Scaling LLC

Side wing · off Room 9, the Old Growth Gallery · optional

The Old World Rooms

Older companies, older money, and no shared scale.

Seven rooms: a Dutch book of 1602, an English charter, a guilder peak, Paris and London in 1720, and two modern listings that set a price on a sliver of a company. Nothing in here is converted into today’s money.

OW01The Case With No Scale
Period case, emptyA glass case with no scale on it, lit and empty. There are no ticks, no baseline and nothing to read against.Period case ·demonstration

This case shows what was written down, in the money it was written in, on the day it was written. It does not show what that would be worth now. No lens in this building reaches back this far, and the one we have is American and begins in 1913.

Old World · room 1 of 7

The Case With No Scale

This wing was built first, and it shows. Varnished boards, brass fittings, a lamp over every case. Nothing in here is running.

There are no poles in these rooms for a while. A pole needs a price, and a price needs somebody standing there every day, in public, willing to pay it. For most of the history of large companies, nobody was.

So the old wing uses cases. A case has no scale on it. You can read what is inside one; you cannot hold two of them up against each other. The first one is empty, so that you notice what is missing before anything is put in front of you.

The Old World is a geography here, not an age. The last two rooms are Chinese and Saudi, and their dates are 2007 and 2019.

THE INSPECTORWhere is the scale?

THE LEDGERThere was no price to put on one.

The caretaker“Rulers don’t reach.”

What this room showsFor most of history there was no daily public price for a company, so there is nothing for a ruler to measure.

No figures in this room.

The second case is not empty. It holds a book.

OW02The Book Closed at Midnight
The Dutch East India Company, drawn as a spice ship grown as a plantOW-02
Capital paid in, Amsterdam chamber, 1602 — capital tank, in guildersCapital paid in: 3,674,945 guilders. Money paid in, shown filling a tank once. Not a valuation. Basis: 31 August 1602, subscription register.01m2m3m4mƒ 3,674,945guildersCapital paid in
Capital tank, graduated in millions of guilders. It filled once.
Empty mount: restitution of capitalA plinth with a clean, machined socket and nothing in it. The engraved label reads: Restitution of capital, due 1612. The charter allowed shareholders their stake back after ten years. It was never returned, and the States General supported the directors in not returning it.Restitutionof capitaldue 1612
Empty mount · Restitution of capital The charter allowed shareholders their stake back after ten years. It was never returned, and the States General supported the directors in not returning it.

Amsterdam chamber, subscription register, closed at midnight on 31 August 1602: ƒ 3,674,945 from 1,143 subscribers. Money paid in, in guilders. Not a valuation.

All six chambers: almost ƒ 6.5 million.

Old World · room 2 of 7

The Book Closed at Midnight

In August 1602 a book lay open in Amsterdam and anyone could put their name in it. Its first page is six directors at twelve thousand guilders each. Further in, Dirck van Os’s maid, Neeltgen Cornelis, put in a hundred guilders she had worked a long time for. The last line reads: Barent Lampe for Dignum Jans, fifty guilders. The national archive lists her as a maid.

At midnight on 31 August the notary Jan Fransz Bruyningh wrote down the total, 3,674,945 guilders from 1,143 subscribers, and recorded that nothing had been paid in after twelve o’clock. Across all six chambers it came to almost six and a half million. Amsterdam was fifty-seven per cent of it.

It is in a tank, not on a pole, and it stays in guilders. This is money paid in. It is not what anyone thought the company was worth, because nobody was yet in a position to say.

The charter said the shareholders could have their money back in 1612. The Dutch national archive puts what happened plainly: after ten years there was no return of capital, and the original stake stayed unchanged for the life of the company. The States General backed the directors.

So the mount beside the tank is empty, and what is missing from it is the refund. If you cannot get your money out, the only way out is to sell your share to somebody else. A market grew in the space where the refund had been.

THE LEDGERƒ 3,674,945.

THE PRICEPaid in. Not a price.

The caretaker“Socket’s clean. Nothing ever sat in it.”

What this room showsMoney paid into a company is not a price for it, and a share you cannot redeem needs a market to sell it in.

Check the readings (2)

Each figure keeps its own money, date, basis and sources. Nothing here is converted unless the line says so.

