Dear CEO – You are paying an intelligence tax for a moat that is evaporating

Dear CEO – You are paying an intelligence tax for a moat that is evaporating

Aug 24, 2026

The intelligence tax is crashing.

The cost of intelligence is decoupling from the value of the model provider.

Nvidia is moving from being a pure supplier to a direct competitor.

Their six billion dollar deal to license Poolside technology is a strategic strike against the proprietary dominance of OpenAI and Anthropic.

By absorbing the talent from Poolside into its Nemotron project, Nvidia is building an open weight ecosystem that aims to rival the most advanced frontier models.

This move follows a moment of extreme vulnerability for Poolside, when they faced a narrow window to secure a 40,000 GB300 cluster to maintain their operations.

The importance of this shift for the C suite cannot be overstated.

We are moving from an era of AI as a service to an era of AI as a localized asset.

When you rely on a closed proprietary API, you are essentially renting your intelligence.

You are paying a premium for a black box that you do not control and cannot host yourself.

As open weight models close the capability gap, that premium becomes a wasted cost.

Some argue that frontier models still lead in reasoning and specialized integration.

This is true today. But the gap is narrowing.

1. Audit your current API dependencies to identify high cost intelligence.

2. Test open weight models against your specific data to find the true performance gap.

3. Build a sovereignty roadmap that prioritizes hosting models on your own infrastructure.

Is your AI strategy a subscription or an asset?

#AIStrategy #DataGovernance #Nvidia #EnterpriseAI #AIsovereignty

Dear CEO – The data center warehouse defense is a lie that invites local hostility

Dear CEO – The data center warehouse defense is a lie that invites local hostility

Aug 20, 2026

You cannot build a digital empire on a foundation of physical resentment.

Alberta's technology minister recently faced jeers from citizens at a town hall regarding new AI data centers.

The government's defense was that these facilities are merely warehouses with computers.

This logic is a strategic error.

Treating massive infrastructure as a weightless digital asset ignores the energy, water, and land footprints that local communities actually care about, creating a rift between corporate strategy and community survival.

While Ontario is moving toward strict sustainability guidelines for deployment, Alberta is taking a free for all approach.

This creates a volatile operational environment.

Some argue that efficiency improvements in cooling and power might shrink the impact.

But efficiency does not eliminate the footprint.

The province must act to ensure stability:

1. Mandate energy and water neutrality for all new sites.

2. Provide resource guidelines for municipalities.

3. Codify public input processes into the permitting cycle.

Is your Government planning for digital growth or for physical stability?

#DataGovernance #AIInfrastructure #EnterpriseRisk #Sustainability

Dear CEO – High accuracy does not guarantee high value

Dear CEO – High accuracy does not guarantee high value

Aug 19, 2026

An elite machine learning model that nobody uses is just an expensive math experiment.

The focus must shift from technical excellence to business impact.

Leaders often mistake model sophistication for strategic readiness.

The real tension exists between the data scientist who wants accuracy and the person responsible for the P&L who needs to explain every decision to a customer.

I saw this when we built an AR prediction model that identified overdue invoices within 30 days with 88 percent accuracy (iteration 1).

Despite the high success rate, the business unit leaders simply did not believe the machine.

They distrusted the numbers because the math was a black box they could not explain to their clients.

Here is an action plan:

Prioritize explainability so your managers actually trust the output.

Measure success by business outcomes.

When was the last time an AI tool failed not because the math was wrong, but because your people refused to use it?

#AIStrategy #DigitalTransformation #Leadership #EnterpriseAI