---
title: "Part VII: The Firm"
part_of: "The Commanding Heights"
part_of_url: https://commanding-heights.com/commanding-heights.md
authors: ["Will Manidis", "Joe Roberts"]
date: 2026-10-29
url: https://commanding-heights.com/#part-vii
summary: "Economic activity is organized around firms, their shape evolving with each technological revolution. The originary owner-operator, the manager-employee, and shareholder-disciplined firm each had their era; the last has been made obsolete by intelligence. Tomorrow, firms will be composed of a few operators, paid and governed like capital, directing work done mostly by models. Venture becomes trading, private equity is refashioned to diffuse model capability, and the sovereign re-settles its obligations to the private sector."
---

## Part VII: The Firm

> *Economic activity is organized around firms, their shape evolving with each technological revolution. The originary owner-operator, the manager-employee, and shareholder-disciplined firm each had their era; the last has been made obsolete by intelligence. Tomorrow, firms will be composed of a few operators, paid and governed like capital, directing work done mostly by models. Venture becomes trading, private equity is refashioned to diffuse model capability, and the sovereign re-settles its obligations to the private sector.*

The firm is a tool of human interest organized for the benefit of those who work within it or own it. It has endured throughout every historical and material circumstance. Wherever there is a multitude of humans, there are firms or many in the making. This elemental has taken on a variety of names throughout history, but its function has remained the same.

Let us recapitulate the neoteric conditions that we see: We expect the line to be sustained, from capability growth to capital expenditure to the monolithic value of the market. The technology sector becomes, in reference, the economy, and assumes a new level of centrality. The enduring value of machine intelligence will seep away from its originary genesis in the labs to new points of capture. As capital and its uses agglomerate to those states best positioned to use them, future model progress and its diffusion will be sourced only from China and America, themselves a fearful symmetry.

In China, Sinomodernity will demonstrate diverging forms of governance and the exercise of its power. The centrality of machine intelligence will force the state to treat its labs as peculiar projects, answerable to the sovereign alone. The proliferation of the Neo-Dollar will usher in the superscription of all things within the American digital sphere of influence. As a result, value will primarily agglomerate to distribution and network owners, as the ladder to development is kicked out from beneath the ascending poor world. The K-shaped economy will grow evermore at all levels, internationally, within America, amongst firms and individuals, within firms and individuals. Roads of excess, palaces of wisdom. The shape of things to come.

Amid all this machine intelligence-caused reordering, we foretell a new opportunity to purchase control over the Commanding Heights. Long captured during the old economic regimes, the sands are shifting, and the positions of Command in the intelligence era have not yet been appropriately valued. Amid the simultaneous upheaval and prospering of the world, these Heights will be scarce, controllable positions of economic power whose holders durably capture the gains of increasingly capable and diffused intelligence. But to understand this future, we must first look back to the past. When firms reorder the Heights into a new compact of ownership, they are continuing a centuries-old ontology.

### Guinness

We can anatomize the function of the firm with the case of Guinness. Founded by Arthur Guinness in 1759, his firm has survived long enough to live through every arrangement and material circumstance that the modern Western world has known.[^120]

[^120]: Patrick Lynch and John Vaizey, *Guinness’s Brewery in the Irish Economy, 1759–1876* (Cambridge: Cambridge University Press, 1960).

At first, it was that elemental expression of commerce, a simple brewery at St. James’s Gate, Dublin. Mr. Guinness leased the property with his own capital and held a 100% interest in the firm. He hired his employees to operate the brewery and roll the barrels of product at a wage he dispensed daily. It was a familiar scene, at a human scale, depicted alongside the baker and the butcher by Adam Smith. It did not remain so. The Guinness brewery evolved with the material conditions and scale of the times. For decades, Arthur and his kin toiled away in those early days of firm growth that, on the compounding graph an era or two later, look imperceptibly different from no growth at all.

![Line chart of Guinness sales volume from 1840 to 2024, read from a Diageo chart that gives no scale. Volume grew through the nineteenth century, fell sharply during World War I, sagged in the 1930s, fell from the mid-1970s to the early 1990s, dipped in 2020, and reached its highest level in 2024.](https://commanding-heights.com/img/guinness-volume-chart-print.png)

*Guinness sales volume, 1840–2024. Source: Diageo.*

Source: Diageo, “Introduction,” *Guinness Investor and Analyst Event*, May 20, 2025, 19. Volumes from IWSR, validated against Guinness Archives records for 1841–2003.

By 1886, it was time to float the firm on the London Stock Exchange, an institution younger than the brewery by forty-two years. At this moment, a new settlement in firm organization occurred: the control and capital of Guinness became two distinct things. One hundred years later in 1986, the company’s chief executive executed a takeover of the Distillers Company. A century of evolution had further still broken up the component parts of the firm, this time separating management from capital by ceding decision-making to career-manager-employees. While Guinness’ CEO eventually succeeded in his goal of centralizing industrial control, the ensuing share scandal caused by his nefarious method of doing so resulted in a further reordering of the City of London and sent him to prison. The affair became a herald of the corporate malfeasance and excess that would come to characterize late-twentieth-century firm activity, finding its recompense as the calendar turned to the twenty-first. In 1997, Guinness, by now an international conglomerate, merged with Grand Metropolitan to form Diageo. The Guinness firm had gone from a brewery to a joint-stock company and reached its terminus as a brand in a conglomerate.[^121]

[^121]: Guinness Archive, “Papers of the Guinness Company,” Irish Archives Resource. London Stock Exchange Group, “The History of LSEG.” Criminal Cases Review Commission, “Saunders, Ernest,” September 9, 2024.

The functions of all firms can be traced through such varying arrangements of ownership, capital, labor, and their agglomeration. The tracing must answer who supplies the capital, who makes the final decisions, who carries them out, and to whom each part of the whole is owed. Each unique combination of these functions is its own settlement, a thing impermanent but tangible, a corpus with a shape. Only once formed can it then be referenced and pointed at and put up to trade on the exchange.

