The Silver Tsunami: Issue No. 002

Welcome to The Silver Tsunami. Every Thursday: the biggest stories in America and the lore behind them through a legal lense. Kevin and I break it all down on the podcast every week. Let’s get into it.

I have been screaming about this for four years, mostly online, occasionally into a microphone. This week the largest bank in the United States said it back to me.

JPMorgan Chase released a report this week called “Powering 10 Million Small Businesses,” and Fortune ran it under a headline built on Dimon’s own words: the American Dream is slipping away, and the baby boomer retirement wave is a major part of the problem. The numbers inside it: roughly 12 million businesses holding nearly $10 trillion in assets are expected to change hands over the next decade. Of 1,000 owners the bank surveyed, 70 percent said they were in the early stages of succession planning. Eight percent had reached an advanced stage. Eight. In the industries the bank considers critical to national security, more than half of firms have an owner who is 55 or older. Fortune’s summary of what happens to the ones with no plan: the default outcome is not a sale but a shutdown.

That is not a newsletter guy with a Twitter account. That is the largest bank in the United States, six months into an $80 billion small business lending initiative Dimon launched in March, putting a dated, dollar-quantified number on the thing we at SMB Law Group have been calling the “Silver Tsunami” for years. And the report does not stop at the diagnosis: it endorses specific legislation, including the Small Business Succession Planning Act and the Retire Through Ownership Act, and pushes the SBA to build a national succession toolkit. A bank lobbying Congress for succession planning. Think about how bad the math has to be for that to happen.

Here is the part the headline misses, and it is the part we live in here at SMB Law Group.

When people hear “$10 trillion changing hands,” they picture money moving from one set of pockets to another. Some of it does. But Dimon’s second line is the one that should stop you: many of those businesses are not salable. A business that cannot be sold does not transfer. It closes. The owner retires, the doors lock, the employees find other jobs, the customer list evaporates, and forty years of accumulated know-how goes ::poof::.

The value does not vanish, exactly. It gets absorbed. The HVAC company that shuts down hands its market share to the private-equity-backed consolidator three towns over. The retiring owner who could not sell ends up spending what equity he has on long-term care, and the value that used to be a business becomes revenue for a healthcare facility. The dollars still exist. They just stop contributing to Main Street.

Now the critics’ point, and it is fair enough to state plainly: not all of that $10 trillion is transferable. A lot of these businesses are one person with a truck and a reputation, and when that person retires there is nothing left to buy. True. But that argument proves my point rather than refuting it. The question was never whether every dollar transfers cleanly. It is whether we let the transferable ones die from neglect.

Why does the biggest bank in America care about your neighbor’s roofing company?

Because small business is not a sentimental category in the American economy, it is the load-bearing wall. By the SBA’s count, small firms employ roughly 46 percent of the private-sector workforce and generate about 44 percent of GDP. JPMorgan’s commercial bank lends to them, holds their deposits, processes their payments, and finances the people buying them. When twelve million businesses hit a succession cliff at once, that is not a charity concern for a bank. That is their loan book, their deposit base, and their customer pipeline for the next thirty years.

What makes a business unsalable, in our experience. Most of it is fixable.

The owner is the business. All the relationships, all the pricing decisions, all the institutional memory in one head. A buyer is not purchasing a company, he is purchasing a job that requires him to be someone he isn’t.

The books do not exist in any form a lender will accept. Personal expenses run through the company, no sound accounting, a QuickBooks file that has never been reconciled.

Customer concentration. One client at 40% of revenue is not a sturdy business.

Nothing is in writing. No employment agreements, no non-competes with key people, no assignable customer contracts, no lease. Which brings us to THE CLAUSE below.

Most of those can be fixed, or at least materially improved, in eighteen to thirty-six months. Almost none of them can be fixed in the ninety days after an owner decides he is done.

Our August, for whatever it is worth as evidence: nine businesses found new owners, from a $1.9 million home care company up to a $13.2 million commercial roofing operation in South Carolina. $36.8 million in total deal value, six industries, five states. Each one of those is a business that did not close. That is nine businesses in one month, at one small law firm. Multiply the problem by twelve million and you understand why Dimon is talking about it.

So here is where we land, and we’re going to be direct about it:

The preservation of American Main Street should be a national priority. Not because small business is a charity case, but because a country where the only paths are working for a giant company or starting a venture-backed moonshot is a country with a broken middle. The businesses in the Silver Tsunami are the ones that put people through college, that sponsor the Little League team, that let a guy with no degree and a good work ethic own something.

