HOly Cow Studios

HOly Cow Studios

Research

Research

Independent research on artificial intelligence-what it is, who holds power over it, and what it does to the industries that adopt it.

Independent research on artificial intelligence-what it is, who holds power over it, and what it does to the industries that adopt it.

Research series

Research series

AI Research

AI Research

What the technology actually is. For readers who need to engage with AI as a strategic subject without a technical background.

What the technology actually is. For readers who need to engage with AI as a strategic subject without a technical background.

The three-cluster systems taxonomy (symbolic / parametric / agentic)

Our vendor says their product is AI-powered and our lawyer says it might not legally be an AI system. Who is right, and how would I tell?

Twenty-eight published definitions of AI, coded against eight themes, show the three sectors emphasise systematically different things: 75% of industry definitions invoke human comparison against 14% of policy ones, which foreground autonomy and data-dependence instead. Statistical clustering of 27 real systems built between 1956 and 2025 finds three groups — and the smallest, four high-autonomy agentic systems, is defined by acting in the world rather than by scale or recency.

Read research →

The three-cluster systems taxonomy (symbolic / parametric / agentic)

Making Sense of Artificial Intelligence

Our vendor says their product is AI-powered and our lawyer says it might not legally be an AI system. Who is right, and how would I tell?

Twenty-eight published definitions of AI, coded against eight themes, show the three sectors emphasise systematically different things: 75% of industry definitions invoke human comparison against 14% of policy ones, which foreground autonomy and data-dependence instead. Statistical clustering of 27 real systems built between 1956 and 2025 finds three groups — and the smallest, four high-autonomy agentic systems, is defined by acting in the world rather than by scale or recency.

Read research →

Research series

Research series

AI Order Research

AI Order Research

How power and ownership are being redistributed around the technology. Board-grade strategic analysis for policy and executive readerships.

How power and ownership are being redistributed around the technology. Board-grade strategic analysis for policy and executive readerships.

AI National Power Index (AINPI) · The chokepoint paradox

If the United States has spent twenty-three times what China has, why is the gap between their best models under three percentage points -and what does that mean for where we place our own effort?

Fourteen actors scored across ten dimensions of AI power. The United States leads at 88.1, China at 72.7, and nine further actors sit within nine points of each other. But the ranking hides the leverage: 74.5% of tracked global AI compute sits in the US, and 100% of extreme-ultraviolet lithography comes from one Dutch firm neither superpower owns. The paper argues the frontier leaderboard has stopped measuring power, and that chokepoints, verification and standards are where strategy belongs.

Read research →

AI National Power Index (AINPI) · The chokepoint paradox

Techno-Polarity in the Twenty-First Century

If the United States has spent twenty-three times what China has, why is the gap between their best models under three percentage points -and what does that mean for where we place our own effort?

Fourteen actors scored across ten dimensions of AI power. The United States leads at 88.1, China at 72.7, and nine further actors sit within nine points of each other. But the ranking hides the leverage: 74.5% of tracked global AI compute sits in the US, and 100% of extreme-ultraviolet lithography comes from one Dutch firm neither superpower owns. The paper argues the frontier leaderboard has stopped measuring power, and that chokepoints, verification and standards are where strategy belongs.

Read research →

Who Owns the Intelligence? (ten questions) · The four ownership forms · Five Futures (Models A–E)

Four US firms will spend $700 billion on AI infrastructure this year. If we are renting our intelligence from them, what exactly do we own — and what happens to the value our people and our data produce?

Four US hyperscalers will spend ~$700bn on AI infrastructure in 2026, against $1.25bn for India's entire national AI mission — a 560-fold gap. The IMF finds capital and wealth inequality "always increase" under AI, with high earners gaining seven times what low earners do. Yet the US, EU and India are producing three visibly different economies from the same models. The paper argues ownership, not capability, decides which economy emerges, and scores five candidate systems against eight criteria.

