Because a mature business eventually converges toward the growth rate of its market, at some point you can no longer avoid the question: "Should the cash flow that the core business generates be reinvested into the core business, or into something else entirely?"
10X is not yet at that phase, but ours is a business where one side of the cash flow is fairly easy to see, so I feel we need to start thinking about this relatively early.
To sharpen the wording a bit more, I'd put it this way: "the decision about where to reallocate money once the cash on hand starts to exceed what the core business needs to grow."
A company that pushes this question to its logical extreme would, I imagine, appear to have transformed "from an operating company into an investment company." And in fact several such companies exist. Berkshire Hathaway, led by Buffett, is probably the best-known example.
That said, I don't know Berkshire Hathaway well myself, and I'm not familiar with other cases either. So, wanting to widen my sample and understand this better, I did some research and put it together.
I made full use of Claude for the research, having it gather examples from around the world of companies that shifted "from operating company to investment company," then picking out five that seemed especially instructive for me and organizing them.
Premises
Let me define a few of the slightly harder terms up front.
- Investment holding company (permanent capital type): A company that invests using funds on its own balance sheet, that is, its own equity or operating cash flow. Because the money is not entrusted from outside, it is never chased by redemptions or withdrawals. Berkshire is the archetype.
- External LP fund manager (GP type): A company that raises money from outside investors (limited partners, LPs), invests as the manager (general partner, GP), and earns management fees. The capital raised comes with maturities and redemptions.
- Insurance float: The money an insurer can invest in the interim between receiving premiums and paying out claims. If underwriting is profitable, the effective cost of that money is actually negative.
- Serial acquisition: A model of continuously reinvesting the cash flow you earn into one company acquisition after another.
Even within the shift "from business to investment," the "permanent-capital type that invests its own money" and the "manager type that raises outside money" have completely different structures, even though both are called "investment companies." Berkshire is the former; SoftBank is the latter.
The Big Picture
Before getting into individual cases, let me give an overview of the five companies I picked for this piece.
| Company | Original core business | Shift to investing | Type of investment capital | Current form | Capital-structure type |
|---|---|---|---|---|---|
| Berkshire Hathaway | Textiles (merger in 1955) | From 1967 (insurance acquisition) | Insurance float + own equity | Investment holding conglomerate | Permanent capital |
| Brookfield | Power, infrastructure, real estate | From 2005 (became a manager) | Own equity + external LP | Major alternative-asset manager | Hybrid (moving toward manager type) |
| SoftBank Group | Software distribution, telecom | From 2017 (Vision Fund) | Own equity + external LP | Strategic investment holding company | External LP fund type |
| Tencent | Games, social networking | From 2011 (strategic investment) | Operating cash flow | Operating company + huge investment book | Permanent capital (strategic investment) |
| Constellation | Vertical-market software | From 1995 (since founding) | Free cash flow | Serial-acquisition machine | Permanent capital |
Only two of them, Brookfield and SoftBank, are seriously raising external LP capital. The other three do not solicit money from outside; they invest with the money they earn themselves or with insurance float.
Now let me look at each company one by one.
Photo by Hoover Tung on Unsplash
Berkshire Hathaway: From Textile Mill and Insurer to Own-Money Investor
Berkshire's starting point was a declining textile company.
In 1955, Berkshire Fine Spinning Associates and Hathaway Manufacturing Company merged to create Berkshire Hathaway.
At the time it was based in New Bedford, Massachusetts, with over 12,000 employees and more than $120 million in sales, but pressed by cheaper competitors it kept shutting down plants (Wikipedia: Berkshire Hathaway).
In 1962, Buffett began buying Berkshire shares at around $7.50 each as a "cigar butt" (deeply undervalued stock) through his own fund, Buffett Partnership Ltd. (BPL).
In 1964, angered by a discrepancy over the buyback price with the then-president (verbally $11.50 per share, but $11.375 in writing), Buffett refused the tender and instead bought more, taking control in 1965 (The Rational Walk).
