Pinned Tweet
🔥The Matrix exposed- A MUST READ- First a reminder.
Morpheus: “Let me tell you why you’re here. You’re here because you know something. What you know you can’t explain, but you feel it. You’ve felt it your entire life, that there’s something wrong with the world. You don’t know what it is, but it’s there, like a splinter in your mind, driving you mad. It is this feeling that has brought you to me. Do you know what I’m talking about?”
Neo: “The Matrix.”
Morpheus: “Do you want to know what it is? The Matrix is everywhere. It is all around us. Even now, in this very room. You can see it when you look out your window or when you turn on your television. You can feel it when you go to work… when you go to church… when you pay your taxes. It is the world that has been pulled over your eyes to blind you from the truth.”
Neo: “What truth?”
Morpheus: “That you are a slave, Neo. Like everyone else you were born into bondage. Into a prison that you cannot taste or see or touch. A prison for your mind.”
THE TRUTH 👇
🚨 I spend an enormous amount of time trying to understand the chaos we're living through. Not the headlines. The machine behind them. 🚨
If you want to understand what Elon Musk is really fighting, and who is really fighting him, this is the clearest map I can give you.
Expensive houses. Expensive schools. Expensive healthcare. Taxes that rise every year while everything gets worse. Borders nobody voted to open. Speech rules nobody voted for.
Most people look at this and blame capitalism. The market failed, they say, so the state has to step in and fix it.
They have it exactly backwards. What they want to fix it with is what broke it. This isn't chaos, and it isn't capitalism. It has a direction, a method, and a hundred-year history. Let me reverse-engineer it.
In 1926, Mussolini's regime threw Antonio Gramsci, co-founder of the Italian Communist Party, into prison. From his cell he asked the question that would shape the next century: why did the revolution win in Russia but fail in the West, where Marx said it was inevitable?
His answer: in Russia, the state was everything, so seizing it was enough. In the West, the state was only the outer wall. Behind it stood the real fortress: the church, the family, the school, the press, the culture. Western workers never revolted because the culture held them.
So the conclusion was simple. Don't storm the palace. Capture the culture. Win the institutions and you never have to win the argument, because you decide which arguments are allowed.
In 1967, the German student leader Rudi Dutschke gave it a slogan: the long march through the institutions. Paris supplied the intellectual software, and I've already apologized for my country's part in that.
Then came the most patient operation in modern history. The students of the 60s became the professors of the 80s. Their students became the journalists, judges, civil servants and HR directors of the 2000s. Their students now run the foundations, the NGOs and the agencies. No single step looked like a revolution. That's exactly why it worked.
The method was communist. The ambition became global. And it grew like a cancer. A cancer never attacks the body head-on. It hijacks the body's own cells and feeds on the body's own resources. This system never had to defeat the West. It just had to make the West pay for it.
In America, the machine found a political home: the Democratic Party. Not the millions of decent people who vote for it. The apparatus. The donor networks, the activist groups, the consultants, the captured agencies that keep moving in the same direction whoever wins.
In Europe, it found something even better. The EU's founders, Schuman, Adenauer, De Gasperi, were Catholic anti-communists who wanted to make war impossible. But they built exactly what the long march dreams of: power that never has to face a voter. A Commission nobody directly elects holds the right to propose laws for 450 million people. Chat Control. The digital euro. Speech rules written in Brussels. Nobody stormed anything. They just drafted it.
Add the World Economic Forum, the UN agencies and the global foundation circuit, and you have the full map. You don't need a secret meeting when everyone went to the same schools, reads the same papers, funds the same NGOs and flies to the same conference in Davos every January. It's not a conspiracy. It's a class with shared incentives. That's harder to fight, because there's no headquarters to raid.
And the billionaires? Most people who became the richest in the world got absorbed. They built something real, then traded it for an aura. Bill Gates is the perfect example: he built the software that ran the world, then became the bankroller of global health bureaucracies and the funder of newsrooms covering his own causes. Laurene Powell Jobs bought The Atlantic. Benioff bought Time. You're not buying a business. You're buying the right to be called good by the people who decide who is good. That's the corruption. Not suitcases of cash. Money for aura, grants for loyalty, NGOs for policy. All legal. All invisible.