But such forms are never permanently settled. They are instead ever changing. In 1937, a young Ronald Coase explored the question of why a firm should exist at all, if in fact markets were so very good at coordinating activity between individual actors. If the interest of each is the good of all, why these particular groupings, this distinction of an in-the-firm-group and an out-group, without which the firm would not exist? His answer was that markets were costly to use. The efficient market is an expensive market and inefficiencies can be thought of as savings that would otherwise be expended in the flow of information. Think of Coase’s idea like this: work that is difficult to describe or hard to measure is less expensive to operate from within the firm than from without. To operate outside the firm involves transaction, which, as he described, always includes a cost. Thus, the form of the firm is as a settlement between the composite factors optimized for the cost of transaction it undertakes.

Thus, technology becomes a key determinant of the form of the firm. With each new technological era, a new, more effective form of the firm is unlocked, and commerce migrates there, ultimately becoming the predominant settlement of the component firm variables of capital, control, and labor. Since the eighteenth century, we have observed three different settlements, each driven by a new era of progress, each with their own form and Heights. Today, the era of intelligence is creating a new settlement, only the fourth compact since capitalism left its own Genesis.

### Settlements

Each tectonic shift in the settlement of the form of the firm is preceded by changes in the origins and flows of returns. If one watches the returns with as much focus as one gives the line, they can be just as revealing about the true nature of the corresponding firm.

The first settlement ran from that early modern period in the eighteenth century and terminated in the 1920s. This was the time of the human-sized firm so lovingly described by Adam Smith. Mr. Guinness, the owner-operator of his one-building brewery, his own name on the placard and a plan to pass it down in patrilineal fashion. To speak of management and capital as separate things would have seemed ridiculous. Capital was the justification of the right to manage. Even towards the end of this period, the early 20th century, the two were not so distinct. When J.P. Morgan brought together U.S. Steel as the first $1B corporation in 1901, the convening trust was designed so that Morgan’s own partners, those who supplied capital, would be able to control the operations of the firm.[^122]

[^122]: Ron Chernow, *The House of Morgan: An American Banking Dynasty and the Rise of Modern Finance* (New York: Atlantic Monthly Press, 1990).

But each settlement, by its arrangement, contains in itself evidence of what it lacks. Morgan stretched the limits of what owner-operators could do, which was fundamentally constrained by the scale of the enterprise. These were firms built for individuals. By the 1920s, the scale of everything and the complexity inherent meant that individuals were no longer capable of overseeing the firm in its entirety. Ownership and management became distinct things. To meet the scale of the capital needed, ownership was divided into shares and spread across millions of part-owners who couldn’t possibly force any kind of change of operations. That was left to the managers, who had immediate control over facilities and the levers of power, but held very little in the way of ownership. This was described by James Burnham in 1941 as the managerial revolution.[^123] And it was a successful revolution.

[^123]: Adolf A. Berle Jr. and Gardiner C. Means, *The Modern Corporation and Private Property* (New York: Macmillan, 1932). James Burnham, *The Managerial Revolution: What Is Happening in the World* (New York: John Day, 1941). Alfred D. Chandler Jr., *The Visible Hand: The Managerial Revolution in American Business* (Cambridge, MA: Belknap Press of Harvard University Press, 1977).

Scale brought the firm to its second settlement, which ran from the 1930s to the 1970s. Those corporations that were controlled by managers and owned by multitudes were the best at returning a profit to each and every member of the owning body. These new managers were intimately familiar with the firms, in many cases spending their entire careers at one company and rising up the ladder. They had pensions, like other employees. Management in the second settlement was effectively a privileged form of labor. The firms were run in a way that reflected this dynamic, and managers felt a debt of obligation to offer generous benefits to their workers.[^124] Such was the American Dream.

[^124]: S. R. Dennison and Oliver MacDonagh, *Guinness 1886–1939: From Incorporation to the Second World War* (Cork: Cork University Press, 1998).

What was lacking in the second settlement? Ruthlessness, a willingness to trim the fat. There was too much commonality between the managers and the managed, and when inflation in the 1970s ate the cash sitting on the corporate balance sheets, the next settlement became an adjustment to this leanness.

That hunger sparked the revolt of capital. Like any revolution of the downtrodden, it began with self-organization around common interests. Capital has opinions and interests like any other, and by the late 1970s, it felt that it was not receiving its due. The pensions were too generous; the margins were not fat enough. Michael Jensen and William Meckling described the dynamic at play in 1976, stating that managers were agents whose interests could diverge from those of the owners. The way to ensure good governance, which in this case meant obeyance on the part of managers, was to incentivize alignment with owner oversight and stock-based pay and bonuses.[^125] From the buyout shops to the activist investors, capital began to sing from this hymn with one voice. They sang for new managers, who would swear fealty to those above, rather than below. The firm assumed a new hierarchy, a reflection of the relationship between its constituent parts. The owners disciplined the managers, and the managers, under that guidance, executed the defining trade of this settlement: they moved labor off of the corporate balance sheet and into the hands of vendors.

[^125]: Michael C. Jensen and William H. Meckling, “Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure,” *Journal of Financial Economics* 3, no. 4 (1976): 305–60. Michael C. Jensen, “Eclipse of the Public Corporation,” *Harvard Business Review* 67, no. 5 (September–October 1989): 61–74.

Such were the feats of accounting that produced fantastic returns atop otherwise dreadful stagnation. A key characteristic of the third settlement was the falling cost of capital. Treasury yields peaked near fifteen percent in 1981, towards the beginning of the settlement, and ground down to about half of a percent forty years later.[^126] The increasing availability of money masked a similarly increasing sluggishness and disinterest in the novel in favor of the profitable. This moment encompassed the decades that Peter Thiel describes as the great stagnation.

[^126]: Board of Governors of the Federal Reserve System, “Market Yield on U.S. Treasury Securities at 10-Year Constant Maturity, Quoted on an Investment Basis” (DGS10), FRED, Federal Reserve Bank of St. Louis, observations for September 30, 1981, and August 4, 2020.

This constituted the third settlement, which lasted until quite recently. You lived through the end of one thing and the beginning of a thing quite unlike it. Did you notice? History is back, with a vengeance.