  1. ƒ 3,674,945

    Capital paid in / nominal stock · 1602-08-31 · basis: period-record · confidence: high

    The Amsterdam chamber’s subscription total as the notary Jan Fransz Bruyningh recorded it when the register closed at midnight on 31 August 1602: 3,674,945 guilders from 1,143 subscribers. Money pledged into the company, not a price for it. Read from the historian of the register (Petram), who quotes the notary’s closing entry. Reference works print a slightly higher Amsterdam figure (ƒ 3,679,915); the sources opened do not explain the difference, and the card carries the notary’s midnight tally because that is the document the room is about.

  2. ƒ 6,429,588

    Capital paid in / nominal stock · 1602-08 · basis: period-record · confidence: medium

    Subscribed capital across all six chambers (Amsterdam, Zeeland, Enkhuizen, Delft, Hoorn, Rotterdam). Petram gives 6,429,588 guilders and describes it as “almost 6.5 million”, which is how the narration says it. The widely printed ƒ 6,424,588 does not reconcile: the six chamber totals printed beside it in the same reference works add up to 6,429,588.

    Still to check: The reconciling figure rests on a historian’s blog and on adding up chamber totals printed in secondary reference works. The narration only says “almost six and a half million”, which both totals support. To lift it to high, open Gelderblom, De Jong and Jonker (JEH 2013) or Den Heijer (2005) and read the chamber table.

Across the aisle, England had tried this two years earlier and built it differently.

OW03Sealed Voyages
The English East India Company, drawn as a rack of separate sealed voyagesOW-03
Empty mount: equity valueA plinth with a clean, machined socket and nothing in it. The engraved label reads: Equity value, c. 1610 · none. Each voyage was subscribed and settled on its own, so there was no company-wide stock and nothing continuous to price. A pole was never fitted here.Equity valuec. 1610· none
Empty mount · Equity value Each voyage was subscribed and settled on its own, so there was no company-wide stock and nothing continuous to price. A pole was never fitted here.

Old World · room 3 of 7

Sealed Voyages

England got there first. On the last day of 1600 Elizabeth I granted a charter to the Governor and Company of Merchants of London, trading into the East Indies.

Then it built something else. Its early voyages, from 1601 to 1612, were separate: the subscribers to each voyage bore its cost and took its profits, and the next voyage was subscribed again.

That is why there is a rack here and not a pole. Every pod is one voyage, and every pod is sealed, because each was settled and closed.

Asking what this company was worth in 1610 is like asking for the market value of one film’s financing. There is no wrong answer. There is no answer, so the socket where a pole would stand has never had one fitted.

Two companies, two years apart, in the same ocean. Only one was a thing you could own continuously and sell to a stranger. The English voyages were pooled into joint stock after 1612, and a single joint stock for the whole trade came with Cromwell’s charter of 1657.

THE LEDGEROne voyage, one ledger. No stock to price.

The caretaker“Never fitted one.”

What this room showsA company made of separately financed voyages has no single stock, so there is nothing continuous to price.

No figures in this room.

Back across the aisle, the Dutch one is still standing, and now it has a price.

OW04Seventy-Seven Million Guilders

The VOC’s case

The Dutch East India Company, drawn as a spice ship grown as a plant, fully ladenOW-02
Guilder rule inside the VOC’s caseGuilder rule, wood, graduated in millions of guilders from 0 to 80 million. It reads just over 77 million guilders, marked peak · 1720. Basis: a historian’s multiplication of a quoted price by the nominal capital. It is not converted to any other money.01020304050607080millions of guildersjust overƒ 77 millionpeak ·1720
Wood, in guilders. It does not leave this case.

On the wall · documented absence

7.9 trillion US dollars, as it circulates

The Dutch East India Company “was worth 78 million Dutch guilders, which translates to a whopping $7.9 trillion in modern dollars.”

Not a measurement. A guilder figure attached to 1637 rather than 1720 and converted by a method the page never states: no rate, no index, no working. Its own inputs imply about $101,000 per guilder ($7.9 trillion ÷ 78 million, this wing’s arithmetic). The historian of the price series puts the properly converted figure a factor of about 8,000 lower.

Peak market capitalisation, 1720: just over ƒ 77 million. A quoted price of about 1,200 per cent, applied to the nominal capital. A bubble print, briefly traded, in guilders.

