Structural Analysis 12

The Lord Class's Self-Destruction — The Limits of Big Tech's Enclosure

Nadella could have admitted Copilot failed and moved on. He kept running anyway. That locks in the villain role.

How the Lord Class Responds to New Technology

Historically, whenever a new technology threatens an existing class structure, the lord class has followed a predictable script.

Medieval Lords vs. the Printing Press Big Tech vs. the AI Revolution
Licensing and permits for printed materials AI regulation (with exemptions for their own products)
Censorship of books API rate limits, content policies
Universities placed under church authority AI research absorbed in-house (failed Mistral acquisition, Inflection absorbed)
Inquisitions Lawsuits against former employees, enforced non-competes
Monopoly over pilgrimage routes App Store / Play Store monopolies
Toll taxation API fees, platform commissions
Forced corvée labor Mandatory Microsoft accounts, forced Copilot integration
Church legitimizing theology "Responsible AI" / "Safety" rhetoric

Every move is the classic lord-class playbook for not releasing serfs and the middle class. Historically, this resistance ultimately failed — feudalism was dismantled. There is structural reason to expect the same outcome this time.

But in the course of that resistance, the lord class engages in self-destructive behavior. That self-destruction is now proceeding at global scale.

Nadella's Seven Acts

The clearest embodiment of Big Tech's response to the AI revolution is Microsoft's Satya Nadella. Arranged chronologically, his key acts are:

Nadella's Seven Self-Destructive Acts:
1. Massive investment in OpenAI (2019–2024, cumulative $13B+)
The ignition device for the entire AI bubble. Without this, the AI arms race may never have started.
2. February 2023 — Bing + ChatGPT integration announced
Triggered "Code Red" inside Google; forced the entire industry into AI arms-race mode.
3. November 2023 — Altman reinstatement operation
Immediately after OpenAI's board removed him over safety concerns, Microsoft announced it would hire him external pressure forced his reinstatement.
Destroyed the AI industry's only safety-side governance mechanism.
4. Stargate $500B (2025)
All hyperscalers were forced to follow; concentrated the world's semiconductors, power, and water into AI.
5. Forced Copilot bundling
Systematic imposition into Word, Excel, PowerPoint, Outlook, Teams, Windows, and GitHub.
6. Copilot+ PC (NPU 40 TOPS requirement)
Forces hardware replacement on the existing 3 billion Windows users.
7. Take-or-pay contracts (Kenya, etc.)
Transfers enormous risk onto developing countries under a contract structure that bears no accountability on failure.

Every one of these was Nadella's personal decision. Microsoft holds $750B in cash and draws stable revenue from Office, Windows, GitHub, and LinkedIn. Admitting "Copilot was a failure" would not have sunk the company. He continued anyway.

That is the structural basis for designating him an "intentional villain."

The Kenya Case — Evidence He Did Not Retreat

The clearest evidence that Nadella "chose not to exit when exit was available" came from the Kenya data-center deal.

Microsoft / G42 Kenya Project:
Project scale 1 GW (1,000 MW) geothermal data center
Kenya's total generation capacity 3,000–3,200 MW
Share of national grid roughly one-third of the entire country's power
President Ruto's public statement "Running that one data center would require cutting power to half the country"
Contract demand Take-or-pay (Kenya pays even if it doesn't use the capacity — hundreds of millions of dollars per year)
Kenya's compromise offer Phase in, starting at 200 MW
Microsoft's response Rejected and withdrew ("Partial guarantees make project financing impossible")

This is a case where Microsoft insisted on a contract that transferred all failure risk away from itself, then walked away when its terms were refused. It demanded the right to lease an entire developing country's power grid wholesale; when that was denied, it left.

Yet Kenya was just one country. The same playbook is running simultaneously in Indonesia, Malaysia, the UAE, Saudi Arabia, and elsewhere. A designed extraction structure is being pressed simultaneously across the globe.

Recall Was Not an Accident — It Was Inevitable

Microsoft Recall — the feature that takes screenshots every few seconds and indexes them with AI — was criticized worldwide. It is often read as "Microsoft overreached and made a mistake."