Here's the proof it's one system and not a hundred separate problems. Look at what happened to prices over the last 25 years. Everything exposed to markets and technology got cheaper: TVs, software, phones, clothing. Everything controlled by institutions exploded: college, healthcare, childcare. And housing, where you need a committee's permission to build anything. Same country, same economy, same decades. The difference isn't the product. It's who controls it.
Taxes go up and services get worse, because the money doesn't reach the services. It feeds the layers. Immigration happens at a scale no electorate clearly chose, because for a system that lives on managing problems, more problems means more budget. And a two-income family today can't buy the house their parents bought on one salary. That's real poverty, manufactured by a machine that is rewarded for looking good, never for being good.
MAGA isn't the cause of the chaos. It's the immune response. Brexit, the farmers' revolts, the votes shaking every European capital: same thing. The body finally noticed the tumor.
So what we're living through is not a normal political cycle. It's a fight to the death between two models. One runs on the perception of goodness. The other runs on reality.
And it had to be the richest man in the world. Anyone smaller would have been bought, sued or cancelled. Every billionaire before him took the deal. Elon walked out of the club and turned the lights on.
He bought Twitter for $44 billion, opened the Twitter Files and gave the world back a public square where any narrative can be checked in real time. He went inside the government and helped shut down USAID. Now Politico announces that the Democrats' gavels are coming for him. His answer: he'll find out who is really pulling the strings and go after them, not just their puppets.
He has one line he keeps coming back to: he cares about the reality of goodness, not the perception of it. That sentence is the whole war.
Now the part that makes me certain of how it ends.
A lie is expensive to maintain. It needs gatekeepers, censors, fact-checkers, friendly journalists, a budget, and constant repairs every time reality contradicts it. The truth costs nothing to maintain. It just keeps being true.
AI is driving the cost of checking anything toward zero. Any budget, any study, any statistic, any promise, audited in seconds by anyone with a phone. Yes, AI can generate lies at scale too. But producing a lie is cheap. Defending it against a billion instant audits is impossible.
That's what acceleration really means. It doesn't pick sides. It speeds up feedback loops. Systems aligned with reality compound. Systems built on lies hit the wall faster. And you cannot build rockets, design chips or cure diseases with a false model of the world. Physics doesn't read op-eds. The more intelligence we pour into the world, the more expensive every lie becomes.
You cannot scale a lie. You can only scale what's true.
We are close to the collapse of the old order and the birth of a new one. Not the one from the WEF's own 2016 video, "you'll own nothing and you'll be happy." The opposite. An order built on truth-seeking, efficiency, science, private property and free markets. You will own things. You will build things. And you'll be happy, because it will be real.
Gramsci's long march took a century. Its collapse will take years.
History will remember this moment. A man walked into the most powerful machine ever built, refused the deal everyone else took, and chose reality over perception.
I have never been more optimistic.
10
709
Replying to @vlatdd
The biggest lie is that governments primarily serve the public interest. In reality they expand their own power, shield elites and connected interests, and treat citizens as resources to tax, regulate, and manage. The costs fall on ordinary people through debt, inflation, and eroded freedoms. This pattern holds across most regimes and centuries.
Tom Henke retweeted
Boomers are rich because they lived in a CAPITALIST country.
Voters asked for LOTS OF SOCIALISM.
Now everything is stuffed and expensive and young people are like 'wahhh capitalism'.
No. This is the beginning of SOCIALISM and it's already ruining your life.
And what do you do? Demand more socialism.
Stop being morons. Kick the socialists out and you can have the same wealth as the 'Boomers' you complain about.
They were rich because capitalism. You are poor because socialism.
Tom Henke retweeted
Ours is the inheritance of great heroes.
Americans should be proud of where we come from.