Old World · room 4 of 7

Seventy-Seven Million Guilders

This is the one specimen in the old wing that can be measured, and the Dutch are why. Its shares were quoted as a percentage of what had been paid in, and the capital behind them barely moved after 1602, apart from a few small adjustments. So a price times the capital means something.

The rule is wood, graduated in guilders, and it never leaves this case. It reads just over 77 million, in 1720, when the shares briefly traded at around 1,200 per cent. That is a peak, and the rule says so.

The card on the wall says seven point nine trillion dollars.

That number is not a typo. It is a different kind of object: seventy-eight million guilders, attached to 1637, turned into dollars by a method the page never states. Run it backwards and it implies about a hundred thousand dollars for every guilder.

The historian who assembled the price series ran his own conversion and got about 820 million euros, which he calls a factor of eight thousand. That figure is not on the rule either. It is offered to retire a number, not to install one.

The wrong year is the part worth keeping. In early 1637, at the height of the tulips, the shares stood around 270. The peak was 1720, the year of Paris and London. The story picked the more famous bubble.

THE PRICEI am $7.9 trillion.

THE INSPECTORThe page never states the conversion.

The caretaker“People photograph the card.”

What this room showsA famous number with no stated method is not a measurement; the VOC’s real peak was just over 77 million guilders, in 1720.

Check the readings (5)

Each figure keeps its own money, date, basis and sources. Nothing here is converted unless the line says so.

  1. just over ƒ 77,000,000

    Equity value · 1720 · basis: reconstruction · confidence: high

    The VOC’s peak market capitalisation: “just over 77 million guilders”, in 1720. A historian’s multiplication of a quoted price (“briefly traded at around 1,200”, per cent of the amount paid in) by a nominal capital that barely moved after 1602; 1,200 per cent of ƒ 6,440,200 is ƒ 77.3 million, which is why the source can say “just over”. The only pre-modern figure in the building that is a genuine equity value, and so the only one allowed a calibrated scale: the wooden guilder rule, in guilders, marked PEAK. It is a briefly traded bubble print, not an average.

  2. 1,200 per cent of nominal

    Proportion (no currency) · 1720 · basis: period-record · confidence: high

    VOC shares “briefly traded at around 1,200” in 1720: twelve times the amount originally paid in. VOC prices were quoted as a percentage of nominal, so this is a ratio, not a sum of money, and a quoted “1,200” is not a per-share figure in guilders (guardrail 5).

  3. ƒ 6,440,200

    Capital paid in / nominal stock · 1693 · basis: reconstruction · confidence: medium

    The VOC’s nominal capital after “a few adjustments”, as it stood from 1693. It is the reason the narration says the capital “barely moved” rather than never moved: the national archive says the original stake remained unchanged, Petram says “roughly the same”, and this is the figure behind the “roughly”. Nominal capital, not a value.

    Still to check: One secondary reference site, corroborated by a Wikipedia citation to Den Heijer (2005) p. 60 that was not opened. Open Den Heijer before anything is built on the exact figure; the narration uses it only to say “barely moved”.

  4. €820,000,000

    Equity value · 2020 · basis: reconstruction · confidence: medium

    Petram’s own conversion of the 1720 peak through a published historical price tool: “approximately €820 million (just short of $1 billion)”, which he calls “a factor 8,000 less than reported”. NARRATION ONLY (guardrail 2). It is counter-evidence to a wrong number, not an answer: a single consumer-price conversion across three centuries is exactly what this building declines to do on an instrument.

    Still to check: Deliberately capped. The figure is correctly quoted, but the method (a price index across three centuries) is one docs/08 §1.1 rejects for instruments. It may only appear in narration, as the historian’s estimate, and never on the rule.

  5. 270 per cent of nominal

    Proportion (no currency) · 1637 · basis: period-record · confidence: high

    “At the height of the tulip craze, in early 1637, VOC shares traded around 270” — 2.7 times the amount paid in. The year the popular figure attaches the peak to, and not the peak.

1720 was a busy year. It was busier in Paris.

OW051720

Paris: about 10,000 livres a share at the turn of 1719–20; the file’s high is 10,100 on 8 January 1720. The company’s own bid made its capitalisation up to 4.96 billion livres. That is an argument, not a quotation.