You are now living through the formation of the fourth settlement, one essentially neoteric in its attributes. It began around 2023, breaking from several prior variables. The tailwind from falling interest rates gave out in the same few years that the transformer became central to the market. In such conditions, technology once again became the source of margins, as it had been in the first settlement. Such novel technology needs operators, not disinterested management. **The role of management itself will be replaced by the fusion of operators and capital.**

In some sense, this is a return to the eighteenth century, when a small elite with distinct knowledge about the latest developments in technology managed their own capital. There is a parallel dynamic to the labor of the first settlement, which was severely disadvantaged and had little in the way of bargaining power. Except now, that labor is done primarily by models, which are more easily managed by charismatic individuals. There still must be capital, of course, and operators. But management is more of a function than a role, and it has been subsumed. The fourth settlement is one sympathetic to great men and leaves little room for labor as you know it.

![Line charts, 1947-2026, of corporate profits after tax at a record 13.1 percent of GDP, labor’s share of nonfarm business income at a record low of 93.4, and CEO pay at 325 times a typical worker’s in 2025, with the second, third and fourth settlements shaded; and a table of the four settlements of the firm, showing who supplied the capital, made the decisions, did the work and was owed in each, with average profits, labor share and CEO pay for the second, third and fourth.](https://commanding-heights.com/img/three-settlements-chart-print.png)

*Corporate profits, labor’s share of income and CEO pay, 1947-2026.*

*Who supplied the capital, made the decisions, did the work and was owed, by settlement.*

Source: U.S. Bureau of Economic Analysis, “[Corporate Profits After Tax (without IVA and CCAdj)](https://fred.stlouisfed.org/series/CP)” and “[Gross Domestic Product](https://fred.stlouisfed.org/series/GDP),” via FRED, Federal Reserve Bank of St. Louis, for 1947-1969, and via FactSet from 1970. U.S. Bureau of Labor Statistics, “[Nonfarm Business Sector: Labor Share for All Workers](https://fred.stlouisfed.org/series/PRS85006173),” index 2017 = 100, via FRED. Profits and labor share are quarterly, from 1947 through the second quarter of 2026. Josh Bivens, Elise Gould, and Jori Kandra, “[CEO Pay Has Skyrocketed since 1978](https://www.epi.org/publication/ceo-pay/),” Economic Policy Institute, September 15, 2026. CEO pay is realized compensation at the 350 largest publicly owned U.S. firms as a multiple of a typical worker’s pay, with gaps before 1992. In the table, profits and labor share are averages, and CEO pay a range, over each settlement’s years in the data, from 1947 for the second settlement and through the second quarter of 2026 for the fourth. The first settlement ended before the data begin. Calculations by Commanding Heights.

### Labor Leaves the Firm

The departure of labor from the firm is not happening all at once. It first went off of the balance sheet during the third settlement, but it was still contracted in the form of humans by the firm. In the fourth settlement, labor is leaving the firm altogether.

Legibility is the fulcrum of understanding where labor goes from here. Recall that Coase’s justification for the existence of firms was the comparative advantage of conducting certain operations and labor within the firm, rather than outside, because of the cost of transactions. Increasing legibility, which lowers the cost of transactions, dramatically shrinks the size of firms by headcount, while increasing capabilities because more and more of their operations can be externalized. The offshoring boom of the third settlement took an enormous amount of interior labor that was now suddenly legible and low-cost enough that a highly profitable transaction could be conducted. This legibility was the precursor to where we are today with models. Once various human functions could be commoditized with service levels and per-ticket costs, and conducted simultaneously from across the world, we were not far off from those functions being made legible to a machine intelligent model. Ravi’s dismissal is the case in miniature. His time spent reviewing agent output and sending it back to the orchestrator model was never management in any sense that mattered. He was in the training environment for the agent that replaced him.

But prior to the arrival of models, there were certain inefficiencies within in-firm labor that were tolerated for political reasons. We expect these to continue, though their stakes will move from in-firm politics to matters of national import. Think of the allocation of payroll as a sort of political settlement within the firm, especially the larger ones. Each of the greatest technology companies relied on small teams that really printed the money, which allowed the manifold other parts to operate and putter around. If a key contributor to Google’s core search business went on the market, one of those few essential engineers, how many billions would his salary have to be? Shareholders, still believing themselves in the third settlement, and staff, unwilling to acknowledge their obsolescence, would have revolted. The political equilibrium within the firm is to hire many and pay them all reasonably well, so that the import and pay of the few, the indispensable few, look more reasonable in retrospect. The result was a salaried bourgeoisie in which the great majority were overpaid and the small minority, while highly remunerated, were still not paid according to their true value. This is the dynamic we expect to endure in some form, even in the intelligence economy.

We expect that this new settlement will be primarily determined by what happens to labor within the interior structure of the firm. What remains of labor will be split in three ways. A large class of licensed professionals will go on continuing to collect their rents in exchange for their signatures. A larger class goes on taking home a salary that is rapidly being debased. Its new form may allow the fusion of capital and management into a thing not easily separated out again. A small class of labor, selected for their extraordinary abilities, will become operators, who will be paid more like capital.

### No-Collar Labor

A defining characteristic of the fourth settlement is that once the few true producers are out in the open, the firm will have to pay and govern them as if they were capital.

Doing so will be especially difficult for publicly traded companies. They are historically bad at paying their top talent and even worse at keeping them. A compensation committee cannot really pay a highly valuable employee what he is truly worth because there is an entire remainder of the company to account for, its opinions and objections to the unvarnished exchange of value that they have leeched off of, made legible.

In the inverse, the most valuable employees are poor fits for bureaucratic publicly traded companies, with their widely applied procedures and requirements. The kind of person who could make the firm a fortune will not idly sit through a managerial review, or even worse, the online annual training program. Before the Volcker Rule, banks were accommodating of such singular figures, who could run proprietary books right up to the edge of the risk and leverage permitted. After the Volcker Rule, they retreated to hedge funds and odd trading pods.[^127] Value has followed them.

[^127]: Board of Governors of the Federal Reserve System, “Agencies issue final rules implementing the Volcker rule,” December 10, 2013. Stephen Gandel, “Wall Street’s trading shuffle bankrolls the casino,” *Reuters Breakingviews*, October 8, 2025.