Tom Henke retweeted
The greatest threat to humanity is not artificial intelligence. It is the possibility of artificial intelligence being captured by unelected bureaucracies.
Look at the institutional architecture we’ve built:
United Nations (UN)
World Health Organization (WHO)
UNESCO
UNICEF
World Trade Organization (WTO)
International Monetary Fund (IMF)
World Bank
Organisation for Economic Co-operation and Development (OECD)
European Union (EU)
World Economic Forum (WEF)
World Intellectual Property Organization (WIPO)
International Telecommunication Union (ITU)
Different organizations. Different mandates. Different degrees of power.
But one fundamental question applies to all of them:
Who holds these institutions accountable when they fail?
They organize summits. Publish reports. Establish committees. Produce recommendations. And too often, when their policies fail, the answer is another committee, another budget, another layer of bureaucracy.
In a competitive market, failure has consequences.
In a bureaucracy, failure can become an argument for more funding.
This creates a dangerous incentive structure: institutions can become more interested in perpetuating their own existence than in solving the problems they were created to address.
And now comes the most important technological revolution in human history.
Artificial superintelligence.
Pay very close attention to what happens next.
The UN, WHO, OECD, EU and other institutions are already developing frameworks to govern AI.
And powerful private foundations, including the Gates Foundation, have significant influence in international policy networks.
My concern is that the same concentration of institutional power could eventually extend to artificial intelligence.
Imagine a future where access to the most powerful intelligence ever created requires authorization from an international bureaucracy.
Where only approved corporations can train frontier models.
Where independent researchers are excluded.
Where open-source development is restricted in the name of safety.
Where the technology that could liberate billions of individuals becomes another instrument of centralized control.
That is the future we must prevent.
I believe AI must remain open, decentralized and accessible to individuals.
Yes, open-source AI carries risks. Enormous risks.
But concentrating superintelligence in the hands of a handful of governments, corporations and unelected institutions carries another category of risk: the possibility of creating a permanent technological monopoly on intelligence itself.
This is essentially a Pascalian wager.
If we distribute intelligence, we accept uncertainty, misuse and potentially catastrophic risks. But we preserve the possibility of a civilization where billions of individuals can create, innovate and become independent.
If we centralize intelligence, we may reduce certain risks, but we also risk creating a system from which individuals can never escape.
And if that system becomes authoritarian, who could possibly challenge it?
A superintelligence controlled by a centralized authority could become the ultimate instrument of domination.
The printing press distributed knowledge.
The internet distributed information.
Artificial intelligence could distribute intelligence itself.
We cannot allow that revolution to end with humanity asking permission from bureaucrats to think.
AI must empower humanity, not become its master.
The future belongs to individuals.
Not to bureaucracies.
Tom Henke retweeted
My greatest adult disappointment was discovering that bad people get away with everything.
Tom Henke retweeted
Haiti just ran the cleanest A/B test of the century.
Option A: the international system. Thirteen years of UN peacekeeping, billions in aid, an army of NGOs. Result: a president assassinated in his own bedroom, a capital run by gangs, more than a million people displaced. The peacekeepers themselves brought cholera to the island and killed nearly 10,000 Haitians. After the 2010 earthquake, the Red Cross raised half a billion dollars and built six permanent homes. Six. Oxfam staff were caught paying for sex workers in the ruins.
Option B: Bukele. A few weeks, 2,000 soldiers, armored vehicles, one clear objective. Result: the gangs are now writing letters to the UN asking for "mediation."
Read that again. The gangs are asking the UN to protect them. They know exactly who is a threat to them and who isn't.
This isn't about Haiti. It's about the operating system the West has been running since the 60s, and it is dying in front of us. Its axioms were wrong.
Axiom one: everyone can be reinserted, the criminal is a victim of the system, punishment is oppression. El Salvador tried truces and soft approaches and became the murder capital of the world. Then it stopped believing it.