Structurally, however, once AI is placed inside a fused stack, this outcome becomes inevitable:

The structure that makes Recall inevitable:
Windows' stack is fused (kernel + shell + AI + apps are inseparable)
Every layer can observe every other
Privileged processes hold the right to see the entire screen from the start
AI embedded in the OS a channel for passing screen data to AI is created
Microsoft account and cloud are already wired together no guarantee data stays local
Recall = fused stack × AI, a logical conclusion

Building the same kind of feature on Linux would be technically difficult — Wayland structurally restricts cross-application snooping, the kernel has no awareness of the screen, and the culture keeps AI external, so no single process ever starts with authority to aggregate that information. The impossibility is a consequence of design, not philosophy.

Recall became inevitable the moment Microsoft decided "our stack maintains a fused architecture." Even if Recall is disabled, as long as the stack remains fused, another isomorphic feature will necessarily emerge.

Why Retreat Is Impossible — Structural Lock-In

With this much self-destruction visible, why can Nadella not retreat?

Barrier Content
Irreversibility of CAPEX Data center construction runs on 2–3 year cycles; stopping means the project still completes
Promises to customers Got them to buy Copilot+ PCs and sign Azure contracts — reversal means loss of trust
Promises to shareholders Said "AI is the next Azure"; retreating collapses market capitalization instantly
Competitive structure OpenAI, Anthropic, and Google keep running; Microsoft alone stopping changes nothing
Personal accountability Admitting every major decision to date was a failure

He lit the fire himself, then moved the extinguisher far away. Charles Mitchell — New York banker of the 1920s and symbol of the Great Depression — left his post after the collapse and ended his career in a tax-evasion trial. He had the option to exit as an individual.

Nadella does not. He had the capacity to exit (Microsoft is flush with cash), but he dismantled the exit structure himself.

The Collision of the AI Bubble and the Resource Shock

Here is the core of the historical context. Nadella's self-destruction collides with other historical forces and becomes irreversible:

Converging dynamics:
AI CAPEX Concentrated consumption of power, water, copper, lithium, semiconductors
simultaneously
Oil shock Iran tensions, Hormuz crisis, crude oil at $98–$200 scenarios
simultaneously
Entry-level hardware scarcity Memory and SSD prices soaring, "16 GB / 500 GB" machines becoming unaffordable
simultaneously
Food and fuel inflation Rising cost of living for ordinary households
simultaneously
Microsoft Still selling "next-generation AI PC, ¥200,000+"

The scene of "selling cake during a famine" materializes

When that happens:

  • The justification for AI infrastructure investment collapses ("AI grows the economy" is rewritten as "AI is inflating food and fuel costs")
  • Power and water allocation become a political issue (data centers vs. housing, agriculture, transportation)
  • Enclosure strategy faces headwinds (selling ¥200,000 Copilot+ PCs alongside people who cannot afford food)
  • Nadella's statements will later be cited as "the moment that person said exactly that"

Structurally, Nadella occupies the same position as Charles Mitchell saying "markets are healthy" in October 1929.

Big Tech's Self-Destruction Spans the Entire Hierarchy

Nadella is not alone. All of Big Tech has entered the same pattern:

Figure Nature of Acts Retreat Possibility
Nadella Design and systematic execution High (chose not to)
Musk Flashy execution, unstable High (retreats occasionally)
Zuckerberg Algorithmic harm + political realignment Medium (depleted by metaverse failure)
Altman Driving force but subordinate Low (Microsoft-dependent → working to break free)
Bezos Intervention from a stepped-back position High (effectively retired)
Pichai Reactive, narrow in vision Medium (follower)
Cook Closer to passive victim Low (paralyzed by China dependence)

Yet Nadella stands out for seven structural reasons:

  1. Origin of causality — everyone else is living in the wake of his first move
  2. Scale difference — annual AI-focused CAPEX of $80B+
  3. Systematic execution — unglamorous but methodical, unlike Musk's flamboyant whims
  4. Intentional destruction of safety mechanisms — the Altman reinstatement operation is something no one else did
  5. Customer captivity — a difficult-to-escape crowd of Windows 1.5B + Office 400M
  6. Highest retreat potential, yet did not retreat — Microsoft had the greatest capacity to exit
  7. "Bowling CEO" Musk strategy — while Musk absorbs criticism, Nadella quietly advances the maximum destruction

The last point matters most. The greatest harm, yet the least scrutiny — this is the worst combination. Musk is attacked and faces corrective pressure; Trump faces impeachment and elections. But who besides shareholders can stop Nadella? Those shareholders are prospering from AI stock prices and have no motive to intervene.

The Category of "Quiet Maximum Perpetrator"

Historically, the greatest perpetrators are often not the flamboyant dictators:

  • I.G. Farben management in 1930s Germany (an unglamorous chemical company, the largest-scale wartime collaborator)
  • Robert McNamara in the 1960s Vietnam War (rational technocrat, systematized mass killing)
  • Stan O'Neal in the 2000s Lehman collapse (unassuming investment banker, drove subprime securitization)

Flashy figures attract media attention, but it is the quiet technocrat type that builds the structure. Nadella fits this lineage precisely. While Trump monopolizes media attention, Nadella is rewriting the architecture of the physical world.

The Relationship Between the Lord Class and the Emerging Class — The Twin Nature of Trump and Big Tech

The relationship between Big Tech lords and Trump can also be made structurally explicit:

The structure that produced Trump:
Big Tech optimizes for engagement economics
Polarization manufactured at scale (anger, fear, friend/enemy sorting)
Local news exterminated (Google + Facebook absorbed advertising)
An information environment where Trump-type politicians have selective advantage
Trump rises
2024: Big Tech publicly endorses Trump (Musk, Zuckerberg, Bezos, major tech CEOs in the front row at the inauguration)
Big Tech created Trump, and Big Tech exploits Trump politically

Yet both parties can only do their respective "narrow jobs":

  • Big Tech relied on Trump to destroy regulation → achieved (FTC, AI regulation rollbacks)
  • Simultaneously, the accompanying trade war, immigration restrictions, and Iran war destroy Big Tech's own supply chains
  • "Narrow transaction" transformed into "broad subordination"

Cook making personal pilgrimages to Mar-a-Lago to beg for tariff exemptions, Musk clashing with Vance over H-1B, Nadella quietly keeping his distance — these are all symptoms of a coalition where both sides misread the scope of their own capabilities.

Chapter Conclusion — Inability to Retreat Confirms the Villain Role

Nadella could have retreated. He did not. Microsoft was financially more than capable of retreating. He still carried out seven self-destructive acts.

This is the factor that decisively changes the outcome. Choosing not to retreat structurally confirms the villain role. In historical memory, the depth of the record will likely exceed Mitchell's.

Nadella's weight of responsibility lies in the fact that he did not exit when exit was available.
The seven self-destructive acts — massive OpenAI investment, the Altman reinstatement operation, Stargate, forced Copilot bundling, hardware requirements, take-or-pay contracts, refusal to retreat — were all deliberate management decisions.
Recall was not an accident; it was the logical conclusion of embedding AI in a fused stack.
Just as medieval lords could not suppress the printing press, Big Tech cannot suppress the AI revolution.
The collision of the AI bubble and the resource shock makes this self-destruction irreversible.

The next chapter examines another movement running in parallel with the lord class's self-destruction. We analyze how the dynamics of the AI revolution are dismantling this feudalism — and how a new class, the builder, is rising to take its place.

← Prev: The Contradictions Modern Feudalism Produced — Fakes, Attention Extraction, Mythos, and the Material Blind Spot Next: AI and the Individual →

We are watching the lord class destroy itself in real time.

Just as medieval lords could not suppress the printing press, Big Tech cannot suppress the AI revolution. Unable to retreat, they will be recorded in history as the villains.

AISeed — 生物多様性・食料・AIと暮らし(Facebook)