A turning point came in 1967.
Berkshire acquired the insurance company National Indemnity for about $8.6 million.
This was the first full-scale acquisition under Buffett's control, and the starting point of his move into insurance. The target was the roughly $19.4 million of insurance float.
Buffett began to use this money, which he could hold from the time premiums came in until claims were paid out, as capital for buying stocks and businesses (Tontine Coffee-House).
From then on, Berkshire shifted its center of gravity away from textiles entirely. In 1972 it acquired See's Candies, moving from bargain-hunting toward a "buy great businesses" style.
See's is a maker and retailer of premium chocolates and candy long beloved on the West Coast. Thanks to its strong brand, customers didn't leave even as it raised prices every year, and it generated steady cash with little capital investment.
Through this acquisition, Buffett is said to have learned the value of "an excellent business with pricing power" over cheap assets.
In 1985 he closed the last of the textile operations.
In 1996 he acquired the remaining 49% of GEICO for about $2.3 billion to make it a wholly owned subsidiary, and in 1998 he acquired the reinsurer General Re for about $22 billion.
Here the insurance float swelled from about $19.4 million in 1967 to about $171 billion by the end of 2024 (Berkshire 2024 Letter to Shareholders).
Buffett describes float as "as if others deposit money with us, even pay us a fee to hold it, and let us invest that money for our own benefit" (Berkshire 2010 Letter to Shareholders).
| Year | Event |
|---|---|
| 1955 | Two textile firms merge to form Berkshire Hathaway |
| 1962 | Buffett (BPL) begins buying at about $7.50/share |
| 1965 | Buffett takes control |
| 1967 | Acquires National Indemnity for about $8.6M, enters insurance |
| 1972 | Acquires See's Candies (shifts toward quality businesses) |
| 1985 | Final closure of textile operations |
| 1996 | Fully acquires GEICO for about $2.3B |
| 1998 | Acquires General Re for about $22B |
The important point here is that Berkshire is not a fund that raises money from external LPs.
From 1956 to 1969, Buffett did run a "fund" called BPL that took in money from outside investors. It charged management fees (25% of returns above 6% a year) and allowed investors to redeem. It was a classic GP/LP-type fund.
But in 1969, Buffett dissolved BPL.
He let investors choose between Berkshire stock or cash, while he himself kept holding Berkshire stock.
After becoming chairman in 1970, Berkshire became his permanent investment vehicle (Wikipedia: Warren Buffett).
In other words, Buffett wound down the fund that held outside money (BPL) and switched to a holding company (Berkshire) that invests with its own equity and insurance float.
There are no external LPs, no management fees, and no redemption pressure. Buffett's compensation is still just $100,000 a year.
Something else that surprised me: Buffett himself looks back on his decision to take control of Berkshire as a "monumentally stupid decision."
He has even called choosing a declining textile company as his starting point a "$200 billion blunder" (CNBC). The starting point wasn't a planned success story but rather a big mistake and the cleanup that followed.
Brookfield and SoftBank: Moving to the Side That Aggregates Outside Money
Unlike Berkshire, the next two are examples of companies that moved toward being "managers" that raise money from outside investors.
Brookfield: From Stable Business to a Management Business
Brookfield's origins go back to 1899.
The starting point was a holding company (later Brascan, formerly Brazilian Traction, Light and Power) built when Canadian investors invested in Brazilian electricity and streetcar businesses (Brookfield official history).
For a long time it was an operating holding company that owned and operated power, infrastructure, and real estate on its own balance sheet.
In 2001 it formed its first private equity fund, and when Bruce Flatt became CEO in 2002, it began steering from a conglomerate toward "an asset manager that operates real assets with institutional-investor money."
In 2005 it changed its name from Brascan to Brookfield Asset Management.