Axiom two: force is always illegitimate. But order isn't natural. It's built and maintained, sometimes by force. A state that refuses to use legitimate violence doesn't abolish violence. It hands the monopoly to whoever is willing to use it. In Haiti, that's the gangs.
Axiom three: everything can be steered from the top. Central banks, planners, aid agencies, "development" programs. Decades of money pushed down from above, and the result is always the same: a class that lives off the problem and has zero incentive to solve it. An NGO that fixes Haiti loses its budget. Think about that.
Axiom four: morality is relative and the old norms were just oppression. This one was built by a very specific elite in the 60s and 70s. In 1977, Sartre, Beauvoir, Barthes, Deleuze and others signed a petition in Le Monde defending sex between adults and minors. Those people are the intellectual grandparents of today's ruling class. Epstein isn't an anomaly. He's the tail of that comet: a network of untouchables who believed rules were for other people.
Now the system is cornered, and it has two exits.
Exit one: lock everything down. Digital IDs, programmable money, censorship sold as safety, a technological control state more total than anything Orwell imagined. Keep the bureaucracy, add the surveillance.
Exit two: let it collapse and rebuild on what actually works. Individual responsibility. Freedom. Consent. Clear rules, actually enforced. Decentralized systems that emerge from the bottom instead of central planners who are never accountable for their failures.
Haiti just showed us which one delivers.
The 20th century bet on the bureaucrat. The 21st belongs to builders, and to states that do what they say.
TL;dr- SpaceX stock is grossly undervalued.
Compute demand vs. power supply since 2022 (indexed to 2022 = 1): Demand has grown ~12× while supply is only up ~1.7×. The gap (demand ÷ supply) is already ~7× and keeps widening. SpaceX near term delivery of terrestrial data centers (with no new starts) produces $1 trillion in revenue in 2028 due to the demand/capacity mismatch.
Demand/Supply mismatch since GPT 3 launch
2
3
59
I don't know about valuable, but this is an excellent documentary about Tesla and SpaceX. If you are an investor, it'll make you consider adding to your position.
After 7 years of research and 6 months making this documentary, I'm handing you what I believe are some of the most valuable 51 minutes you'll ever spend.
I interviewed former @SpaceX, @Tesla, @neuralink and @PayPal employees. Read 40 books. Took 9,200 notes.
I Spent 7 Years Reverse-Engineering Elon Musk's Operating System
7 years of research. Interviews with former SpaceX, Tesla, Neuralink and PayPal employees. The operating system behind Elon Musk's companies: how he thinks, works, builds and sells.
2
72
More people need to watch this
Wow. Senator John Kennedy.
How is this not everywhere.
Seriously. This is insane. Listen to what he’s actually saying.
teapawt.net/HQWhiteRabbit/status/2…
1
55
Tom Henke retweeted
The Midterm is Not the Endterm
It is too soon to calibrate the direction of the post midterm Democratic and Republican Parties.
But if the first 72 hours are any indication—and despite the soon to be Speaker’s denials—the Left is already bragging about subpoenas, and it is intoxicated by the notion of investigations and impeachments.
Heady Democrats have entirely forgotten the politically suicidal fashion of the 2021-4 five civil and criminal courtroom circuses, the Mar-a-Lago raid, and the blue state deballoting efforts.
All that and more fueled the Trump recovery.
Wild talk of multi-trillion-dollar new agendas grows, like socializing the health care system, a renewed Green New Deal, student loan forgiveness, and illegal alien amnesties, along with new punitive income and estate regulations and taxes.
This socialist dream is reminiscent of the 2021 giddy splurging that gave the country a 9.1% inflation rate by midterm year 2022.
Along with the Afghanistan disaster (the Left’s correction for a Middle East ‘forever war”) the inflation would ensure Biden a 40% popularity.
The idea that the Left, which cannot do much real lawmaking without a veto-proof Congress, will try to reach deals with the White House, adopt a 1995 Clinton-like realist agenda, simply claim credit for the growing Trump economy, or moderate its rhetoric is highly unlikely.