London: £950 per £100 of stock, 10 July 1720, the London list’s maximum. Nominal stock in March 1721: £37,802,203, of which £13,301,781 was held by the company itself. Dates converted from Old Style.

Under the Asiento the South Sea Company held a contract to deliver enslaved Africans to Spanish America. It made ninety-six such voyages. Just over thirty-four thousand people were put aboard. Just under thirty thousand came off.

Old World · room 5 of 7

1720

Paris first, because Paris went first. A share in the Mississippi Company stood around 250 livres in 1718. On 8 January 1720 it was quoted at 10,100.

The bank that issued the notes was run by the same people, and in February 1720 it was merged into the company. Its notes went from 38 million livres in April 1719 to 2,359 million in June 1720. From October 1719 the company was buying its own shares; the support was halted on 22 February, and on 5 March the price was fixed at 9,000.

An economist at the Chicago Fed has since done the sums the period did not. On the company’s revenues he put a defensible price at 1,875 livres, a fifth of the peg; revisiting them later, at 2,660 to 2,860.

London is smaller and worse. South Sea stock opened the year at £128 and change, reached £950 on 10 July, and was £180 by 9 October. London wrote those dates eleven days earlier; this building uses one calendar.

There is no market capitalisation on either side, and the empty mount between them is why. The price peak is July 1720; the capital, £37.8 million of nominal stock, is March 1721, after seven subscriptions whose stock had not been issued in July. Two different companies. A price is a fact. A market capitalisation is an argument.

A Treasury memorandum from before the scheme survives. It describes a company that did “depend chiefly if not intirely on their annuity out of the Exchequer and their trade is so small and doubtfull as to the profit”. The scheme went ahead.

The chamber behind the London case is open. Its plaque says what it held.

What this room showsA price is a fact and a market capitalisation is an argument; in 1720 the price and the capital belonged to different moments.

Check the readings (17)

Each figure keeps its own money, date, basis and sources. Nothing here is converted unless the line says so.

  1. 10,100 livres per share

    Price per unit of stock · 1720-01-08 · basis: period-record · confidence: medium

    The highest quotation in the Yale 1720 file for the Mississippi Company, 10,100 livres on 8 January 1720 (Gregorian, as Paris dated it), from Murphy’s series. A price for one share, not a value for the company. Velde’s own papers put the peak at “10,000L in January 1720” (2003) and at 9,525 livres on 2 December 1719, “possibly close to 10,000L just before Christmas” (2014 draft).

    Still to check: The quotation is read correctly from the file, but the file’s own authorities disagree about the exact peak by a few hundred livres and a few weeks. The card prints the file’s figure and date; the plaque says “about ten thousand”. Leave it at medium.

  2. £950 per £100 of nominal stock

    Price per unit of stock · 1720-07-10 (London, Old Style: 1720-06-29) · basis: period-record · confidence: high

    The London list’s 1720 maximum, £950 per £100 of stock, quoted on 29 and 30 June 1720 as London dated them (Old Style), 10 and 11 July in the Gregorian calendar this building uses. A price per unit of stock, not a value for the company (guardrail 5). The famous four-figure number is real but Amsterdam’s: the Leydse Courant has 1,100 on 6 July 1720.

  3. up to 4,960,000,000 livres

    Equity value · 1720-02 · basis: reconstruction · confidence: medium

    An UPPER BOUND and an argument, not a quotation: 300,000 original shares at 9,925 livres — the price the company itself was paying for them — plus “up to 300,000 subscriptions” valued at 6,600, after Du Hautchamp (1743), as Velde gives it. The price in this multiplication was the company’s own bid.

    Still to check: Correctly read, but the count of subscriptions is itself a ceiling (“up to”), so the product can only be shown as “up to”. Keep it off every instrument.

  4. £37,802,203

    Capital paid in / nominal stock · 1721-03 · basis: reconstruction · confidence: high

    Existing South Sea stock in March 1721, after the 1720 subscriptions and conversions: £37,802,203 nominal, of which £13,301,781 was held by the company itself (Velde 2025, Table 9). Nominal stock, not money paid in at market and not a value. It belongs to March 1721; the price peak belongs to July 1720, when most of this stock had not been issued. That gap is why no market capitalisation is computed for the South Sea Company in this building.