These are the circumstances that have given birth to the acquihire/hackquisition: a new proto-form of M&A that acquires talent by comping operators as though they were the firm.[^128] These days, operators are becoming capitalized cash flows. Meta’s purchase of half of Scale AI in 2025 was really just a $15B offer to get Alexandr Wang to run their superintelligence effort.[^129]

[^128]: Miles Kruppa and Lauren Thomas, “Google Paid $2.7 Billion to Bring Back an AI Genius Who Quit in Frustration,” *Wall Street Journal*, September 25, 2024. Maxwell Zeff, “Windsurf’s CEO Goes to Google; OpenAI’s Acquisition Falls Apart,” *TechCrunch*, July 11, 2025.

[^129]: Ivan Mehta, “Scale AI Confirms ‘Significant’ Investment from Meta, Says CEO Alexandr Wang Is Leaving,” *TechCrunch*, June 13, 2025.

![Dot chart, on a log scale, of AI pay packages and prices paid for AI teams against one year of median S&P 500 CEO pay of $17.3 million, from Ruoming Pang’s package of more than $200 million to Nvidia’s $17 billion Groq deal.](https://commanding-heights.com/img/ai-talent-prices-chart-print.png)

*Prices paid for AI talent, 2024-2025, against a year of median S&P 500 CEO pay.*

Source: AFL-CIO, “[Executive Paywatch - 2026](https://aflcio.org/paywatch),” median S&P 500 CEO pay of $17.3 million in 2025. Mark Gurman and Riley Griffin, “[Meta Poached Apple’s Pang with Pay Package over $200 Million](https://www.bloomberg.com/news/articles/2025-07-09/meta-poached-apple-s-pang-with-pay-package-over-200-million),” *Bloomberg*, July 9, 2025. Mike Isaac, Eli Tan, and Cade Metz, “A.I. Researchers Are Negotiating $250 Million Pay Packages. Just Like N.B.A. Stars,” *New York Times*, July 31, 2025. Berber Jin and Keach Hagey, “Thanks for Your $1 Billion Job Offer, Mark Zuckerberg. I’m Gonna Pass,” *Wall Street Journal*, August 2, 2025. Meta disputed the figure. Shirin Ghaffary and Rachel Metz, “Microsoft to Pay Inflection AI $650 Million after Scooping Up Most of Staff,” *Bloomberg*, March 21, 2024. Lora Kolodny, Jordan Novet, and Kif Leswing, “Nvidia Just Spent over $900 Million to Hire Enfabrica CEO, License AI Startup’s Technology,” *CNBC*, September 18, 2025. Julia Love et al., “Google Is Said to Pay $2.4 Billion for Windsurf Assets, Talent,” *Bloomberg*, July 11, 2025. Marina Temkin, “[More Details Emerge on How Windsurf’s VCs and Founders Got Paid from the Google Deal](https://techcrunch.com/2025/08/01/more-details-emerge-on-how-windsurfs-vcs-and-founders-got-paid-from-the-google-deal),” *TechCrunch*, August 1, 2025. Kruppa and Thomas, “Google Paid $2.7 Billion.” Jonathan Vanian, “Scale AI’s Alexandr Wang Confirms Departure for Meta as Part of $14.3 Billion Deal,” *CNBC*, June 12, 2025. Dan Primack, “[Scoop: Nvidia Deal a Big Win for Groq Employees and Investors](https://www.axios.com/2025/12/28/nvidia-groq-shareholders),” *Axios*, December 27, 2025. Packages are multiyear totals as reported. Team deals licensed technology and hired staff without buying the company, except Meta’s 49 percent of Scale AI. Nvidia booked $14.4 billion of the Groq price as goodwill, “primarily attributable to the workforce and future development of the licensed technology,” and put the total consideration at about $17 billion. NVIDIA Corporation, [Form 10-K for the fiscal year ended January 25, 2026](https://www.sec.gov/Archives/edgar/data/1045810/000104581026000021/nvda-20260125.htm), note 2.

But compensating outlier talent appropriately remains a difficult thing for large corporations to accommodate. It remains scarce, and, properly priced as capital, operators will demand corresponding governance. Hardly the employment agreement the American enterprise is used to. Equity, as defined in business-school textbooks, begins to lose its meaning.

Take Anthropic, for instance. Ahead of its public offering, the firm asked shareholders to approve an arrangement where its seven co-founders, themselves some of the most desirable talent in the world, would control a collective 50.1 percent of the votes on key corporate matters through a Founder LLC. In the strict (and far less interesting) economic sense, the seven co-founders own about two percent of the company each.[^130] Command is what matters. Talent can make such demands, which were once to be reserved for capital, and capital alone, in the old worlds of the second and third settlements.

[^130]: Connie Loizos, “Anthropic’s Founders Seek Voting Control ahead of IPO,” *TechCrunch*, September 25, 2026. Echo Wang, Ross Kerber, and Jeffrey Dastin, “Anthropic Leaders to Control AI Lab via ‘Founder LLC’ to Promote Public Good over Market Forces,” *Reuters*, September 28, 2026, updated September 29, 2026.