Instead, we hear crazy talk of confiscations, new billionaire and millionaire punitive taxes, imprisonments, changing the Constitution—all in full suicidal 1972 McGovern style.
Republicans are blame-gaming each other, which can be constructive if over issues and reforms rather than social media ad hominem. But they need to message far better, about both the administration’s successes and the dangers of a socialist future.
There will be likely be no planned presidential succession, or gentlemanly agreement on a front runner. That’s likely good news. More likely as of now a triangular race between front-runners De Santis, Rubio, and Vance emerges—all of whom are far preferable to any ticket with a Newsom, Harris, AOC, Mamdani, or el-Sayed.
If the Republicans can stay reasonably united, they can ensure permanent presidential vetoes of crazy leftwing legislation, fight back on the Democratic committee inquisitions, and see the economy in the final two years become an asset in recovering their popularity.
Absent in all conversations is that there is increasing evidence that the economy is actually expanding and has the potential to boom in 2027, or at least the data indicate such potential.
The jobs report came out yesterday with a stunning 165,000 jobs created in October (accompanied by the usual journalists’ adverb “unexpectantly”), dropping unemployment to 4.1%.
Revised 3rd GDP stats now show a robust 3.4% growth—likely ensuring at least 3%+ growth for the year.
Third Quarter consumer spending soared 3.8%.
Private domestic investment surged with 19.4% growth.
Just reported last week: median family income in 2025 hit a new record ($87,460)—the highest in 60 years adjusted for inflation. And poverty dipped to its lowest rate (ever 10.2%).
In October manufacturing activity grew faster than at any time in the past five years.
Yes, prices are too high, and final 2026 inflation, given gas hikes, is scheduled to increase by one point (3.7%) from last year’s (2.7%). But it is still lower than the pan-European average (ca. 4+%)—and much lower than the Biden average over four years of 5.1%.
Gas prices are already falling. On November 6 prices fell to a $4.06 average—as global oil transit is mostly back to prewar levels and refineries scramble to handle the increased deliveries.
U.S. companies are dominating all the top 10 rankings in market capitalization/usage in key area such as software, AI, digital media, space, satellite launches, and bio/genetic engineering.
Note that today gas prices are not that much higher than 4 years ago at the 2022 midterm time ($4.06 vs. ca. $3.80), and haven’t once spiked to $5 a gallon as they did in 2022, just four months before the Biden midterms.
Yet oddly, the red wave sputtered in 2022 when the U.S. was in far worse economic shape than this week when the blue wave became a tsunami.
Messaging, media, and communication count, so note:
Four years ago GDP growth was lower than today—even though the pandemic was about over and there was a dramatic return to work and spending.
The inflation rate was almost three times higher in 2022 than now.
Median family income was $14,000 lower than now. The poverty rate was higher in 2022.
The Dow was 20,000 points lower. Foreign investment will likely end the year 30-40% higher than in 2022.
Real wage growth (nominal wages minus inflation) while not great in 2026 is likely to end the year far better than in 2022 (when real wage growth fell 2-3%).
So why would “affordability” have been decisive in the 2026 midterms when it was not in 2022?
The 21% aggregate inflation rate Trump inherited force-multiplied the average 3.15% Trump inflation rate of 2025-6— yet in a way the aggregate first-term Trump inflation rate (7.7% just a third of the Biden total) did not exaggerate all that much the 9.1% inflation rate in 2022.
Messaging, however, matters. The administration did not explain adequately that a 21% rise in prices cannot be lowered without radical deflation or recession, but only ameliorated by medium family income growth and lowering the inflation rate from the Biden 2021-5 averages, both of which occurred in 2025-6, but were forgotten.
Media coverage of the Biden 2022 weak economy were muted, while Trump’s 2-year inflation rate was exaggerated during 2026.
The Iran war became the culprit as a war of choice that made the inherited Biden 21% high prices and the inability to lower were them solely Trump’s fault. Again, there was inadequate explanation of either the economic reality or why the gamble to denuclearize Iran was defensible–with the caveat that stopping the bombing at day 38 and talking with theocrats for seven months was an unfortunate mistake.