  5. 250 livres per share

    Price per unit of stock · 1718 · basis: period-record · confidence: high

    “During the year 1718, the price of a share in the Company of the West stood around 250L.” Approximate, as the source says.

  6. 38,000,000 livres

    Banknotes in issue (narration only) · 1719-04-30 · basis: reconstruction · confidence: high

    Banque Royale notes issued less notes burned, 30 April 1719 (Velde, Table 4). A monetary liability, not capital: narration only.

  7. 2,359,000,000 livres

    Banknotes in issue (narration only) · 1720-06-30 · basis: reconstruction · confidence: high

    Notes issued less notes burned, 30 June 1720 (Velde, Table 4): 2,359.0 million livres. Narration only.

  8. 5,000 livres per share

    Price per unit of stock · 1719-10-05 · basis: period-record · confidence: high

    On 5 October 1719 the company ordered its treasurer to buy any share offered at 5,000 livres: the start of the support operation.

  9. 9,000 livres per share

    Price per unit of stock · 1720-03-05 · basis: period-record · confidence: high

    On 5 March 1720 the company opened an office buying and selling shares at a fixed 9,000 livres, after price support was halted on 22 February.

  10. 1,875 livres per share

    Price per unit of stock · 1720 · basis: reconstruction · confidence: high

    Velde (2003): annual revenues of 75.5 million livres at a factor of 15 give 1,132 million, “or a share price of 1875L, which is one fifth of the peak share price of 9000L”.

  11. 2,660 livres per share

    Price per unit of stock · 1720 · basis: reconstruction · confidence: medium

    Velde’s revision (2014 draft): “a share price of 2660–2860L … overvalued by a factor of at least three.” Lower end.

    Still to check: An unpublished draft on a third-party course page. Quoted as the author’s revision; do not promote it over the 2003 working paper without the published book.

  12. 2,860 livres per share

    Price per unit of stock · 1720 · basis: reconstruction · confidence: medium

    Upper end of the same revised range.

    Still to check: As the lower end: an unpublished draft.

  13. £128.375 per £100 of nominal stock

    Price per unit of stock · 1720-01-12 (London, Old Style: 1720-01-01) · basis: period-record · confidence: high

    The first London quotation of 1720 in Castaing’s list, £128⅜ per £100 of stock. Dated 1 January 1720 in London (Old Style); 12 January in this building’s calendar.

  14. £180 per £100 of nominal stock

    Price per unit of stock · 1720-10-09 (London, Old Style: 1720-09-28) · basis: period-record · confidence: high

    £180 per £100 of stock on 28 September 1720 as London dated it, 9 October in the Gregorian calendar.

  15. 96

    Count · 1713 · basis: reconstruction · confidence: high

    Slaving voyages made by the South Sea Company under the Asiento: “96 voyages in all”, London-origin sailings counted from the Eltis et al. trans-Atlantic slave trade database.

  16. 34,000

    Count · 1713 · basis: reconstruction · confidence: high

    People put aboard on those voyages: “just over 34,000”. A lower bound, stated as the source states it.

  17. 30,000

    Count · 1713 · basis: reconstruction · confidence: high

    People landed alive: “just under 30,000”. An upper bound, stated as the source states it.

The next room is in Shanghai, in 2007.

OW06Two Prices, One Company
PetroChina, Shanghai price × every share — valuation poleEquity value: $1.06 trillion. Read on a linear pole calibrated from $0 to $1.2 trillion (US dollars, shared). Basis: Shanghai price × every share, November 2007, this building’s arithmetic. The pole is built in 7 sections, one for each order of magnitude of its range.$0$200B$400B$600B$800B$1T$1.2T$1.06TEquity value$1.06 trillion0 – $1.2TUS dollars,shared
Shanghai price × every share
PetroChina, drawn as an oil organism from well to forecourt, with the small traded part of its shares lit on the plinthOW-06
PetroChina, New York price × every share — valuation poleEquity value: $467 billion. Read on a linear pole calibrated from $0 to $1.2 trillion (US dollars, shared). Basis: New York price × every share, November 2007, this building’s arithmetic. The pole is built in 7 sections, one for each order of magnitude of its range.$0$200B$400B$600B$800B$1T$1.2T$467BEquity value$467 billion0 – $1.2TUS dollars,shared
New York price × every share

Shanghai: the November 2007 high close of RMB 43.96 × 183,020,977,818 shares = RMB 8,045,602,184,879; at the filing’s 2007 average rate of RMB 7.5806 to the dollar, about US$1.06 trillion.

