Skip to content
OpenAI · ChatGPT · GeminiEpisode 094 · 25 January 2026 · 51:38

1.9 GW for OpenAI: AI Companies Are Beginning to Resemble States in Energy Use and Influence

What to watch for

1Compare “1.9 GW for OpenAI: AI Companies Are Beginning to Resemble States in Energy Use and Influence” with “XAI and Colossus 2 Supercomputer (1 GW)”: they provide different criteria for judging the same issue.
2Test the conclusion from “Bandcamp prohibits AI music” in your own use case—what actually changes in the process and what remains a promise.
3Before choosing a product or approach, record the constraint identified in “The British wants to ban social media until 16 years”.
4Define the owner of the outcome and the quality metric for the situation described in “1.9 GW OpenAI load (energy as a country)”.
Signals to track afterwards
Watch for actions by Bandcamp and Anthropic that confirm or challenge the episode’s central claims.
Compare new launches and policy changes with “Bandcamp prohibits AI music”: have access, quality, price, or constraints changed?
Check whether the scenario in “1.9 GW OpenAI load (energy as a country)” becomes repeatable practice rather than a one-off demonstration.
Most useful for
EntrepreneursLegal professionalsProduct leadersAI usersInvestorsStrategy teams

Key takeaways

00:00OpenAI’s 1.9-gigawatt load shows that modern AI is not merely code in the cloud

The “1.9 GW for OpenAI: AI Companies Are Beginning to Resemble States in Energy Use and Influence” scene leads to a working conclusion: this section clarifies the mechanism behind the topic and preserves a constraint that would otherwise be lost in an overly simple conclusion.

03:20Why an announcement is not enough: robot 1X NEO: autonomy instead of telecontrol

The “Robot 1X NEO: autonomy instead of telecontrol” scene leads to a working conclusion: a benchmark measures a narrow capability; working value requires repeatability, a clear cost, and control over errors.

07:14The market tests it through use: risks of privacy of Gemini personal intelligence

The discussion of “Risks of privacy of Gemini personal intelligence” yields a practical test: a benchmark measures a narrow capability; working value requires repeatability, a clear cost, and control over errors.

14:06The boundary between value and constraint: new ChatGPT Go $8

The “New ChatGPT Go $8” topic becomes clearer once this point is included: a launch matters only when it changes access, quality, price, or user behavior in a real workflow.

16:44Who owns the outcome: 1.9 GW OpenAI load (energy as a country)

The practical meaning of “1.9 GW OpenAI load (energy as a country)” is that a benchmark measures a narrow capability; working value requires repeatability, a clear cost, and control over errors.

21:00What changes in real work: anthropic and Teach For All: training of educators

The working conclusion from “Anthropic and Teach For All: training of educators” is that the important signal is not one funding number: the next round, available runway, and closure rate show whether a company can survive the new cost of capital.

22:44Why context matters more than one metric: big Tech Olympic power

For the “Big Tech Olympic power” scene, the decisive point is this: the important signal is not one funding number: the next round, available runway, and closure rate show whether a company can survive the new cost of capital.

24:17The eight-dollar ChatGPT Go plan shows the other side of scale: companies want to expand the audience in countries where Plus is too expensive

The discussion of “XAI and Colossus 2 Supercomputer (1 GW)” yields a practical test: this section clarifies the mechanism behind the topic and preserves a constraint that would otherwise be lost in an overly simple conclusion.

32:04Anthropic works with Teach For All because a model’s influence is not determined by infrastructure alone

The boundary of the “Bandcamp prohibits AI music” case is defined by this point: teachers will decide how AI is used in the classroom and what children learn. Bandcamp, in contrast, bans AI music to protect human authors and the culture of the platform. These are two different attempts to set a boundary.

50:10Deepfakes and UK plans to restrict social media for people under sixteen show that AI companies already participate in governing society

The “The British wants to ban social media until 16 years” topic becomes clearer once this point is included: this section clarifies the mechanism behind the topic and preserves a constraint that would otherwise be lost in an overly simple conclusion.

What this episode is about

OpenAI consumes power on the scale of major infrastructure, xAI is building Colossus 2, Anthropic trains educators, Bandcamp bans AI music, and countries discuss social-media bans for children. Models can no longer be treated as ordinary software: they require resources and create rules for society.

OpenAI’s 1.9-gigawatt load shows that modern AI is not merely code in the cloud. That scale requires power plants, grids, land, cooling, and long-term agreements. The company begins to influence regional infrastructure in the same way as a major industrial business.

xAI is building Colossus 2 with roughly the same logic. More compute means models can be trained and served more quickly. Energy does not guarantee a good product, however. An enormous supercomputer provides the right to participate in the race, not an automatic victory.

The eight-dollar ChatGPT Go plan shows the other side of scale: companies want to expand the audience in countries where Plus is too expensive. Millions of new users increase load while creating data, habit, and future revenue.

Anthropic works with Teach For All because a model’s influence is not determined by infrastructure alone. Teachers will decide how AI is used in the classroom and what children learn. Bandcamp, in contrast, bans AI music to protect human authors and the culture of the platform. These are two different attempts to set a boundary.

Deepfakes and UK plans to restrict social media for people under sixteen show that AI companies already participate in governing society. They determine what can be created, who gets access, and how age is verified.

When a private platform consumes a country’s energy and influences the information seen by millions, “Who is in charge?” cannot be answered by market capitalization alone. Rules, transparency, and accountability have to match the scale.

When a private platform consumes a country’s energy and influences the information seen by millions, “Who is in charge?” will not be answered by market capitalization alone. As a result, rules, transparency, and accountability have to match the scale.

Episode transcript

The episode is in Russian; below is an English reading guide to the transcript (the full EN transcript is a machine translation). Voice matching applied to 125 segments: 78 identified, 4 mixed, 34 marked with ✓, and 9 unresolved.

Loading…