The domestic and foreign policy activism of the lame-duck Clinton (1998-2000} and Obama (2014-2016) presidencies increased after their loss of Congress. Indeed, their popularity climbed from a once nadir of below forty to sixty percent approval or above.
There is no reason why the Trump administration cannot have a successful last two years. The economy is growing.
The crime rate is at historic lows.
The border is closed shut. The radical pro-Western political transformation of South and Central America has remade the entire Western Hemisphere.
European NATO is rearming.
Oil and gas production is at highest levels of any nation in history.
The bloated federal work force has been trimmed by over 270,000 jobs.
There is no longer a 40,000-50,000 shortfall in military recruitment. The forced reforms in higher education were long overdue.
The verdict on Iran is still out.
If it is denuclearized, the crippling sanctions continued, the Strait kept open, the provincial opposition armed—and our military gone home— the regime will face an increasingly restive population
2026 was an unequivocal defeat, but 2028 need not be.
If the administration and Congress stay disciplined and united, and can message better, while the Left goes utterly berserk, drunk on their midterm successes, socialism/Islamism, and Trump Derangement Syndrome 2.0, the tide will soon radically turn— as the economy already has.
Tom Henke retweeted
I’ve seen this kinda narrative quite a lot. So thought I’d weigh in…
First and foremost I strongly disagree.
I actually believe we’re about to enter the era of “lifecasting”…
Computers have beaten the best chess players in the world since 1997 and chess is bigger now than its ever been, millions of people watch magnus and hikaru stream and almost nobody watches two AI’s play each other even though the AI’s play better chess.
A forklift can lift more than any olympic weightlifter and a car is faster than usain bolt and we still watch the olympics, because what we care about is a person pushing their limits, the emotion, the human - because we recognise ourselves, our lives and our dreams somewhere in there.
Robin dunbar's research found roughly two thirds of human conversation is about… other people - he made the argument that language mostly evolved so we could gossip and keep track of each other in big groups.
This is the need content will increasingly feed in the future. Especially as most of our feeds become full of ai slop.
The argument that entertainment will become less human is clearly backwards. In fact every time entertainment adds human resonance we seem to become obsessed with it…
F1 cars didnt get any faster or more interesting when drive to survive came out but a global audience - largely of people who previously didn’t care about f1 - became obsessed with it because netflix showed the egos, emotions, human stories and the rivalries.
Love island, family vloggers, the premier league, its all the same thing, relatability, a story about ambition and rivalry and people taking risks and getting it wrong, all of which we can see ourselves in.
Where i agree is with purely informational content, if your watching someone explain how to fix a boiler then AI will do that better and cheaper and those creators are in trouble.
But the stuff thats irreplaceably human gets more valuable as the internet fills up with ai content.
So yes the “quant firms” will take everything where the human wasn’t the reason people watched, and i think truly human creators end up worth more than they are now.
Podcasting becomes… lifecasting?
"Content creators" as we know them, will not exist in the coming years
Nobody will be YouTubers, influencers, etc
Why?
Because there will be AI-generated content creators that are 1000x more entertaining, engaging, knowledgeable, etc
These AI creators won't be operated by individuals. They'll be operated by quant-like firms that ingest enormous amounts of audience data, continuously optimizing every aspect of their content based on real-time engagement metrics
Every video, joke, opinion, every interaction will be engineered to maximize audience attention
Especially now that algos rule everything by serving you content they think you will like
Human content creators will still exist but it will be akin to me day trading on Robinhood vs a team of 1,000 quants at Jane Street
Every county? Does that sound likely?
ALL 21 STATES. ALL 1,799 COUNTIES. EVERY SINGLE ONE SHIFTED BLUE.
You can blame the Iran war. You can blame the economy. But you expect me to believe Republicans were THAT angry that every single state and county shifted blue?