New York: the November 2007 high close of US$255.06 per receipt for 100 shares, × the same shares = about US$467 billion.

Both products are this building’s arithmetic from the filing’s own inputs. Monthly high closes: same month, not necessarily the same day. The Shanghai float was 2.18 per cent of the shares.

Old World · room 6 of 7

Two Prices, One Company

5 November 2007. PetroChina lists four billion new A shares in Shanghai, issued at 16.70 yuan. They close the first day at 43.96.

Apply that price to every share the company has and you get about 8.05 trillion yuan. The wires called it the world’s first trillion-dollar company.

Here is the company’s own filing. Shares after the issue: 183,020,977,818. The state parent held 86.29 per cent of them. The shares actually trading in Shanghai were the four billion new ones, 2.18 per cent: one share in every forty-six.

The same company was also trading in New York, in dollars, as receipts for its Hong Kong shares. At New York’s best close that month, the same share count gives about 467 billion dollars. At Shanghai’s best close, converted at the filing’s own average rate for 2007, it gives about 1.06 trillion.

Same company. Same month. Two filed prices, and a factor of 2.27 between them. Both poles are calibrated alike, and both are reading honestly.

A price is what somebody paid for the shares in front of them. It is the lamp maker’s hundred-share trade again, done in two cities at once, and this time the filing prints the whole of it.

THE PRICEAbout $467 billion.

THE LEDGERAbout $1.06 trillion.

THE INSPECTOROne share count. A factor of 2.27.

The caretaker“Both poles are true.”

What this room showsA market capitalisation is one price applied to every share; where only a sliver trades, one company can carry two very different ones.

Check the readings (8)

Each figure keeps its own money, date, basis and sources. Nothing here is converted unless the line says so.

  1. $1,061,341,079,186

    Equity value · 2007-11 · basis: reconstruction · confidence: high

    THE ONE EXCHANGE RATE IN THIS WING, AND A LABELLED APPARATUS CHANGE. RMB 8,045,602,184,879 divided by RMB 7.5806 per US$ — the 2007 average noon buying rate stated in the same 20-F (an average of month-end rates) — gives US$1,061,341,079,186. Converted so that both readings can stand on two identically calibrated brass poles; the conversion is contemporary and the filing’s own, never a restatement across time. The dollar headline of the day (“first $1 trillion company”) was a press convention built this way.

  2. $466,813,306,023

    Equity value · 2007-11 · basis: reconstruction · confidence: high

    THIS WING’S MULTIPLICATION, printed in no source: the November 2007 high ADS close, US$255.06 per 100 H shares, times all 183,020,977,818 shares = US$466,813,306,023. Already in dollars; no conversion. Both poles use monthly high closes, which need not be the same trading day: “same month”, never “same day”.

  3. RMB 8,045,602,184,879

    Equity value · 2007-11 · basis: reconstruction · confidence: high

    THIS WING’S MULTIPLICATION, printed in no source: the November 2007 high A-share close, RMB 43.96, times all 183,020,977,818 shares = RMB 8,045,602,184,879. The Shanghai price extended to every share, including the 97.82% that were not trading in Shanghai. The primary reading of the room, in the money it was struck in.

  4. RMB 43.96 per A share

    Price per unit of stock · 2007-11 · basis: public-market-cap · confidence: high

    The highest A-share closing price in November 2007 on the Shanghai Stock Exchange, RMB 43.96, from the 20-F’s market price table. The press reports the same figure as the first day’s close, 5 November. A closing price, not an intraday high.

  5. $255.06 per ADS (100 H shares)

    Price per unit of stock · 2007-11 · basis: public-market-cap · confidence: high

    The highest closing price of PetroChina’s New York ADSs in November 2007, US$255.06, each ADS representing 100 H shares.