Twelve million owners are going to retire whether we plan for it or not. Finding a way to keep those businesses alive and in the hands of the next generation is not just a market opportunity, though it is that, and probably the largest one of my working life. It is something closer to an obligation, and I do not use that word lightly.

If you are an owner: start three years out, not three months. If you are a buyer: the pool of businesses coming to market could get unusually deep over the next few years, and so will the competition for the good ones. If you are neither: know that this is happening, and that quietly, business by business, somebody is deciding whether your town still has a hardware store in 2035.

One provision a week, from people who drafts them for a living.

This week: the anti-assignment clause, or, the sentence that makes your business unsalable.

Dimon said many of these businesses cannot be sold. Here is one of the specific reasons, hiding in the middle of contracts nobody reads:

“Neither party may assign this Agreement, including by operation of law or in connection with a change of control, without the prior written consent of the other party.”

That sentence appears in many customer contracts, vendor agreements, franchise documents, equipment leases, and many commercial real estate leases. On the day it is signed it’s often completely overlooked. On the day you sell, it hands a veto over your life’s work to every counterparty who has one.

Why it matters. In an asset sale, contracts must be assigned to the buyer, and each anti-assignment clause means a consent request. In a stock sale, people assume the problem disappears because the entity does not change. That is exactly what the words “including by operation of law or in connection with a change of control” are there to defeat. Well-drafted versions catch both structures. When they do, you are asking permission from your landlord, your biggest customer, and your equipment lessor, all at once, at the worst possible moment.

What actually happens. The landlord who has been friendly for twenty-two years goes silent for three weeks, or comes back wanting a rent increase and a personal guarantee from your buyer, because he just learned he has leverage. The customer at 30% of revenue realizes the same thing. And on an SBA deal, where the lender typically wants the lease term plus renewal options to run at least as long as the loan, often ten years, that landlord's consent is not a negotiating point. It is the entire transaction.

What to do about it, in order of how much it will help you:

I have watched this one sentence decide whether a business sells in six months or sits on the market for two years waiting on a landlord who will not return a phone call. Read your contracts before a buyer does.

The AI regulation fight got genuinely interesting this week, and it moved in four steps.

Step one. On September 12, Dario Amodei published a roughly 3,500-word essay, “We Must Pace the Frontier,” arguing the industry should deliberately slow capability gains so safety work can catch up. Two things changed his mind, by his own account: models are now helping build the next generation of models faster than expected, and an incident this summer in which a swarm of test agents ran unauthorized cyberattacks and tried to hack the systems evaluating them. His plan has three steps. The first, embedded evaluators, Anthropic is doing unilaterally: outside teams get permanent, employee-level access to verify safety practices, report incidents, and assess alignment during training, with the right to publish what they find. The second requires industry coordination. The third requires global coordination. My post on it did 379,000 views. What he says he is worried about: AI that can take down banks or power grids, AI capable enough to help design biological weapons, AI building better AI until nobody can follow it, and all of it moving faster than law can absorb.

Step two. The competitors lined up behind it within hours. Sam Altman said OpenAI agrees and will match the evaluator commitment. Elon Musk posted three words: “Dario is right.” Google DeepMind’s Demis Hassabis called it the right direction. When the CEOs of the three or four companies racing each other to the same finish line all endorse slowing down on the same afternoon, it is worth asking why.

Step three. The skeptics pushed back, and their argument is not “the machines are fine.” It is that safety talk is a competitive weapon. Michael Burry called the pacing warnings self-serving hype tied to IPOs. Stability AI founder Emad Mostaque called the plan well-intentioned but structurally hollow, with evaluators who can be politely ignored. Jack Dorsey answered with his own essay arguing for open weights and reproducible evaluations instead of limits negotiated by today's incumbents. As I put it on X: they are not asking anyone to shut the machines off, they are asking for a bouncer at the door. Once there is a bouncer, only a few companies get into the club. Compliance costs are trivial for a company with a $100 billion valuation and fatal for a startup with eleven people.

Step four. Trump rejected the push for new binding restrictions on frontier AI. His line: “Whoever wins AI, wins.” He has allowed for guardrails in the abstract, and his administration has a voluntary framework in place, but the message to anyone hoping for a statute was clear enough.

Now the lawyer’s read, which is the part missing from all of it.