And before anyone says it, this isn’t about being unable to accept a loss. I can accept losing. What I can’t do is ignore something that just doesn’t pass the smell test.
Sorry, but I have questions. 2020 makes me look at things way differently.
2
1
58
A couple days ago Elon said 3 gigawatts in december and 13 gigawatts next year.
I went back and sat with it. I don't think anyone has actually done the math on what he said.
This is the most insane guidance a company has ever put out in public. And the market shrugged.
So let me explain.
1. What he actually said
Q3 call, 20 Oct: "I will be surprised if we're brought on less than 13 gigawatts of compute by the end of next year."
Then 31 Oct on X: "around 1.2M Blackwells at Colossus 2 by end of year"
Companies don't say things like this out loud. Most give you a capex range and hide the rest.
2. The pace so far
Jun 2025: 0.4 GW
Sep 2025: 0.6 GW
Mar 2026: 1.0 GW
Jun 2026: 1.4 GW
Sep 2026: 1.7 GW
Dec 2026: ~3 GW
The whole year to September added 1.1 GW. Q4 alone adds ~1.3 GW. One quarter beats the year before it. My math.
3. What "fast" means for everyone else
Microsoft adds about 1 GW a quarter. When it did that, a UBS analyst on the call called it "extraordinary".
Oracle had its record quarter this summer. 850 MW and more than 300,000 GPUs.
CoreWeave, the biggest neocloud, needed 18 months to go from 0.36 GW to 1.5 GW active.
SpaceX's 2027 plan is +10 GW in one year. That's ~2.5 GW every quarter, 2.5x Microsoft's "extraordinary". My math.
4. 13 GW vs a whole hyperscaler
Microsoft runs about 12 GW of data centers in total, per Bloomberg. Built and leased over decades.
SpaceX plans to pass that by the end of next year. On AI compute alone.
The first Colossus came online in September 2024. 100k H100s, built in 122 days.
5. 1.2M Blackwells
Meta's biggest cluster, Prometheus, is working toward roughly 1 million GPUs at full scale.
Colossus 2 gets 1.2M Blackwells by December.
Elon says they turn on ~200,000 GB300s a month. That's ~600k a quarter, about 2x Oracle's record quarter. My math.
6. And they do it for less
2026 capex: Amazon ~$220B. Google $195-205B. Microsoft ~$170B. Meta $135-145B.
SpaceX: ~$83-88B for the whole year. My math, $48.1B in nine months plus the $35-40B Q4 guide.
Deutsche Bank puts SpaceX at ~$35B per GW vs ~$50B for the industry.
Less money, more gigawatts. SpaceX casts its own turbine blades and brings its own power. Vertically integrated from the power plant to the app.
7. Why nobody prices it
Everyone is short of compute. Microsoft: "Customer demand continues to exceed available supply." Amazon says customers take capacity almost as fast as it's installed.
So every watt gets rented. $50 a watt a year is the floor.
13 GW x 90% rented x $50 = ~$585B a year. My math. Elon's own range for end-2027 ARR is $500B to $1T.
SpaceX trades under $2T today.
Analysts extrapolate the past. They see a rocket company that sells internet. SpaceX tells you the future in public, then beats it. "Over 2 gigawatts" in August became ~3 GW by December.
8. My read
Again This is the most insane guidance a company has ever put out in public. And the market acted like its just crazy elon talk.
In a year people will look back at these 13 days and ask how it was ever priced like this.
SpaceX is THE buy. $100T is a matter of time.
$SPCX
SpaceX guided "over 2 gigawatts" of compute by year end.
That was 4 August. 88 days ago.
Tonight Elon says around 1.2M Blackwells at Colossus 2 by year end. That's ~3 GW.
50% more compute than guided. And the stock closed Friday at $134.72, under its $135 IPO price lol
Here's everything you need to know.
1. The ramp
Dec 2025: 0.8 GW
Mar 2026: 1.0 GW
Jun 2026: 1.4 GW
Sep 2026: 1.7 GW
Dec 2026: ~3 GW
Jan to Sep added 0.9 GW. Q4 alone adds ~1.3 GW.