At that point assets under management were about $40 billion. From here it increasingly took on the character of an asset manager that not only owns its own assets but also earns fees by managing outside investors' money.
The shift was completed as a structure in December 2022.
Brookfield spun off and listed 25% of its asset-management business, splitting the company in two.
The parent, Brookfield Corporation (BN), holds permanent capital and its own investments, while the pure manager, Brookfield Asset Management (BAM), handles the management of external LP capital and fee income.
BN holds 75% of BAM (Wikipedia: Brookfield Asset Management).
Assets under management surpassed $1 trillion in 2025, of which fee-bearing capital reached about $549 billion (BAM Q1 2025 Shareholder Letter).
Why carve out the manager?
In his 2022 shareholder letter, CEO Bruce Flatt gave three reasons:
- a pure manager is easy to value and commands high multiples in the market,
- the management business requires almost no equipment or working capital, so most of its profit can be paid out as dividends,
- and it generates pension-like, long-term cash flows
(Q2 2022 Letter to Shareholders).
It seems fair to call this a case where a company with a stable business base "went on the offensive," moving into a more capital-light, higher-margin management business.
What strikes me is that it wasn't cornered into this. Using the stable cash its business generated as leverage, it moved of its own accord toward a form the market values more highly.
| Year | Event |
|---|---|
| 1899 | Origins in Brazilian power business (later Brascan) |
| 2002 | Flatt becomes CEO, shifts toward asset manager |
| 2005 | Renamed Brookfield Asset Management (AUM about $40B) |
| 2022 | Spins off and lists manager BAM, moves to BN/BAM two-company structure |
| 2025 | AUM surpasses $1 trillion, fee-bearing capital about $549B |
SoftBank: Turning Telecom Cash Into an External-Capital Fund
SoftBank, a household name in Japan, was founded in 1981 by Masayoshi Son as a software distribution company.
After that, as everyone knows, it built up an internet and telecom business base through acquisitions: Yahoo Japan (1996), Japan Telecom (2004), Vodafone's Japanese operations (2006, about 1.75 trillion yen), Sprint (2013, about $21.6 billion), and ARM (2016, about 3.3 trillion yen) (SoftBank Group - Wikipedia).
On the foundation of this operating cash flow and these assets, in May 2017 it formed the SoftBank Vision Fund (SVF1).
Against a target of $100 billion, it secured more than $93 billion at first close.
The breakdown of commitments: Saudi Arabia's Public Investment Fund (PIF) up to $45 billion, SoftBank itself up to $28 billion, Abu Dhabi's Mubadala $15 billion, and the rest contributed by Apple, Foxconn, Qualcomm, Sharp, and others (SoftBank Group "First Major Closing").
This is where the decisive difference from Berkshire lies.
In addition to its own money ($28 billion), SoftBank raised capital from external LPs such as PIF and Mubadala, and runs the fund as the manager (GP) itself.
The official release explicitly states that "an overseas wholly owned subsidiary of SBG oversees the fund as the GP."
Son has described this strategy in various ways: the "cluster of number ones" strategy, "a company that grows for 300 years," and "a strategic holding company for the information revolution" (Logmi Finance).
That said, the difficulty of raising external capital was also exposed.
For Vision Fund 2 (target $108 billion), announced in 2019, the Gulf sovereign LPs that had been the largest investors in SVF1 dropped off the initial list, and as a result the fund ended up centered on SoftBank's own money (CNBC).
I think a model that raises external capital tests something different from the core business's cash: your trust and track record as a manager.
| Year | Event |
|---|---|
| 1981 | Founded as a software distribution company |
| 2006 | Acquires Vodafone Japan for about 1.75 trillion yen |
| 2016 | Acquires ARM for about 3.3 trillion yen / agrees on Vision Fund concept with PIF |
| 2017 | Forms Vision Fund 1 (over $93B, centered on external LPs) |
| 2019 | Announces Vision Fund 2 (external LPs don't gather, centered on own money) |
Tencent and Constellation: Turning Software Cash Into Investment
Finally, since 10X is a software business, let me take up examples of similar companies.