  6. 183,020,977,818

    Count · 2007-12-31 · basis: period-record · confidence: high

    Total share capital after the A-share issue: 183,020,977,818 shares. CNPC 157,922,077,818 (86.29%); public A shares 4,000,000,000 (2.18%); H shares 21,098,900,000 (11.53%). A count from the filing, not a money figure.

  7. 2.18%

    Proportion (no currency) · 2007-12-31 · basis: period-record · confidence: high

    The public A-share float as a share of all PetroChina shares: 4,000,000,000 of 183,020,977,818, “approximately 2.18%”. FLOAT, not class: after the issue CNPC’s shares were also registered as A shares, so the A-share class was 88.47% of capital. The 2.18% is the part a Shanghai buyer could actually trade.

  8. RMB 16.70 per new A share

    Price per unit of stock · 2007-11-05 · basis: offering-size · confidence: high

    The issue price of the 4,000,000,000 new A shares, RMB 16.70, listed in Shanghai on 5 November 2007.

The last room sold one and a half per cent, and stands beside the offering the route ends on.

OW07Two Trays
Saudi Aramco, drawn as an oil field grown as a plant: wellhead, separators, pipelines and a tanker at the berth, with the small offered part of its shares lit on the plinthOW-07
Saudi Aramco, December 2019: offering tray, two compartmentsSaudi Aramco · December 2019 · riyals. To the company: nothing; this compartment is empty. To selling holders: 96 billion Saudi riyals. Counted in blocks, one block = SAR 10 billion · riyals, not to scale with the dollar tray. Basis: gross proceeds of the offering.To the companyemptyNothingTo selling holdersSAR 96 billionSaudi Aramco · December 2019 · riyalsone block = SAR 10 billion · riyals, not to scale with the dollar tray
SpaceX, June 2026: the route’s booster specimen, dressed for its listing, with bunting and the bell
SpaceX, June 2026: offering tray, two compartmentsSpaceX · June 2026 · US dollars. To the company: 86.2 billion US dollars. To selling holders: nothing; this compartment is empty. Counted in blocks, one block = US$10 billion · dollars, not to scale with the riyal tray. Basis: gross proceeds of the offering.To the company$86.25 billionTo selling holdersemptyNothingSpaceX · June 2026 · US dollarsone block = US$10 billion · dollars, not to scale with the riyal tray

Saudi Aramco, base offering: 3,000,000,000 existing shares at SAR 32, all sold by the government; SAR 96 billion to the seller, nothing to the company. With the over-allotment option the government sold 3.45 billion shares in all, 1.73 per cent.

SpaceX: 638,888,888 newly issued shares at $135; $86.25 billion gross to the company, nothing to selling holders.

Both gross. Different money, different years, no shared scale.

Old World · room 7 of 7

Two Trays

The route ends at a bell: SpaceX in June 2026, every share in its offering newly issued, all of the money to the company and the selling holders’ compartment empty. That tray is here. The one beside it is from six and a half years earlier, and it loads the other way round.

December 2019. Saudi Aramco sells three billion shares at 32 riyals. The prospectus says it twice, in case anyone misses it: the share capital will not change, and the number of outstanding shares will not change. The seller is the government, which owned all of it. “Accordingly, Saudi Aramco will not receive any of the proceeds of the Offering.”

Thirty-two riyals across all two hundred billion shares implies 6.4 trillion riyals, about 1.71 trillion dollars at the peg the prospectus prints. That price was agreed for one and a half per cent of the company, before there was a market for any of it.

Two trays. One compartment full on each, and it is not the same compartment. Each is counted in its own money, riyals and dollars, and they are not to scale with each other. What they compare is which side is empty.

At the end of 2024, by the company’s own count, the public held 2.46 per cent of Aramco. The old wing has seen this before: a price struck on a sliver, and a number for the whole.

THE INSPECTORAsk who was paid.

The caretaker“Dust both. Fill neither.”

What this room showsAn offering’s size says how many shares were sold, not who was paid; Aramco’s money went to the government, SpaceX’s to the company.

Check the readings (8)

Each figure keeps its own money, date, basis and sources. Nothing here is converted unless the line says so.

  1. SAR 0

    Offering proceeds · 2019-12-11 · basis: offering-size · confidence: high

    Zero, in the prospectus’s own words: “Accordingly, Saudi Aramco will not receive any of the proceeds of the Offering.” Share capital and share count did not change.

