

Month
1
Systems connected, fields indexed. Structural layer maps what exists across CRM, ERP, and data lake. Definitions drafted.
Month
3
Logical layer governed. Key metrics (revenue, pipeline, cost-to-serve) have approved, versioned definitions. Teams aligned on canonical terms.
Month
6
Cross-system conflicts resolved. "Revenue" across SAP, Salesforce, and the data lake now resolves consistently by context. Drift detection running continuously.
Month
12
Contextual layer live. Meaning shifts by team, quarter, regulatory context, all captured and version-controlled. New queries cost a fraction of month-one queries.
Month
24
Full institutional memory. Every new system integrated immediately inherits the governed definition graph. The organization’s AI capability is model-independent and compounding.
Unlike a content wiki
Bounded cost
A wiki’s maintenance cost scales with corpus size and query volume, both of which grow indefinitely. The semantic layer’s cost is governed by the definition surface, which is bounded. There are only so many ways to define "revenue."
Nadella’s framing
Compound the firm
"Bring these four together and you create your own continuous learning loop that will allow your AI investments to compound the value of your firm." The compounding happens at the definition layer, not the model layer.
Control

When Learning Flows One Way

Fair use lets providers train on public data. The status quo then restricts what you can do with your own usage. That imbalance decides who compounds.

Nadella flags an irony worth sitting with. The industry won broad rights to train on public data as fair use, which enabled the current generation of models. The same regime then turns around and imposes restrictive terms on distillation while reserving the right to learn from customer usage and interaction data. Learning is permitted in one direction and constrained in the other, and direction is what determines where value accrues.

This is not abstract. In 2026 OpenAI began closing self-serve fine-tuning: no new organizations, and an end to new fine-tuning job creation for existing customers by January 2027. Whatever the business reasons, the practical effect for a buyer is that the ability to shape a model on your own data can be narrowed on a timeline you do not set. If your institutional knowledge only exists as adaptations inside that provider, your leverage erodes exactly as your dependence grows.

This is the test every enterprise should run, and it is the one your own teams already worry about. If any single model you rely on were taken away tomorrow, could you still operate and still hit your evals using another? Does the capability your organization built, the veteran, stay with you, or does it leave with the generalist? If the honest answer is that it leaves, you do not own your learning loop. You are renting it.

Does your business context leave your infrastructure?

Every re-query with raw organizational context is a training signal you donate. LazyFox abstracts your data at indexing, so your actual business data never reaches a model, at setup or after.

Standard AI deployment
Raw organizational context re-enters the model on every interaction
🗄️
Enterprise Data Systems
CRM, ERP, Data Lake, with real metric definitions and business logic
Raw context passed to model on every query
☁️
Model Provider API
Receives your actual business definitions and context each call
⚠ Each API call with raw context is a signal about your domain. At scale, the model learns your business. That’s not yours anymore.
LazyFox architecture
Your actual data never reaches the model. Queries are governed code, not business data.
🗄️
Enterprise Data Systems
CRM, ERP, Data Lake, with real metric definitions and business logic
Indexed and abstracted once at setup
🔒
LazyFox Semantic Layer
Governed definition graph. Versioned. Owned by you. Never leaves your infrastructure.
Governance boundary
Only governed queries (code) pass through
☁️
Model Provider API
Receives governed, abstracted queries only. Never sees your actual business data, ever.
✓ Your actual business data never reaches the model, at indexing or after. It is abstracted probabilistically first. Your definitions stay yours, and private evals run on your own definition graph.

Nadella’s defense is to keep the definitions of "good" on your side: private evals that run against your real internal outcomes, plus ownership of your memory, traces, and the right to use model outputs from your own tasks. But leakage is not the only failure that ownership has to catch. There is a quieter one.

Semantic drift. The model’s understanding of a term diverges from what the term really means, and no one notices until the outputs are wrong in ways that are hard to trace. "Revenue" in the CRM is not "revenue" in the ERP is not "revenue" in the board deck. Without a governance layer, the model reconciles the three implicitly, picking the most statistically frequent interpretation and running with it. The board deck, the CRM, and the warehouse all cite "the AI." None looks wrong. All of them conflict.

"What the technical customers want is control over their compute, their models, their data stack, and their alpha. They want to know they own the means of production, and it’s not being transferred to someone else."

Alex Karp, Palantir, quoted in Nadella’s essay

This is where LazyFox’s architecture earns its place. Definitions versioned and governed outside the model make drift measurable: which term is diverging, and in which system. When a schema change in SAP alters the data feeding a metric, LazyFox detects the divergence and re-enriches only the affected slice. The evals, the definitions, and the corrections stay yours, which is the whole point of a boundary.

Choice

One Governed Answer, Any Model

Nadella’s Choice test: if any one model were taken away, could you still operate and optimize for your evals using another? That requires decoupling meaning from the model.

The way to keep the model interchangeable is to move the layer that holds your meaning out of the model entirely. Definitions, evals, and context become governed artifacts you own, and models become engines that run against them. When that layer also reconciles across systems, the same query returns one governed answer regardless of which model is cheapest and best that week. That is Choice and Cost in Nadella’s terms: no single provider can hold your operations hostage.