![Dot chart of economic ownership against voting control. Anthropic’s seven co-founders would hold 50.1 percent of the vote with about 14 percent of the shares, compared with Snap, SpaceX, Meta, Alphabet, OpenAI, Tesla and Nvidia.](https://commanding-heights.com/img/founders-voting-control-chart-print.png)

*Voting control against economic ownership at selected companies.*

Source: Snap Inc., [Form 10-K for the fiscal year ended December 31, 2025](https://www.sec.gov/Archives/edgar/data/1564408/000156440826000013/snap-20251231.htm), filed February 5, 2026. Space Exploration Technologies Corp., prospectus, June 11, 2026, and Elon R. Musk, Schedule 13G, filed August 13, 2026. Meta Platforms, Inc., [2026 proxy statement](https://www.sec.gov/Archives/edgar/data/1326801/000162828026025532/meta-20260416.htm), April 16, 2026. Alphabet Inc., [2026 proxy statement](https://www.sec.gov/Archives/edgar/data/0001652044/000130817926000342/goog-20260424.htm), April 24, 2026. Loizos, “Anthropic’s Founders Seek Voting Control.” Wang, Kerber, and Dastin, “Anthropic Leaders to Control AI Lab.” OpenAI, “[Our Structure](https://openai.com/our-structure/),” accessed October 4, 2026. Elon R. Musk, Schedule 13G/A for Tesla, Inc., filed June 17, 2026. NVIDIA Corporation, [2026 proxy statement](https://www.sec.gov/Archives/edgar/data/0001045810/000104581026000036/nvda-20260512.htm), May 12, 2026. Anthropic’s arrangement is proposed ahead of its IPO, and the co-founder stakes of about 2 percent each come from one report. Snap’s economic share counts only Class C shares. Alphabet’s 52.7 percent combines Page’s 27.4 percent and Brin’s 25.3 percent. Meta’s economic share is calculated from the proxy’s share counts. SpaceX’s range runs from the shares Musk holds outright to the as-converted total in his Schedule 13G, which adds restricted shares and options. Tesla’s 19.9 percent excludes unearned performance award shares, which are voted in proportion to other holders.

An examination of trading firms further outlines a characteristic of the fourth settlement. These were the early leaders of shedding talent, keeping only a small team that produced enormous profits.[^131] It is effectively a partnership, a common endeavor among a select few capable of undertaking it. There are no shareholders to account for, nor even any stock that one should worry about. There is only the margin, fairly distributed among the partners.

[^131]: Gregory Zuckerman, *The Man Who Solved the Market: How Jim Simons Launched the Quant Revolution* (New York: Portfolio/Penguin, 2019).

There is a deeper reason that the partnership model works well now and will continue to do so in the future. In 1970, Roger Conant and Ross Ashby proposed what is known as the good regulator theorem: every good regulator of a system must in fact be a model of that very same system. If you apply that formulation to markets, it means that the best participant within a market will be the one most nearly isomorphic with it. It is for this reason that a great trading firm aspires to this end, becoming indistinguishable from the market itself. Conant and Ashby described what is effectively a form of evolutionary pressure within the market. Like a hedge fund, any successful firm will work out how a market produces value, then get better iteratively at capturing it. Over time, all thriving firms get closer and closer to this end. The hedge-fundification future of every industry is a sign of the times and one rewarded in the fourth settlement.

The market-applied good regulator theorem also governs and predicts the behavior of labs. Data is everything, superseding what are the more traditional markers of value for a firm. If, at a certain scale, the best and most profitable way for a lab to prepare long-horizon tasks whose results are hard to verify is to own and operate a business, it will do so. A reinforcement-learning environment at the scale of a whole firm. There’s a reason it was Google and Mercor bidding for the operational data of Spirit Airlines.[^132] Soon it will be important enough to buy the business outright without waiting for it to become a distressed asset.

[^132]: Nathan Bomey, “Google Wins Bankruptcy Auction for Spirit Airlines Emails, Chats, Documents,” *Axios*, August 17, 2026.

This idea taken to its outer limit, which we are not quite so far away from, sees the labs begin to resemble chaebols. As the pseudonymous roon once put it, the labs’ Coasean boundary will keep moving outward until every interesting industry sits inside of it.[^133] Beyond measure, incoherent when examined as distinct businesses but together serving the higher arrangement of another purpose incomprehensible to those outside of the firm.

[^133]: roon (@tszzl), post on X, June 13, 2026.

In such a world, bound to the fourth settlement and trampled under the heel of operators with the weight of capital, its allocators will similarly need to adjust to the times.

### Occident

It bears repeating that modern venture capital has always been an access business dressed up as an investing business. A shiny smile and a poorly tied tie with a term sheet in hand. The new transparency, characteristic of the fourth settlement, will bend it into a trading business. For those who can do it, it will be a more honorable line of work. And for the many venture capitalists now put out of work, this too will be more honorable.

Over the last decade, the returns in venture went to networks and those clear-eyed and unimaginative enough to blindly write large checks into them, rightly assuming that the check would be paid back many times over.[^134] For all the trappings of insight, the money was always destined to pay up for a place in the Elon trade. The managers who understood this were rewarded.

[^134]: AMD, “AMD Acquires Taalas to Advance Compute Solutions for Rapidly Growing AI Inference Market,” press release, August 6, 2026. Ryan Hoover (@rrhoover), post on X, August 6, 2026.

But this is not the order of things to come. Intelligence creates transparency, and transparency will always turn investing into trading. Too many are ill-suited to the times. We expect that seed will stop being an institutional product. That was a dislocation that somehow lasted long enough to be mistaken for an asset class. Only so briefly. No more small funds tantalizing the steady stream of founders who only needed that first check. We live in a world of operators like capital, and the operators will trade like it.

One only needs to look towards the wild west of SPV promoters to see the future. Increasingly it’s clear that venture—too—has gone K-shaped. For the basket of non-premier assets, direct equity access is possible only because demand is insignificant enough to welcome it. For the real sterling stuff—Madison Avenue quality real estate—merely being able to consider yourself “an investor in” it is enough. Seemingly most SPVs marketed today have at best tangential exposure to the underlying they claim to be holding, and when they do, it’s obscured under so many layers of fees that you can barely make it out.

It’s just increasingly clear that actually holding the underlying equity is not only unimportant for the marginal SPV buyer, but maybe actually undesirable. The future may, in fact, be hyperliquid.

For venture, the surviving value will concentrate in selection networks strong enough to have a treatment effect of their own, so that their composite members are highly selected to do more than just found companies. They’ll found winning, durable companies.

As the necessary scale of capital formation required to build a durable and lasting company inside of the intelligence trade reaches and surpasses the capitalization of public markets, the network of capital providers will become smaller and with it the cast net of total players smaller still.

Large platforms are drifting away from venture as popularly understood and towards holding companies and buyouts, because they recognize that the money in the intelligence economy will be made from operating businesses. Look at Thrive, for instance. It is popularly known as a venture capital firm, but such appearances are wildly deceiving. Whatever Thrive does now, it is not venture capital.[^135] But they certainly had to pass through being a seed investor to end up earning the right to do it. You may call it king-making or special situations. The firm has no right descriptor, at least not quite.