Read research →

Who Owns the Intelligence? (ten questions) · The four ownership forms · Five Futures (Models A–E)

From Platform Capitalism to What?

Four US firms will spend $700 billion on AI infrastructure this year. If we are renting our intelligence from them, what exactly do we own — and what happens to the value our people and our data produce?

Four US hyperscalers will spend ~$700bn on AI infrastructure in 2026, against $1.25bn for India's entire national AI mission — a 560-fold gap. The IMF finds capital and wealth inequality "always increase" under AI, with high earners gaining seven times what low earners do. Yet the US, EU and India are producing three visibly different economies from the same models. The paper argues ownership, not capability, decides which economy emerges, and scores five candidate systems against eight criteria.

Read research →

Research series

Research series

Luxury Hospitality Research

Luxury Hospitality Research

Research on the intelligent future of luxury hospitality.

Research on the intelligent future of luxury hospitality.

The Elegance Threshold

Where should a luxury hotel deploy AI first — and where should it never deploy it at all?

Orchestrating human judgement and artificial intelligence for sustainable competitive advantage in luxury hospitality. Prepared for boards, owners and senior operating executives. Its argument: AI creates the most value, and the least risk, when deployed first against the parts of luxury hospitality guests never see.

Read research →

The Elegance Threshold

Nobody Tips a Chatbot

Where should a luxury hotel deploy AI first — and where should it never deploy it at all?

Orchestrating human judgement and artificial intelligence for sustainable competitive advantage in luxury hospitality. Prepared for boards, owners and senior operating executives. Its argument: AI creates the most value, and the least risk, when deployed first against the parts of luxury hospitality guests never see.

Read research →

AI Visibility Matrix · 5E Elegance Chain

Where, specifically, should artificial intelligence sit in a luxury guest journey — and where should it not?

An evidence-based landscape of 22 real, named AI deployments across Four Seasons, Ritz-Carlton, Hilton, Marriott, Rosewood, Aman, Mandarin Oriental, Accor, IHG and Hyatt — each coded on two questions: how much can it decide on its own, and how much does the guest notice it? Those two axes sort the successes from the failures better than anything about the technology itself, and they produce a decision tool that fits on a napkin.

Read research →

AI Visibility Matrix · 5E Elegance Chain

Intelligent Elegance

Where, specifically, should artificial intelligence sit in a luxury guest journey — and where should it not?

An evidence-based landscape of 22 real, named AI deployments across Four Seasons, Ritz-Carlton, Hilton, Marriott, Rosewood, Aman, Mandarin Oriental, Accor, IHG and Hyatt — each coded on two questions: how much can it decide on its own, and how much does the guest notice it? Those two axes sort the successes from the failures better than anything about the technology itself, and they produce a decision tool that fits on a napkin.

Read research →

16-category AI taxonomy · Two reference stacks · KPI framework · Five-stage AI Maturity Model (canonical)

What does an AI stack for a luxury hotel actually consist of, what does it truly cost, and how do you know how far along you are?

Sixteen categories of AI, hundreds of vendors, and no shared reference architecture. This paper builds the taxonomy, prices it, and sets out the five-stage maturity model this series treats as canonical. Three findings carry the argument: vendor list prices understate true first-year cost by 40 to 60 per cent; operators routinely place themselves one to two stages above where an independent assessment would; and the workforce, not the budget, is the binding constraint.

Read research →

16-category AI taxonomy · Two reference stacks · KPI framework · Five-stage AI Maturity Model (canonical)

The Stack Behind The Stay

What does an AI stack for a luxury hotel actually consist of, what does it truly cost, and how do you know how far along you are?

Sixteen categories of AI, hundreds of vendors, and no shared reference architecture. This paper builds the taxonomy, prices it, and sets out the five-stage maturity model this series treats as canonical. Three findings carry the argument: vendor list prices understate true first-year cost by 40 to 60 per cent; operators routinely place themselves one to two stages above where an independent assessment would; and the workforce, not the budget, is the binding constraint.

Read research →