Capital went global. Origination stayed local.

A good company should not be limited to the institutions its owner happened to have the chance to meet.

On Thursday, Citi and Apollo announced a US$ 25 billion program for private credit and direct lending. The design deserves attention because it separates two competences that lived inside the same institution for a long time. Citi brings corporate relationships, origination, and the capacity to structure operations. Apollo brings an extraordinary quantity of private capital, alongside Athene and Mubadala. The program will start in North America, with a declared possibility of expanding to other geographies.

It would be easy to look at the news and see merely more money.

The interesting part is elsewhere.

One of the largest banking institutions in the world and one of the largest alternative managers decided to formalize something I have been slowly noticing while working with credit: finding a company and owning the balance sheet that will finance that company are distinct activities. Sometimes they belong to the same organization. They do not have to.

That changes how I think about the business.

For years, when a Brazilian businessman brought me a capital need, I naturally started from the map available around him. Banks I had relationships with. Funds I knew. Institutions interested in that kind of collateral. Some private investor. Family offices in specific situations. My advantage grew as the quantity and, above all, the quality of those relationships grew.

That model works.

It also contains a limitation so obvious it took us a while to notice: the client's financing universe ends up depending largely on the intermediary's universe of relationships.

A good company can pay dearly for capital simply because the right institution never knew it existed.

That is an inefficiency.

It does not mean any Brazilian company should finance itself abroad. The idea that foreign money is automatically better belongs to the same set of superstitions that makes people consider a fund better when it requires an invitation. Currency does not improve the quality of capital. Sometimes it merely changes where the risk is hidden.

A company that earns in reais and takes on dollar debt because the nominal rate looks lower may discover, some months later, that it financed its plant and simultaneously bought a currency position against itself. The spread saved in the presentation looks small when the liability grows twenty percent because of a variable that has nothing to do with the company's productivity.

Foreign capital requires structure.

A hedge, when necessary. Tenor. Jurisdiction. Collateral. Taxation. Documentation. Legal costs. The capacity to provide information. Compatibility between the currency of the debt and the currency in which the cash is produced. There is no sophistication in crossing a border merely to import a fragility that did not exist before.

Even so, ignoring foreign capital because it adds complexity would be an equally lazy conclusion.

The private credit market has grown to a scale that can no longer be treated as a detail of the financial system. The IMF estimates more than US$ 2 trillion between assets and committed capital globally. The concentration is still in the United States and Europe, but the consequence is already visible: there is an entire industry of non-bank capital looking for companies, assets, receivables, infrastructure, real estate, and other risks that can be financed outside traditional structures.

An enormous quantity of money came to need origination.

That may be the detail that most interests whoever is on the other side of the table.

A fund with tens of billions does not have the problem the small businessman imagines. It is not sitting on a mountain of money taking pleasure in turning people away. It has to put capital to work without destroying the criteria that justified its fundraising.

The challenge is finding assets.

Good assets.

And a good Brazilian private company is, from the financier's point of view, potentially an asset.

The businessman rarely thinks that way because he learned to walk into the bank in the psychological position of someone asking a favor. He fills in forms, hands over documents, and waits for the institution to decide whether it considers him worthy.

In situations of urgency and fragility, that position is real. Whoever desperately needs money has little bargaining power.

A solid, organized company seeking capital for a coherent economic purpose occupies another position. It is offering the creditor a possibility of return. There are two sides producing value in the operation.

A loan is not charity from whoever has money to whoever does not.

It is an exchange.

That change of perspective matters because it widens the question. Instead of simply looking for whoever is willing to lend, I start thinking about which balance sheet should carry that risk.

Perhaps a Brazilian bank.

Perhaps a FIDC.

Perhaps a private credit manager.

Perhaps a securitization company.

Perhaps a family with experience in that sector.

Perhaps, in an operation large enough and properly structured, a pool of foreign capital.

Geography should be a consequence of structure, not an intellectual border.

That looks particularly important in Brazil because we live with a curious contradiction. Business families have been internationalizing the asset side for years. They invest in the United States, keep foreign accounts, buy international funds, equities, bonds, properties, and structures outside the country. International private banks know those families and compete for part of their wealth.

The company that created the fortune, however, may still be financed by the same three or four domestic relationships.

The businessman can send US$ 5 million to invest in an American fund and may not know how to present his own company to an international credit fund.

There are legitimate economic reasons for that. There is also an absence of infrastructure.

The difference became clearer as I started observing operations that could, in theory, travel.

Money travels easily.

Risk does not.

For a foreign lender to analyze a Brazilian company, translating a PDF is not enough. The local context has to acquire institutional form. Corporate structure, ultimate beneficiaries, financial history, consolidated indebtedness, use of proceeds, projections, collateral, cash generation, sector, currency, compliance, KYC, legal documents, and a minimally intelligible explanation of why that company deserves capital.