Most RAG implementations retrieve from one system at a time. A user gets context from the CRM, or from the data warehouse, or from the documentation repo, but not a semantically reconciled view across all three at once. When the same concept exists in multiple systems with different definitions, the model sees the inconsistency and resolves it in-context, usually incorrectly, always expensively, and always in a way you cannot audit later.

The problem is not retrieval. The same term ("revenue," "customer," "open," "qualified") means something different in each system, and each of those differences is load-bearing. Finance recognizes revenue when the deal closes. Sales books it when the contract is signed. Both are correct for their purposes. The conflict is real and structural, and resolving it once, in a layer you own, is what lets you swap the model underneath without re-litigating what your numbers mean.

Four systems. Four definitions.
One governed truth.

LazyFox reconciles meaning across all connected systems simultaneously at runtime, sitting above the data stack without requiring migration, and resolving definitional conflicts inside the boundary before any query reaches a model.

CRM · Salesforce
"Revenue"
Contracted ARR at deal close, before invoicing or recognition
Conflicts with ERP
ERP · SAP
"Revenue"
Recognized revenue under IFRS 15, invoiced and delivered
Conflicts with CRM
Data Lake
"Revenue"
SUM(orders.amount), raw transactional, includes cancelled orders
Unvalidated source
BI Tool · Looker
"Revenue"
MRR × 12, annualized from current active subscriptions only
Forecast, not actuals
Semantic governance layer
Sits above all systems
No migration required
Reconciled at runtime
Governed definition, resolved by context
"Revenue" [context: CFO Q1 board report]
= Recognized revenue per SAP S/4HANA, IFRS 15 basis, excluding inter-company. Excludes cancelled orders and annualized forecasts. Source of truth: ERP.
↳ Version 4.2 · Approved by Finance · Last updated 2026-03-01 · 3 other definitions available by context

This is the "queryable institutional memory" Nadella describes. Not retrieval from a single system, and not retrieval from several with conflicts left in the output. The contextual layer resolves the definitional conflict inside the boundary, so what runs against any model is a governed, consistent, abstracted query: one answer, by context, versioned, auditable, and independent of which model produced it.

The Takeaway

What to Evaluate Now

Separate what you already own from what you are quietly renting. Five questions do most of the work.

  • If your model provider disappeared tomorrow, would your institutional knowledge stay? If it lives only as adaptations inside their system, the answer is no, and that is the whole risk.
  • Where do your metric definitions and evals live? In code you own and version, or inferred by a rented model from prompt history you cannot export.
  • Is your business context re-sent to a model on every query? Each call that ships raw context is a payment in knowledge; the fix is a layer that abstracts your data so the model never receives it at all.
  • Who owns the corrections your experts make? Every correction is training data. Decide deliberately whose asset it becomes before a year of usage decides for you.
  • Can you get one governed answer across systems, or three conflicting ones? Reconciliation decides whether an agent’s number holds up in a board deck or in front of an auditor.

Score these separately from model quality. A better model does not answer any of them. The layer that does is the one that decides whether you are using a model, or paying it in the knowledge that makes you unique.

The Conclusion

Own the Loop, Not the License

In the cloud era enterprises accumulated data. In the AI era they accumulate learning. The boundary has to evolve from protecting information to protecting the mechanism that learns.

Nadella is right that in the AI era enterprises accumulate learning, not just data, and that the trust boundary has to protect the mechanism through which a company learns, adapts, and compounds intelligence. If learning flows in only one direction, value converges toward the owners of the learning infrastructure rather than the creators of the knowledge. That is the trajectory the current regime is on, and it is already visible.

The part left to you is implementation. A trust boundary is only real when it is a layer you own: one that governs meaning, indexes context once, and keeps your definitions and evals as company assets while the models underneath stay interchangeable. A clause in a contract does not do this. A layer does.

"In the cloud era, enterprises accumulated data. In the AI era, they accumulate learning. The trust boundary must evolve accordingly."

Satya Nadella, CEO of Microsoft

LazyFox is built to be exactly this layer: a semantic governance layer that sits above a company’s existing data stack, encodes institutional knowledge in a form the organization owns and controls, and makes it compound with each use without ever requiring it to leave their infrastructure. Your data is indexed once and abstracted with probabilistic methods, so your actual business data never reaches a model. Every request runs from governed code. Cross-system conflicts are resolved inside the boundary before any query goes out. Semantic drift is detected before it propagates into outputs. The model executes. Your data and the meaning stay inside.

That is the difference between using a model and giving up the knowledge that makes you unique. It is also the reverse information paradox, answered: you can consume intelligence without handing over the intelligence you create.

"A company should be able to use a model without giving up the knowledge that makes it unique. That is the reverse information paradox we need to confront."
Satya Nadella, CEO of Microsoft. July 2026
Share this article
About this page

What does this post argue?

This post takes Nadella's claim that the learning loop is the IP and turns it around: if the loop that learns from proprietary knowledge is where the value accrues, an enterprise that lets AI train on its data without governed context is handing that value to the vendor, not keeping it.

A governed semantic layer is what lets an enterprise keep the benefit of that learning loop — deterministic, audited, and portable across models — instead of ceding it.

Own the Loop, Not Just the License.

LazyFox holds your definitions, evals, and context inside a boundary you control, indexed once from your existing stack, while the model underneath stays interchangeable. It connects read-only. Nothing migrates.