[^135]: Kirkland & Ellis, “Kirkland Advises Thrive Capital on Formation of Thrive Holdings,” May 2, 2025.

This “final settlement” will yield to a newly financialized private market infrastructure rivalizing and soon surpassing the scale of public markets. The rectification of names in the fourth settlement is soon to be at hand.

### Orient

Meanwhile, the public markets have beaten the private markets at their own game. Leopold Aschenbrenner’s fund outperformed the entire venture industry by buying listed equities, securities that anyone could have bought at any point, while many venture books from the halcyon 2021 days still sit at around one times cost. Some of the best returns of the cycle went to someone who read the line correctly and bought what was on the screen on the basis of his faith.

More faith will be necessary soon. The third settlement papered over dismal scientific and civilizational progress by using financial wizardry, aided by a golden age of forty years of falling interest rates. We will have no such advantage at this time.

If you examine a chart of American bond yields reaching back to 1871, you will find that for every true crossing above their long-run average rate of roughly four and a half percent, a monetary regime followed that would last decades. On September 25, 2026, the ten-year Treasury yield was 5.17 percent, near a nineteen-year high and well above that long-run average.[^136] Here’s another line to look at, right there on the chart. There is much less to hope for, though. Have faith or fear; we’re looking at a forty-year bull market in bonds coming to a close.

[^136]: Kirsten Chang, “Treasury Yields Snapshot: September 25, 2026,” *Advisor Perspectives*, September 25, 2026. Topdown Charts, “History of US Bond Yields,” 1870 to the present. We owe this reading of the chart to Zack Baker. Board of Governors of the Federal Reserve System, “Market Yield on U.S. Treasury Securities at 10-Year Constant Maturity, Quoted on an Investment Basis” (DGS10), FRED, September 25, 2026. Council of Economic Advisers, *Long-Term Interest Rates: A Survey* (July 2015), 7, fig. 1.

![Line chart of long-term US bond yields, 1871-2026, against a long-run average of 4.48 percent, peaking at 15.32 percent in September 1981 and at 5.17 percent on September 25, 2026.](https://commanding-heights.com/img/us-bond-yields-1871-2026-chart-print.png)

*Long-term U.S. bond yields, 1871–2026.*

Source: Authors’ calculations from Robert J. Shiller, “Stock Market Data Used in *Irrational Exuberance*,” online data, “Long Interest Rate GS10” column. Board of Governors of the Federal Reserve System, “Market Yield on U.S. Treasury Securities at 10-Year Constant Maturity, Quoted on an Investment Basis,” GS10 (monthly) and DGS10 (daily), FRED, Federal Reserve Bank of St. Louis, accessed October 1, 2026. Earlier observations use historical bond proxies, including municipals. The average covers January 1871–August 2026.

There are bright spots in the future, however. Machine intelligence has arrived just in time to save the collective NAV as the cost of capital rises. Put bluntly, the entire private equity industry has been bailed out by the power of AI. But deliverance does not come cheap. The intelligence economy has rescued the old economy in exchange for buying it out and rebuilding it around models. This comes at a far higher cost of capital and likely little tolerance for the worthless software projects.

Because the market cannot re-rate the entire economy at once, it strikes at improvisation. A sort of mimetic function, where a nameable category or theme is bid up before it loses interest. Memory stocks ran hot for six months before falling back. Everyone likes AI transformation companies, but they’re becoming increasingly illegible as the work begins to take shape in new forms unanticipated by the unimaginative. It is the madness of crowds, somehow gripped by terror yet maintaining their decorum. People know this is the largest trade of all time, at a scale beyond any of the prior settlements. But choosing one vertical at a time is much easier to accept, and conduct, than behaving as if the entire market is up for grabs. If we’re right, then nearly everything is mispriced.

For a narrow set of companies, the market has decided there is no price too high—and for the rest, no price low enough. We expect the number of public companies to fall precipitously as functionally the entire market is at estate-sale pricing at present. It will lead to a generation of conglomerations, roll-ups, take-privates, and assorted other private market transitions.

For the challenge of the present, a new kind of private equity must be built. The trade at present is the diffusion of machine intelligence. The barriers to doing so will need to be overcome with new vehicles.[^137] We all know that the models are well past the point of being capable of reshaping much of the function of companies and the economy at large. But there is inertia to be overcome. This is where new private equity comes in. Someone will have to buy these companies, refurbish them while keeping the wheels moving, and make the necessary adjustments to the workforce.

[^137]: Karl Angelo Vidal, “Muted Exits Push Private Equity Continuation Funds to 8-Year High,” S&P Global Market Intelligence, May 22, 2026.

### Jurisdiction

Remember the sequence. Intelligence increases transparency, and transparency creates evolutionary conditions that point towards trading mechanisms. The world is flat and firms may now choose the law that governs them.

Jurisdiction has become a product with a Pareto frontier; choose one depending on your use case.[^138] Firms can be picky, while countries must make themselves attractive to whatever capital is jurisdiction-shopping.[^139] Understand that the King Abdullah Financial District is more a state of mind than a physical place in the world. As the agglomeration continues, all countries other than China and America are being sorted into increasingly distinct roles. All against all among the exarchates, which retain nominal independence but are now separated by a lack of true sovereign frontier intelligence capability. Any exarchate must operate at the dispensation, tacit or otherwise, of a sovereign, who supplies the technology necessary for remaining functions of the state and economy. There is an American tech stack and a Chinese tech stack and there will be no arbitrage or balancing between the two.

[^138]: *Saint Regis Mohawk Tribe v. Mylan Pharmaceuticals Inc.*, 896 F.3d 1322 (Fed. Cir. 2018).

[^139]: Tunica-Biloxi Tribe of Louisiana, “Tunica-Biloxi Tribe Launches the First Tribal Nation Prediction Market App,” September 18, 2026.

