The $65 Billion Question

Freely ye have received, freely give. — Matthew 10:8

An Absurd Amount of Money

Begin with the number, because the number is the thing that makes the question worth asking. Yesterday, on the twenty-eighth of May in the year twenty twenty-six, the artificial-intelligence laboratory Anthropic announced the close of a sixty-five billion dollar Series H financing at a post-money valuation of nine hundred sixty-five billion. The round was led by Altimeter, Dragoneer, Greenoaks and Sequoia. The company reported a revenue run rate of forty-seven billion, up from thirty billion earlier in the same year and ten billion last year. The valuation now exceeds that of the rival lab OpenAI for the first time. By every standard measure of what a private technology company can be worth before going public, this is the largest such figure in the history of the category.

Sixty-five billion is an absurd amount of money, and the absurdity is the relevant feature. It is more than the annual gross domestic product of about a hundred and ten of the world's nations. It is more than the entire venture funding deployed across all of Latin America in any year of the last decade. It is enough to pay every teacher in the United States for a year, or to fund the World Food Programme at its current scale for seven. To call this an investment is technically correct and morally insufficient. It is a coronation. The market looked at the field of laboratories building artificial intelligence and decided which one it most wanted to bet on, and it bet at a scale that has no obvious peer. The bet was on a particular orientation toward the technology and a particular set of people, and the size of the bet is the point. Smaller commitments can be hedged. Sixty-five billion is a thesis.

The Fork at the Crossroads

What thesis. The technology in question has been pulling, almost since its public emergence, in two opposite directions at once, and the bifurcation has only sharpened. Call them the two poles of the crossroads and they are recognizable enough that anyone reading this can name a company on each side.

On one pole there is dominance. The thesis is that artificial intelligence is the most powerful instrument of pattern recognition, surveillance, prediction and control ever built, and that the highest-value applications are the ones that concentrate that power: catch the suspect, predict the customer, model the battlefield, replace the worker. The business models cluster around government contracts, defense, surveillance fleets, automated license-plate readers on every corner. In my neighborhood this week, a wave of billboards has gone up objecting to one of the better-known surveillance products in this category, the residents alarmed at the prospect of having their movements logged and queried by police without warrant. That is the dominance pole made local. Palantir, Anduril, Flock, the rest of the cohort — these are competent companies built by serious people, and what they are competent at is concentrating power. The total cost paid by the rest of us for that competence is not yet on the invoice.

On the other pole there is enablement. The thesis is that the same technology, pointed differently, is the largest extension of human capability ever offered, and the highest-value application is the one that removes the penalty for being poor or remote or unconnected to the institutions that previously gated competence. The business model is direct: sell the tool to anyone who can pay a few dollars a month, and lower the price as fast as the cost of compute will let you. The customer is everyone. The product is the lifting of a ceiling that, for the entire history of the species before this moment, was lifted only for the lucky and the well-credentialed. The wager on this pole is that abundance, made universally available, produces returns no dominance strategy can match — not because abundance is morally superior, although it is, but because the addressable market for making a billion people smarter is bigger than the addressable market for surveilling them.

Both bets have been on the table for several years. Yesterday's sixty-five billion is the market saying, with the largest check it has ever written, that the enablement bet has won the round. The laboratory that closed the round has been the most consistent in its public posture on which pole it occupies, and the validation is not subtle.

What the Other Pole Did This Week

The contrast is sharpened by what happened on the other side of the fork in the same calendar week. Peter Thiel, the most prominent funder of the dominance-pole archetypes — Palantir as the surveillance backbone, Anduril as the defense layer — relocated his family to Buenos Aires, purchased a twelve-million-dollar home in Barrio Parque, enrolled his children in Argentine schools, and is reported to be in discussions with the Milei government about citizenship. He retains his United States citizenship and the bulk of his assets remain stateside, so the move is partial; but partial moves of this kind, by men of this stature, are not nothing. They are a public posture. The man who built much of the infrastructure of the dominance pole is, in the same week the enablement pole closes the largest round in technology history, audibly reconsidering where he wants to live.

The details matter. Thiel spent much of last year delivering lectures about the antichrist, building out a public theology of imminent eschatological threat, identifying the figure he expects to appear and the apparatus he expects that figure to wield. The detail I found arresting is the one the New York Times reported almost in passing: that in Argentina he has begun hosting dinners with local economists on the same subject. A man does not pack the antichrist into the moving truck unless the antichrist is doing some load-bearing work in his picture of the future. And in the same week, the same man was reported as suggesting that he might have, in his own words, overstated artificial intelligence's impact on society, particularly on employment.