Software businesses have a structure that tends to generate abundant cash, thanks to high gross margins and low incremental reinvestment needs.
Let me look at two examples that redirected that cash into investment.
Photo by Jakub Żerdzicki on Unsplash
Tencent: Investing in Over 800 Companies With Gaming Profits
Tencent was founded in Shenzhen in 1998, growing into one of the world's largest game companies through QQ, WeChat, and its gaming business.
For full-year 2025, revenue reached 751.8 billion yuan, net profit attributable to shareholders 259.6 billion yuan, and free cash flow 182.6 billion yuan (Tencent 2025 Annual Results).
Note: the current exchange rate is around 24 yen per yuan.
Using this abundant cash as capital, Tencent has built an enormous investment portfolio.
Riot Games (majority stake in 2011, fully acquired in 2015), JD.com (15% stake in 2014), Supercell (majority stake for about $8.6 billion in 2016), and investments in Tesla and Spotify. Its investees exceed 800 companies (Wikipedia: Tencent).
As of the end of 2025, the fair value of listed investees was 672.7 billion yuan and the book value of unlisted investees was 363.1 billion yuan, together amounting to about 1 trillion yuan (on the order of $140 billion).
What's notable is that all of this is proprietary investment.
Whereas Berkshire leveraged insurance float, Tencent's is strategic investment via pure operating cash flow and its own balance sheet.
Founder Pony Ma (Ma Huateng) uses the phrase "half a life" (半条命) to describe an investment philosophy that doesn't seek controlling stakes, entrusts investees with autonomous management, and helps grow the ecosystem (Sina Finance).
It looks like investment aimed less at financial returns per se and more at strategic synergy with the core business. The order of operations, where a high-margin cash cow like gaming comes first and its surplus is used to expand around the core business, struck me as textbook and instructive.
| Year | Event |
|---|---|
| 1998 | Founded in Shenzhen (QQ → WeChat → games) |
| 2011 | Acquires majority of Riot Games |
| 2014 | Takes 15% stake in JD.com |
| 2016 | Acquires majority of Supercell for about $8.6B |
| 2022 | Distributes about $20B of Meituan shares to shareholders as a special dividend |
Constellation: Serial Acquisition as the Core Business
Constellation Software may be the most extreme example.
In 1995, former venture capitalist Mark Leonard raised about C$25 million from pension funds such as OMERS and founded the company in Toronto (Colin Keeley).
Its business model is to relentlessly keep acquiring small and mid-sized companies that make vertical-market software (VMS) for specific industries.
VMS businesses, with high switching costs and sticky customers, generate stable cash with high gross margins and little incremental investment.
That free cash flow is reinvested into the next acquisition.
After acquiring, it almost never sells; it holds forever.
It acquires more than 100 companies a year, and over 1,000 cumulatively (Constellation FY2024 Results).
Leonard has a thorough discipline of capital allocation.
The hurdle rate for acquisitions (the minimum return required) is set at an IRR of 20-30% depending on scale, and he won't make investments whose returns fall short of the bar.
Under this discipline, since its IPO in 2006 the stock has compounded at about 36,000%, or roughly 30% annually (Cantech Letter).
Constellation, too, is not an external LP fund. It is the permanent-capital type that keeps acquiring with the cash it generates itself.