Such a world is perhaps in some ways familiar to students of history, because it empowers the singularly capable exarch. And if we return to the possibilities enabled by the combination of civil and martial authority? The upheaval among exarchs will be something to behold. More authoritative revivals like in San Salvador. There is only one Haiti, but there will be a multitude of Princes.

It is said that there are just six jurisdictions left in the world, and four of them exist for what could be unsympathetically called crime.[^140] Of the rest, the capital class prefers America over China for a venue. The new firms will increasingly decide jurisdiction based on their freedom to align management compensation with the value provided to the firm. Musk’s Delaware pay package was the shot heard ’round the world, and you can bet that mistake will not be made again, by any significant market participant.[^141] Legacy firms like Meta and Alphabet have been forced to resort to increasingly shameless “acquisitions” of “operating companies” to properly compensate the few employees who are actually capable of driving their businesses forward. Through the fiction of JVs and reverse triangular mergers, these firms can wire key contributors sums that would otherwise make ISS, their index buddies, certain activist investors, and the Chancery judges break out in a rage. This problem will appear increasingly antique as a new venue is erected, a rose-tinted psycho-Delaware doing nothing but supplying accommodating but legible law.

[^140]: Will Manidis (@WillManidis), post on X, September 18, 2026.

[^141]: *Tornetta v. Musk*, 310 A.3d 430 (Del. Ch. 2024). Lora Kolodny, “Tesla Says Shareholders Approve Musk’s $1 Trillion Pay Plan with Over 75% Voting in Favor,” *CNBC*, November 6, 2025. Lora Kolodny, “Musk’s $56 Billion Tesla Pay Package Must Be Restored as Court Rules Cancellation Was Too Extreme,” *CNBC*, December 19, 2025. S.B. 21, 153rd General Assembly (Del. 2025), enacted March 25, 2025.

It is useful to think about the set frontier being developed by the labs as both technological and jurisdictional. One only needs to look towards the OpenAI non-profit flip to understand that the bitter lesson might be that finding the right slices of jurisdiction to allow for corresponding drag-free organization structures is the real advantage the modern lab has.

Mobility also will change the tax base. If an exarchate taxes only machine intelligent companies that move offshore as a result, they must tax somewhere else. That targeted class could be the salaried middle class, for as long as the class has salaries, because they cannot move as easily as the firm. It will not be an explicit tax. Form and function. The function of the Neo-Dollar and its superscription showed that a tax will be collected in the form of currency spread to every corner of the earth and controlled in America.

Layered onto the question of jurisdiction is the increasingly pesky question of local cities and their ability to extract tolls. In late October 2026, Mayor Mamdani of New York extracted a 15% equity concession from New York University’s Quantum Lab. The legal justification for such a claim is dubious, but it’s not hard to imagine other world cities (London, New York, San Francisco) promising campaigns of sustained harassment for companies that don’t fall in line. If the president of the United States is mayor of the U.S. and the dictator of the rest of the world—then the only American Dictators left are our mayors.

If the frontier of crime, the “prediction markets” née sportsbooks, is any indication, questions of jurisdiction will become increasingly slippery. Maybe by 2035, the only lawyers left practicing will be practicing Tribal law.

At the end of this geographical trade is a coalition between the firms and a favorite exarchate that freezes its legal regime, perhaps constitutionally, and raises a subscription to participate in it. Imagine productized Delaware, in the eternal good old days of 1993, and there are robots everywhere. Even the idea of a constitutional provision is not entirely suited to the new role of the jurisdiction. It will develop in degrees, the first big wave being model-driven private arbitration forums, which will displace the courts as a venue for many disputes, a further privatization of state capacity. For these new firms, the ideal regulator looks less like an act of Congress or the rulemaking of an executive agency, and more like a machine intelligent Lloyd’s of London, a market of underwriters pricing risk and refusing whatever they cannot profitably write.

### The Vanishing Point

Something strange happens at the vanishing point, where from a sufficient distance the lines between capital and sovereign power become indistinct.

We believe that the latitude afforded to firms is correlated to the security felt by the sovereign. When the sovereign feels secure, without existential threats at home or abroad, businesses are given greater leeway and the profit motive is the primary determinant of firm behavior. In this way, the separation between firms and states that characterized the post-Cold War period was very much an exception, contingent on fleeting historical circumstances, rather than the rule. We are well past it, even if there has been no act of Congress nor a Presidential memorandum.

![Bubble chart of federal stakes in 35 companies from June 2025 to September 2026, by date announced and sector, sized by dollars committed, led by Intel at $8.9 billion for 9.9 percent.](https://commanding-heights.com/img/federal-stakes-chart-print.png)

*Federal stakes in companies, June 2025 to September 2026.*

Source: Stakes taken by the Commerce, Defense, Energy and Treasury departments, by first announcement, from company filings and agency releases. Intel Corporation, “[Intel and Trump Administration Reach Historic Agreement to Accelerate American Technology and Manufacturing Leadership](https://www.sec.gov/Archives/edgar/data/50863/000005086325000129/a08222025form8-kex991.htm),” press release, August 22, 2025. U.S. Department of Commerce, “[Department of Commerce Announces Letters of Intent with 9 Companies for $2 Billion to Accelerate U.S. Leadership in Quantum Computing](https://www.nist.gov/system/files/documents/2026/05/21/Quantum%20CHIPS%20Press%20Release-2026-05-21-Updated_Final.pdf),” May 21, 2026, and “[Department of Commerce Announces Letters of Intent with 7 Companies for $874 Million to Accelerate Semiconductor R&D for the Compute Supply Chain](https://www.nist.gov/system/files/documents/2026/07/29/CHIPS%20Compute%20Supply%20Chain%20Press%20Release%202026-07-29_Final.pdf),” July 29, 2026. Ahn Sung-mi, “Korea Zinc to Build $7.4b Critical Minerals Smelter in Tennessee,” *Korea Herald*, December 15, 2025. Several stakes have not closed, including those announced in letters of intent. Dollars are as announced, including converted grants, and Korea Zinc’s $1.4 billion is conditional. Other counts differ. Tad DeHaven and Chad Smitson, “Trump Administration’s Equity Portfolio Reaches 34 Deals,” *Cato at Liberty* (blog), Cato Institute, October 2, 2026, counts 34 companies. Jonathan E. Hillman, “Washington’s Growing Portfolio: Tracking U.S. Government Investments,” Council on Foreign Relations, accessed October 4, 2026, counts 39 deals since January 2025, including the Development Finance Corporation’s.