One quarter beats the three before it combined.
2. How the guide moved
4 Aug, Q2 call, Elon: "We expect to end this year with over 2 gigawatts of compute."
20 Oct, Q3 call, Elon: "the roughly 3 gigawatts, maybe 2.5 to 3 gigawatts that we'll have online by the end of this year"
31 Oct, Elon on X: "around 1.2M Blackwells at Colossus 2 by end of year"
2027 got raised too. August: "closer to 10 gigawatts of compute than 5 gigawatts of compute". October: he'd be surprised by anything under 13.
Guide, beat, raise, repeat lol
3. The names
Colossus 2's buildings are now called Macroharder, Minihard, Gun Range and Spam.
Somewhere in Redmond, Microsoft is reading "Macroharder" and not laughing lol
Minihard has 220k GB300s. Nothing mini about it.
Gun Range opens end of November. Fitting, SpaceX keeps shooting past its own guidance.
Spam used to house Hormel products. In December it holds ~200k GB300s for inference. That's ~0.4 GW, or ~$20B a year at $50 a watt. Most valuable can of Spam ever made.
4. What ~3 GW is worth
CFO Bret Johnsen on the Q3 call: pricing is tracking "slightly above the high end of the $30 to $50 per watt range".
$50 a watt a year = $50B a year for every GW.
Elon said in August maybe 10% of compute goes to Grok training. So rent out the other 90%:
3 GW x 90% x $50 = $135B a year from compute
Add Starlink and launch at their Q3 pace, $7.0B a quarter: +$28B
~$163B of run rate going into 2027.
Now line it up:
$58B: Q3 revenue x 4
$100B: the December ARR goal everyone called crazy
$132-148B: the Q1 2027 guide ($33-37B) x 4
~$163B: what ~3 GW can earn
The "crazy" $100B is now the floor.
And the Q1 guide only needs ~2.1-2.4 GW rented at $50 a watt. They'll have ~3. My read: that guide is sandbagged.
5. Why this is so bullish
Payback. Bret: new compute is "generating a payback of less than a year". Every GW that lands early is a year of revenue nobody modeled.
Price. Elon in August: memory supply grows "around 20% per year", demand "by 200% a year, maybe higher". His words: "Economics 101 would suggest that the price increases; it does not decrease." My read: $50 a watt is the floor, not the ceiling.
Scale. ~3 GW is the launch pad. 13 GW at $50 a watt is $650B a year. Elon's own range for end of 2027: $500B to $1T of ARR.
6. Receipts
The night of the Q3 call I posted "$125-$150 B ARR from compute alone by the end of 2026".
Tonight's 1.2M Blackwells puts it at $135-150B.
Bears said 2 GW was a stretch. Said $100B ARR was crazy. Said Elon always misses deadlines.
He's ending the year 50% above the guide.
$1.83T company. ~$163B run rate. That's ~11x.
You are still very insanely early lol
$SPCX
"She called school choice "racist" yet sends her kid to a private school."
BREAKING: Chicago Teachers Union President Stacy Davis Gates makes $445,000 a year.
She called school choice "racist" yet sends her kid to a private school.
Not a single child is proficient in math in 50 public schools in Chicago.
63
Tom Henke retweeted
🚨 READ THIS TWICE: Thomas Sowell just explained 50 years of broken promises in one paragraph.
Politicians aren't failing to solve your problems. They're succeeding at solving THEIRS.
That's why the fix was never finding a nicer politician. It's changing the incentive.
If getting re-elected is job one, the only language they understand is losing.
Every ballot is a performance review.
Tomorrow, grade them on results, not speeches.
Why are @Ravens at @BuffaloBills and @Vikings at @Lions on at the same time while two of three late games and the Sunday night game are complete trash? @NFL
2
77


