That's why it's called the software version of Berkshire, and personally, I felt that this form, where "core business = investment" is completely fused, is the most beautiful of the five.
| Year | Event |
|---|---|
| 1995 | Mark Leonard founds the company in Toronto |
| 2006 | Lists on the TSX (main purpose: liquidity for VC investors) |
| To present | Acquires over 100 companies a year, over 1,000 cumulatively |
Why They Moved Away From the Real Business: Breaking Down the Inevitability of the Shift
Let me rearrange the five companies by the motive of "why they moved to investing."
| Company | Structure of core business | Nature of capital that enabled the shift |
|---|---|---|
| Berkshire | Textiles (structural decline, low returns) | Insurance float (low cost, long term) |
| Brookfield | Infrastructure (stable, low growth) | Stable CF + management know-how |
| SoftBank | Telecom (mature, stable CF) | Business CF + external LP capital |
| Tencent | Games (high margin, high CF) | Operating CF (proprietary) |
| Constellation | VMS (high margin, low reinvestment) | Free cash flow |
From this table, it seems there are broadly two types of shift.
One is the "no choice" type, which is Berkshire's case.
Precisely because its core textile business was in a structural decline that couldn't produce returns worthy of additional investment, it had no choice but to direct the money it earned not into the core business but into investments with higher returns.
The other is the "offensive" type.
Here the core business is actually generating too much cash, and reinvestment targets within the core business run dry. So investment is chosen as the optimal place to put the surplus cash.
Tencent and Constellation are the archetypes. High-margin software businesses with little additional investment generate large amounts of cash to allocate to investing.
SoftBank and Brookfield, by contrast, are examples of leveraging the stability of the core business to move into the investment business with external capital.
Takeaways: The Levers That Work When an Operating Company Becomes a Fund
Having lined up the five companies, let me finally organize what I feel I can take home.
The first is that the cash-generation structure of the core business determines whether the shift to investing becomes inevitable.
The higher the margin and the smaller the need for additional reinvestment, the more surplus cash piles up, and the more naturally investment surfaces as an option.
Tencent's and Constellation's software businesses are exactly that, while Berkshire's was a shift from the side where the core business does "not" generate cash. The path was the complete opposite.
I suspect that the more a business has "cash flow that is easy to see," as I wrote at the start, the sooner this question arrives.
The second is that the "quality" of the capital allocated to investing determines the strength of the investment itself.
Berkshire's insurance float was low-cost, long-term money that, moreover, is never redeemed.
Brookfield's "pension-like cash flow" has the same nature. The cheaper, longer, and less withdrawable the source, the easier it is to let compounding work.
More than how much you invest, it's what nature of cash you put into investing. My understanding is that this is the lever that matters most.
The third is that the own-money type and the external-money type demand entirely different capabilities.
The own-money type (permanent capital) works if you have the core business's cash-generating power and excellent capital-allocation discipline.
The external-money type (manager), on the other hand, needs, on top of that, enough trust and track record to be entrusted with money by outside investors, and the nerve to keep meeting their expectations.
As SoftBank's Vision Fund 2 showed, no matter how strong the core business is, external capital won't gather without credibility as a manager.
The fourth is that in the end, a manager's very capital-allocation ability becomes the competitive advantage.
Buffett's float management, Leonard's hurdle-rate discipline, Son's vision.
In every case, the essence of the company shows up at a single point: where they direct their cash.
An operating company turning into an investment company is, when you push it to the core, the act of rebuilding capital allocation as the company's central capability.
For the day 10X faces that question, I intend to keep taking these practice swings.
References
- Berkshire Hathaway Shareholder Letters Archive
- Berkshire Hathaway - Wikipedia
- The Rational Walk: Berkshire Hathaway's Great Transformation
- Brookfield official history / Brookfield Asset Management - Wikipedia
- Brookfield Q2 2022 Letter to Shareholders
- SoftBank Vision Fund "First Major Closing" (2017) / SoftBank Group - Wikipedia
- Masayoshi Son's "cluster of number ones" and "300-year vision" - Logmi Finance
- Tencent 2025 Annual Results / Tencent - Wikipedia
- Mark Leonard's President's Letters (Constellation Software)
- Mark Leonard (Constellation Software) Operating Manual - Colin Keeley
Comments
Share your thoughts or questions. Comments are published after approval.
Write a comment