The informal trust that works in the local market loses part of its usefulness when it crosses the border.

"I have known that family for twenty years" can be valuable information to me.

To a committee in New York, it is merely a sentence until what I built over those twenty years can be turned into evidence.

That is exactly where origination becomes more interesting.

A large international fund cannot manufacture local proximity in every geography. It can hire teams, open offices, and build relationships. Even so, there will always be companies that appear first to someone who was already close: an accountant, a banker, a lawyer, an adviser, a businessman, a family office, or an originator.

Global capital has scale.

The local relationship has context.

The value appears in the link between the two.

In recent years I had been treating that link mainly as a relationship. Today I think a relationship is only half the story. The other half is translation.

An originator has to turn a company's reality into an operation that can survive outside the physical presence of the businessman.

That looks trivial and is not.

There are companies with hundreds of millions in revenue whose financial information is still organized in a family way. One number is in the ERP. Another is with the accountant. The bank debt has to be reconstructed from statements. The properties sit in different holding companies. The lawyer knows the corporate structure. The founder knows what nobody documented.

Inside the circle of trust, the disorganization is compensated by human memory.

A distant lender does not have that memory.

Increasing the radius of access requires increasing legibility.

That may be one of the reasons mid-sized companies end up paying more than their economic risk would justify. Not because the financial system necessarily considers them bad, but because they arrive at the system incomplete.

Every unanswered question has to be compensated somehow.

Sometimes with more collateral.

Sometimes with spread.

Sometimes with refusal.

That makes me think organizing a company for capital can produce value even when the final financing stays domestic. A company able to survive the diligence of an institution that does not know its founder has probably become more comprehensible to local banks, investors, possible buyers, and to the family itself as well.

The foreign lender may be merely the stranger who forces better questions.

There is some usefulness in strangers.

Old relationships carry trust, but they also carry indulgence. Whoever has known the businessman for twenty years mentally fills in gaps a third party will force him to explain. Sometimes that is the difference between a relationship and an analysis.

The same holds for family offices.

I have been thinking of them less and less as investment offices and more as institutions that should represent the family before the entire financial system. Wealth does not consist only of assets. It has liabilities, companies, collateral, commitments, liquidity, and future needs.

A family can be a creditor on Tuesday and a borrower on Thursday.

It can join a club deal and, months later, need to finance an acquisition in the operating company.

Capital moves in two directions.

A structure that sees only the investments sees half the family.

That observation opens an interesting possibility: some families already have international relationships on the wealth side that could, in certain situations, widen the capital map on the corporate side. Not to mix everything irresponsibly, but to treat the family's financial position as a system.

The larger the wealth, the less sense it makes to accept that each part lives inside an institution that ignores the rest.

That view also changes the role of an origination house.

I had been thinking my function was to find money.

That is an insufficient definition.

If the client needs R$ 30 million, finding R$ 30 million may be precisely the fastest way to solve the wrong problem.

First I have to understand the purpose, the tenor, the cash flow, the wealth, and the options. Only then does the capital market come in.

The right house should not start from the shelf.

It should start from the client's balance sheet.

That brings the work far closer to a capital solutions boutique than to a traditional correspondent.

A correspondent represents an institution's products inside a specific relationship.

An independent boutique should represent the businessman's problem before several institutions.

The difference is one of direction.

One side starts at the product and looks for a fit.

The other starts at the client and looks for capital.

That second model interests me.

Perhaps because it preserves something I consider fundamental: the freedom to conclude that the best answer lies outside what I initially imagined.

If a local bank offers the best structure, excellent.

If a FIDC solves it better, we use the FIDC.

If there should be no debt at all, that possibility has to exist.

If foreign capital creates more risk than value, it disappears from the table without any embarrassment.

Financial independence does not mean the absence of incentives. It means, at a minimum, building an architecture in which the economic result does not depend on pushing one specific solution.

That is a hard standard.

It is also where reputation becomes capital.

If I present a bad operation to ten institutions, I can say I have a large network. On the eleventh, the network will have learned something about me.

The originator's reputation is an unpublished rating.

The lender opens the file differently depending on who brought it.

That fact will matter even more if we start operating with international pools of capital. The distant fund has less context of its own; it therefore tends to assign more value to the filter that brought the deal.

That creates a responsibility.

I do not want access to mean the capacity to spread a request around the world.

I want it to mean knowing where the request belongs.

It is the opposite of the vulgar marketplace.

Distributing to everyone is easy. The internet solved that.

The skill lies in not distributing to whoever should not receive it.

Confidentiality has economic value in private credit. A company seeking funding may not want suppliers, competitors, and dozens of institutions interpreting its need out of context. An originator who circulates operations indiscriminately can destroy the perception of quality before managing to build a structure.

More distribution can produce less value.

Technology frequently forgets that sort of thing because its natural religion is scale.

Private markets do not exist to maximize clicks.

They live on selection.