Regulators have revealed a preference in how they metabolize bubble unwindings. You might blame these agents of the state for having only sentenced one man to prison during the Global Financial Crisis, but that misses the point. The financial system could not be allowed to fail, but more operatively, its firms had the power to tithe non-core participants to the regulators. The properly positioned players grew stronger from the settlement, and the vulnerable firms were gobbled up or left to rot. Bobbie Lehman was long dead. The scapegoats, saving their constituencies, variously equity and credit and bank and insurer, are ceremonially beheaded with the assent of the sovereign. This blood offering is a tithe to the politic, the cost paid to avoid a terminal event. Progress may then continue.

As each cycle develops, certain firms enjoy sustained outperformance riding the frontier. Be ye a frontier lab, crypto, FAANG, financial, dotcom, junk-bond, oil patch, or Nifty, if you outperform too long, eventually your cycle will carry you to the sun. You will become the face of your epoch, public enemy No. 1. They simply had to put Mike Milken away; he had done it too well, taken it too far. And in the settlement, his inventions were distributed widely, and at the cost of twenty-two months at Pleasanton, the alternative investments industry renewed its progress, growing to a size perhaps even he might not have imagined at the time. His former colleagues can attest.

Firms associate into industry to hedge these events. It happens time and time again, in defense, pharma, finance—pick your contested industry. When the firms are called to account, the pack will eject the weak, so that the core may survive. This pre-Christian truth has proven itself eternal. You can only wonder how many credit agreements allow force majeure by Girardian scapegoating.

The sovereign will afford protection to its peculiar labs. But protection always has a price, and you are far too small-minded if you believe that it will be paid in taxes. Labs want federal preemption of state laws, favorable permitting and copyright regulations. They want, somehow, to continue pushing at the frontier while their artifact is being commoditized and published daily by Chinese labs.

The larger the protection, the larger the bill. But there are precedents for actions of this importance. In 1947, the Atomic Energy Commission took ownership of the Manhattan Project’s plants and laboratories.[^142] Along with those, the commission assumed the project’s mission. It was out in the open, sanctioned by court functionaries, blessed in the public eye.

[^142]: Executive Order 9816, December 31, 1946, §§ 1–3, 10, 12 Fed. Reg. 37.

For the same reason that we prefer mechanical Delaware as the ideal corporate jurisdiction, we much prefer the activity of the sovereign to be legible, rather than whispered. The openness of the American sovereign, litigated not in courts but in public and among voters, is still a preferable environment for sovereign action when the alternative is Chinese. The Chinese system is actively hostile to this transparency. If you accept that sovereign power must be exercised, then it is good to steer it in the proper direction if possible.

What path you believe this sovereign action takes depends heavily on your ideological tribe, and increasingly your dining preference. The Butterworth’s crowd, young staffers, virile and stimulant-addled, can barely keep the words “Defense Production Act” and “Youngstown Steel” out of their mouths long enough to make it through their burger. In their minds, the sovereign will approach nearly Singaporean capacity to compel a detailed program of compute offtake and allocation from the labs to release pressure from the system and smooth token prices before some kind of hackneyed nationalization under a democratic administration.

The Cafe Milano crowd prefer much more nuanced regulatory schemes. Trump Accounts, and their imminent expansion to include all pre-retirement Americans, provide a convenient mechanism to pass—voluntary or not—large chunks of frontier equity to the American taxpayer. Pseudo-privatization of Social Security is a welcome and unexpected side effect of such regimes.

Left-of-center elements, now displaced as war tribes originally far outside of the walls of the imperial core (Beltway), move much closer. Centrolina, Cosmos Club, and the halls of Bushwick Bars and Brad Lander fundraisers are haunted by talks of a “private/public synthesis.” Much more comfortable with informal regulatory agreements paired with aggressive taxation schemes. The state-capacity-establishment-left will settle for infiltrating the labs with thousands of would-be Obama staffers as a modern Project Snow White, while the DSA-left will tax and regulate the entire thing out of existence.

We think that the most likely form of sovereign action will be a license for frontier training runs and a form of nationalization,[^143] though as we have previously stated, it could be but a single golden share.[^144] The peculiar lab will house the research and development of the sovereign frontier, which will be run like a cost-plus program. The lab splits into two firms under one sovereign roof. On the one hand is a commercial company, which makes and meters intelligence at whatever margin the Chinese floor allows. On the other hand is a frontier lab with a line direct to the Oval Office.

[^143]: *Youngstown Sheet & Tube Co. v. Sawyer*, 343 U.S. 579 (1952).

[^144]: Nippon Steel Corporation and United States Steel Corporation, “Nippon Steel Corporation and U. S. Steel Finalize Historic Partnership,” press release, June 18, 2025.

The antichrist is in the details. It is still unclear what kind of settlement between the firm and state is emerging, beyond what has been outlined above. Are we in eighteenth-century England, where the state provided the security of last resort and ultimate authority, but left most other functions for firms to carry out? Or are we entering Buyid Persia, where rival sovereigns warred and collected their rents through concessions afforded to whoever held the ground at that given point? From afar, quite so far from the vanishing point, the two look much alike. Up close, the difference is found in whether the firm is a subject of the state or a partner in the dispensation of limited power.

The Anglomodern answer to this question, which is to say the old English answer, is that the sovereign deigns to allow the private firms to service it, while retaining its final and ultimate authority. This, for now, has been the American answer to the question of state capacity and authority. Ask those in the know just how much state capacity already resides inside Goldman Sachs, Citadel, and Apollo. If the remit of this arrangement continues to expand, we will have a Palantir-style future on our hands. The vision of cumulative state capacity will be expressed daily, in the shape of a firm. The idea as the firm, the firm as its management and capital fused into it and the same small group of men. There is no name for it yet, but its function has already found its place.