I will not condemn Thiel. He is a brilliant man, and the impulse to look for the antichrist is at least an impulse that takes the eschatological seriously, which is more than most of his peers manage. But the world does not, in fact, need a new antichrist. That ship sailed in the days of Nero, and the long pattern of identifying a new one in every generation has more often been a way of avoiding the work in front of us than a way of preparing for it. We do not need a new boogeyman to resolve to do better and finish well, one soul at a time, with the humility that recognizes we were all in this together. The man who is moving the antichrist conversation south, while the laboratory on the other pole closes the largest round in history, is showing us something about which pole expects to need a villain and which one expects to need only a customer.

Factory Settings

The case for enablement is the case for what I would call factory settings — the conviction that the human being, as delivered, is fundamentally good. Not perfect; not without the capacity for cruelty and self-deception; but built with the equipment for cooperation, creativity, attention to the other, gratitude, and the kind of joy that does not require another person to be worse off for one to be better. The corruption of those settings is real, and it is what theology calls the fall and what economics calls fear and greed and what every parent calls the moment a child learned, by being hurt, that the world was not safe. The corruption happens early and it happens to everyone, and it produces the world we have. But the original settings are still there underneath, and most adults can recognize, in the moments when they are not afraid, that the better version of themselves is the one closer to the equipment they came with.

The scarcity argument against this anthropology has always been that there are only so many seats at the top, only so many places at the best school, only so many promotions available, and therefore the rational human strategy is to compete hard and protect what you have. That argument is true in a scarcity economy, and the scarcity economy has been the only economy for the entire history of the species before now. What artificial intelligence does, at the enablement pole, is collapse the scarcity that the argument depended on. The best school had limited seats because the best teachers had limited hours. The best teachers had limited hours because the best teachers were embodied humans who could only be in one room at a time. The barrier was not malice or even gatekeeping; it was the bandwidth of attention itself.

A tool that can give every person on earth the equivalent of a patient, brilliant, infinitely available tutor breaks that barrier in a way no education reform ever could. The ten thousand hours that Malcolm Gladwell named as the cost of expertise might compress, at the enablement pole, to one thousand for the cutting edge and to on-demand for ordinary competence. The penalty for being poor is, in the largest part, the penalty for not having access to the people who already know how — and that penalty, for the first time, has a route around it. This is what I have heard called WorldMaxxing in some of the corners that take it seriously: the idea that any person, equipped with the right tool and willing to put in the practice, can become enormously more capable than the previous era allowed. Not equally capable — people have different gifts and different intensities, and the spread will remain — but the floor lifts in a way that drags the ceiling up with it.

If the factory settings are good, then a tool that lets a billion people return closer to them, while doing more in the world than they could have dreamed of doing in any prior decade, is the largest moral good a technology can produce. The economic argument is the same argument restated in numbers. A person whose existence contributes five hundred dollars to global gross domestic product per year, because they live where the institutions that would have multiplied their labor never reached, is potentially enabled to ten or a hundred times that figure by a single tool. Not as a charity case. As a rational, self-interested participant in an economy that suddenly has room for them. The total addressable market for making the bottom half of humanity dramatically more productive is so much larger than any surveillance or defense market that it is not obvious how the dominance pole survives the comparison once the enablement pole has the capital to scale.

The Question the Money Asks

Which brings us back to the sixty-five billion, and to the question the title is built around. The enablement bet has now been validated, in capital terms, at a scale that should silence any remaining doubt about whether the thesis was correct. The market has spoken at the largest volume it knows how to speak. The question that immediately follows is the one no announcement answers, and the one I think anyone watching this should hold in the front of their mind for the next twelve to eighteen months. Does the governance that comes with sixty-five billion dollars accelerate enablement, or impair it?

The question is not idle. Large rounds come with board seats, with covenants, with information rights, with the slow gravity of investor expectation. Some of the investors in this round are sovereign wealth funds. Some are public-market crossover funds with quarterly reporting cultures. Some are firms whose historical relationships with their portfolio companies have been excellent and some whose historical relationships have been less so. The character of the laboratory that closed the round has, until now, been remarkably consistent — published research on alignment, public posture on safety, a co-founder who stood at the Vatican this Monday and argued that the building of these tools cannot be left to the builders alone. That posture is what won the round. The honest question is whether the round will, in time, change the posture.

There is a fork inside the fork here, and it is the one worth watching. Sixty-five billion dollars can be deployed in two very different ways. It can be deployed on five or ten overpriced acquisitions, each one bringing a team and a product and a layer of organizational complexity, in pursuit of becoming the kind of conglomerate that defends a market position by getting bigger. Or it can be deployed on enablement infrastructure at scale — on the model improvements and the deployment economics that let a billion people, not five hundred million enterprise seats, become absurdly smart in two thousand twenty-six. The first path is what most companies do with this kind of money, because most companies have run out of ideas by the time they raise this much. The second path is the one that justifies the valuation on the only ground that matters in the long run, which is whether the bet on factory settings was true.