That may be why I am beginning to believe the infrastructure we need to build for this work should not be a public shop window of credit.

We need something else.

A system that knows the company, organizes documents, preserves history, and records capital relationships. That can look at an operation and reduce the universe of possible counterparties instead of widening it indiscriminately.

A private map.

The more I operate, the more I see that the most valuable knowledge about capital is not necessarily on the lender's website.

It is in the history.

The institution claims to do operations between R$ 20 million and R$ 100 million but has not looked seriously at anything below R$ 50 million for a year. It likes a certain collateral on paper but refused the last three similar operations over a specific detail. A fund says it looks at several sectors, but there is one team that really understands only a few.

That is information produced by repetition.

A boutique that operates a great deal should accumulate that memory.

If it does not accumulate it, it performs work without building an asset.

That is a mistake.

For years, my own memory and a few tools were enough. As the map grows, they stop being enough.

Every new source of capital increases optionality and simultaneously increases the complexity of the operation.

Foreign capital multiplies that problem.

It would be possible to keep solving everything with more people. Hire analysts, assistants, back office professionals. Build a traditional boutique.

Perhaps it works.

But there is something intellectually unsatisfying about using expensive human labor to administer tasks that should produce memory automatically.

Who received which document.

When.

What information is missing.

Which lender analyzed it.

Why it refused.

What proposal appeared.

Which collateral was discussed.

How long it took.

That is not merely back office.

It is raw material for the next decision.

If stored correctly, every deal should make the house smarter.

I am starting to suspect technology can be far more valuable to whoever provides the service than to whoever buys a license.

That is an important shift in how I have been thinking about software.

Over the last few years it became almost automatic to look at an inefficient process and imagine a platform other people will use. It looks like the modern way of building companies: find a profession, digitize its workflow, charge a subscription.

Perhaps that is right in many markets.

In mine, I am starting to doubt it.

If the software lets an originator produce a R$ 20 million operation and capture a commission of hundreds of thousands of reais, why would the best economics necessarily lie in selling him the tool for a small fraction of that?

Perhaps the tool is more valuable closed.

Perhaps the engine should serve the house itself.

It is still a question, not a decision.

But it is a question I can no longer ignore.

A traditional capital boutique has an evident disadvantage: it grows by adding people. More deals require more analysts, more coordination, more administrative work. Revenue rises and the human cost follows.

A proprietary engine can change that relationship.

Not by eliminating the financier.

By eliminating what wastes the financier.

The work of judgment should stay human. Knowing the businessman, reconstructing the need, negotiating structure, interpreting risk, noticing an inconsistency, knowing when a relationship deserves to be used. That is hard to turn into automation without destroying exactly what creates value.

Looking for a file, updating a status, comparing versions, organizing a data room, and remembering a follow-up are far less noble and far more frequent tasks.

Perhaps the financial house of the next generation looks less like a fintech selling tools and more like a boutique using technology the client never sees.

The client buys the result.

The machine stays with us.

I like that idea because it avoids a frequent perversion of enterprise software: transferring work to the user and calling it efficiency.

I do not want the businessman to have one more dashboard.

I want him to have to think less about financing.

He should be able to arrive with an economic problem and receive an architecture of capital.

If the best structure requires a local bank, we go get it.

If it requires private credit, we go get it.

If the operation can access foreign capital, we do the work needed for it to exist in that market.

The client does not have to learn our map.

That is precisely the reason to hire the house.

Perhaps the right analogy is closer to a family office than to a technology platform.

The family does not hire a family office to receive a list of five hundred funds.

It hires one so that someone knows five hundred and puts three on the table.

The value is in the filter.

In credit, it should be similar.

Not "we have 300 institutions."

That is a database.

I want to be able to say something far simpler: we know the market well enough to know where your operation should be.

The difference between those two sentences is the difference between information and judgment.

If we ever manage to operate that way, access to foreign capital will stop being an exotic product.

It will be merely one more dimension of the map.

That also strikes me as the right way to keep the low profile I always associated with the private market.

We do not have to announce every capital relationship we have. The businessman should not hire us because he saw familiar logos in a presentation.

He should hire us because he knows that, when a serious need arises, we will put the structure in front of the right counterparties.

The institution that has to demonstrate access all the time may be selling status.

I prefer to sell execution.

On Thursday, Citi and Apollo announced US$ 25 billion of a partnership that combines exactly distinct competences. Citi brings its network of companies and its origination capacity. Apollo brings capital.

It is a scale that bears no resemblance to what I do today.

The principle, however, does not depend on scale.

The relationship can stay local.

The capital can be global.

Between the two there is a layer of intelligence, preparation, and coordination.

That may be exactly where I want to build my house.

Not a company that sells the road to other drivers.

A house that knows the roads better than the client needs to know them.

Because the businessman should not spend his life discovering where the money is.

He already has a business to run.

Leo Bentier