  2. SAR 96,000,000,000

    Offering proceeds · 2019-12-11 · basis: offering-size · confidence: high

    Total offering amount: 3,000,000,000 shares at the final price of SAR 32, all sold by the Government. The prospectus’s own figure, “excluding any proceeds resulting from the exercise of the Over-allotment Option” — the BASE offering. With the option the Government sold 3.45 billion shares in all (1.73% of the share capital); this wing does not print a proceeds total for that, because none was found in a primary document. Gross, before costs, which the seller also bore.

  3. $86,249,999,880

    Offering proceeds · 2026-06-15 · basis: offering-size · confidence: high

    Gross proceeds: 638,888,888 shares at the $135.00 offer price, before underwriting commissions and offering expenses. All of it to the company, because every share in the offering was newly issued by it. TWO NUMBERS FOR TWO MOMENTS, and the scene must not blur them. At pricing on 11 June the deal was 555,555,555 shares — 555,555,555 × $135.00 = $74,999,999,925, which is the "$75 billion" of the headlines. At closing on 15 June it was 638,888,888, because the underwriters took the whole 83,333,333-share over-allotment option. The prospectus prints the larger figure itself, in a dilution footnote: "the total consideration paid by our new investors would be approximately $86,249,999,880". So the widely quoted $75bn understates what was raised by $11.25bn.

  4. $0

    Offering proceeds · 2026-06-15 · basis: offering-size · confidence: high

    Zero, and the zero is the finding. Not one share in the offering was sold by an existing holder, so no proceeds went to anyone but the company. Established from the document rather than inferred from silence: the cover says "We are offering 555,555,555 shares" and puts the over-allotment shares "from us"; the cover’s proceeds table has a single proceeds line, to the company, where a deal with sellers in it carries two; and the strings "selling shareholder" and "selling stockholder" appear zero times in the whole 1.48-million-character document, whose 77 uses of "selling" are all "selling, general and administrative". Compare Facebook in 2012, where 57.3% of the shares offered were sold by existing holders.

  5. 1.50%

    Proportion (no currency) · 2019-12-11 · basis: offering-size · confidence: high

    The base offering was 3,000,000,000 shares, “representing 1.5% of its share capital”.

  6. SAR 6,400,000,000,000

    Equity value · 2019-12-11 · basis: reconstruction · confidence: high

    THIS WING’S MULTIPLICATION: SAR 32 times 200,000,000,000 shares = SAR 6.4 trillion. A price agreed for 1.5% of the shares, applied to all of them. Never a public market capitalisation: on the day it was set there was no market.

  7. $1,706,666,666,667

    Equity value · 2019-12-11 · basis: reconstruction · confidence: high

    SAR 6.4 trillion at the peg the prospectus prints, SAR 3.75 = US$1.00: about US$1.71 trillion. This wing’s arithmetic; narration only.

  8. 2.46%

    Proportion (no currency) · 2024-12-31 · basis: period-record · confidence: high

    Public shareholders at 31 December 2024: 5,944 million of 242,000 million shares, 2.46%, in the company’s own annual report.

Leaving the Old World Rooms

The exit leads back to the route, to 1901, and a steel company bolted together from parts.

  1. OW01 The Case With No Scale — For most of history there was no daily public price for a company, so there is nothing for a ruler to measure.
  2. OW02 The Book Closed at Midnight — Money paid into a company is not a price for it, and a share you cannot redeem needs a market to sell it in.
  3. OW03 Sealed Voyages — A company made of separately financed voyages has no single stock, so there is nothing continuous to price.
  4. OW04 Seventy-Seven Million Guilders — A famous number with no stated method is not a measurement; the VOC’s real peak was just over 77 million guilders, in 1720.
  5. OW05 1720 — A price is a fact and a market capitalisation is an argument; in 1720 the price and the capital belonged to different moments.
  6. OW06 Two Prices, One Company — A market capitalisation is one price applied to every share; where only a sliver trades, one company can carry two very different ones.
  7. OW07 Two Trays — An offering’s size says how many shares were sold, not who was paid; Aramco’s money went to the government, SpaceX’s to the company.
Back to the route: Room 10, Bolted Together