I hold no position from which to predict which path will be chosen, and the people who do hold such positions are not in the habit of confiding in pseudonymous bloggers. What I can say is what I would watch for. Watch the pricing of the consumer product over the next year — does it fall, dramatically, or does it stabilize at premium tiers that protect margin? Watch the geographic deployment — does access expand into the markets where the per-capita income is too low to support United States pricing, or does it cluster in the markets that can pay full price? Watch the research releases — do they continue to publish alignment work in the open, or does the publication slow because the investors prefer the moat? Watch the executive moves — are the hires coming in the shape of an enablement company that needs to scale distribution, or in the shape of a defense contractor that has decided the easier money is in the other pole after all? None of these will be conclusive on their own. Together, over a year, they will tell the story.

The Better Pole

I will close where I started, because the close is the same as the start. Sixty-five billion is an absurd amount of money, and the question is what it is for. The most hopeful reading of this week is that the market, after a long period of being unsure which pole would prevail, has placed its largest single bet on the pole that scales human capability rather than concentrates it, and has placed it on the laboratory whose public posture has been the most consistent in that direction. The most cautious reading is that capital of this size has its own gravity, and that the laboratory will have to work harder than any laboratory has worked before to keep the posture it had when the capital arrived. Both readings can be true at once. The honest position is to root for the first and watch closely for the second.

The Vatican this week named the choice as the choice between two cities — the tower built to make a name for the builders, and the city built in shared responsibility under God. The Pope's vocabulary is not everyone's vocabulary, but the fork the encyclical names is the same fork the capital just bet on. Build the dominance tower and you will own a great deal of brick and very few souls. Build the enablement city and you may, in this generation, lift more people out of the inherited scarcity than every prior development effort combined. The capital says the latter is the more rational bet. The next year will say whether the people deploying the capital still believe what they said when they raised it.

Make a billion people absurdly smart in 2026. That is the assignment if the bet is real. If it gets done, the sixty-five billion will look, ten years from now, like the bargain of the century. If it does not, the round will be remembered as the moment the enablement pole had every resource it needed and chose, with the slow tragedy that besets every well-funded company, to spend the resources on something else instead. The world is rooting for the first outcome. The world should also be paying attention, because the people watching is what keeps the bet honest. Unity in truth. Love conquers all.

---

The funding facts in this essay are drawn from contemporaneous reporting of May 28, 2026. Anthropic's $65 billion Series H at a $965 billion post-money valuation, led by Altimeter Capital, Dragoneer, Greenoaks and Sequoia Capital, and the company's reported $47 billion annualized revenue run rate, were announced on May 28; the prior Series G ($30 billion at $380 billion in February 2026, led by GIC and Coatue) is included for context. The Vatican appearance of Christopher Olah on May 25, 2026 at the launch of Pope Leo XIV's encyclical Magnifica Humanitas is verified to public reporting and is treated here as a posture event rather than as evidence of any institutional relationship between the laboratory and the Holy See. The Peter Thiel relocation details are drawn from New York Times reporting of May 27-28, 2026: temporary relocation to Buenos Aires, purchase of a $12 million home in Barrio Parque, enrollment of children in Argentine schools, ongoing discussions with the Milei government about possible Argentine citizenship, retention of US citizenship, and the family's hosting of dinners with local economists on the subject of the antichrist. The reported softening of Thiel's prior AI-impact claims, particularly on employment, is from the same reporting and is presented as his stated reconsideration rather than as a verdict on the underlying question. The objection here to the antichrist framing — that the figure was historically fulfilled in the era of Nero, and that the recurring search for a new candidate has often been a way of avoiding present responsibility — is named as the author's confessed theological position and is offered as such. The factory-settings anthropology — that the human person, as delivered, is fundamentally good, and that scarcity-driven fear and greed are the corruptions of that original state rather than its substance — is also the author's confessed conviction. The economic argument that a person currently contributing approximately $500 in annual GDP can be enabled to multiply that figure by an order of magnitude with the right tooling is offered as a plausible upper bound rather than a guaranteed outcome; the actual realized multiplier will depend on infrastructure, language coverage, energy availability, and the deployment choices the named laboratory and its peers make over the coming year. The four observable indicators proposed at the close — consumer pricing trajectory, geographic deployment pattern, alignment-research publication cadence, and executive hiring shape — are offered as practical signals by which any reader can independently track whether the enablement bet survives its own validation. Disclosure: this essay was drafted with the assistance of Claude, an AI system made by Anthropic, which is the subject of the essay. The author has held an admiring view of the laboratory's work since well before this week's announcement and has stated that view in earlier writing under the WellSpr.ing record. The conflict of interest is real and is named here so that readers can weigh the argument accordingly. The argument stands or falls on its own merits, not on the disclosure. In all thy getting, get understanding.