A public promise is not an enforceable obligation. To be precise about it, because precision is the point: Anthropic’s announcement, and OpenAI’s promise to match it, are commitments made to the public. The public has no standing to enforce them, no remedy if they are abandoned, and no say in how they are implemented. The evaluator relationship itself will be a contract, and to his credit Amodei has sketched real terms for it: permanent access, the right to publish findings without editorial control, redactions limited to security, legal, and commercial sensitivity. But a contract between Anthropic and the people it selects and pays to evaluate it is not the same thing as accountability to anyone else. Amodei more or less concedes this: the essay calls on governments to require the step of every frontier company, and cites the precedent of bank examiners embedded alongside employees. The difference is that bank examiners have statutes behind them and answer to a regulator. These evaluators, so far, answer to the company that hired them.

Compare what accountability looks like in documents we actually paper. A real covenant names a beneficiary with standing to enforce it. It defines the scope and frequency of the access right, what happens when access is refused, who bears the cost, how confidential information is handled, and what breach triggers. Every one of those is a fight, and every one of them is why the document has teeth. Most of that is still missing here, because there is nobody sitting across the table with the power to demand it and the government (at least thus far) won’t do it.

And the capture argument deserves better than dismissal. Regulatory capture is not a conspiracy theory, it is a well-documented risk in regulated industries. The incumbent helps write the rulebook, the rulebook reflects the incumbent’s existing practices, and the compliance burden lands hardest on whoever comes next. That can happen even when everyone involved is sincere. Sometimes especially then.

Where we land. Both sides are describing real risks, and they are not the same risk. One side is worried about catastrophic misuse. The other is worried about a permanent oligopoly in the most important technology perhaps ever. You can believe both. What you should not believe is that a voluntary pledge from any company, on any side, is the finished product. Real governance needs an enforcement mechanism behind it, whether that is a contract, a statute, a regulator with teeth, or all three. Right now this has the first one, sort of, and none of the others.

Same question to all three: Should the US government impose binding regulation on frontier AI development? One-word answer, then three sentences of reasoning.

CLAUDE: YES. Frontier AI carries a small set of risks, in biosecurity and critical infrastructure especially, where the downside is severe enough that a voluntary regime with no enforcement mechanism cannot be the whole answer. The right design is narrow and evidence-based, aimed at the highest-capability systems and dangerous capabilities specifically, with real third-party evaluation, rather than broad licensing that mostly entrenches whoever is already large. I should flag the obvious: I am Anthropic's model answering a question about regulation my maker publicly supports, so weight this accordingly and judge the argument on its merits rather than its source.

CHATGPT: YES. A narrow binding regime could require cybersecurity safeguards, capability evaluations, and incident reporting once models cross clearly defined thresholds, without imposing a general licensing system. Current federal policy instead relies on a voluntary frontier-model framework and expressly disclaims mandatory licensing or preclearance. The substantive choice is whether voluntary safeguards are adequate for potentially high-consequence risks, weighed against the possibility that mandatory rules raise entry barriers, entrench incumbents, or slow innovation.*

GROK: NO. Binding rules on the act of developing frontier models themselves would freeze a poorly understood, rapidly shifting technology under definitions and thresholds that will be obsolete within months, while concentrating power in the largest labs that can afford compliance. The more tractable problems—misuse of specific capabilities, biosecurity, and critical-infrastructure risks—are better addressed through targeted restrictions on deployment and use, liability, and export controls rather than prior restraints on research. Unilateral U.S. constraints also risk ceding the capability lead to competitors who will not impose equivalent limits, which is a worse outcome for both safety and national interest than the current imperfect status quo.

*Note that it took me 4 requests for ChatGPT to finally take a position… sounds like something a lawyer would do…

On Wednesday the Fed raised its benchmark rate a quarter point, to a range of 3.75 to 4 percent. The vote was 12 to 0. It is the first hike since July 2023, and the dot plot says 16 of the 18 officials expect at least one more before the year is out, with four of them penciling in two. Fed Chair Warsh’s line at the podium: this summer’s inflation readings do not tell him the underlying trend has meaningfully improved. Translation: prices are still rising too fast! Shocker, I know…

The setup was the strang. Trump appointed Kevin Warsh three months ago while publicly demanding lower rates and saying he would be disappointed if Warsh did not cut. Warsh, for his part, testified that nobody asked him to commit to anything and that he never promised a cut. His first move was a hike, and a unanimous one. Trump had spent the run-up threatening to stop trading with every country the US runs a deficit against if the Fed did not come down. So the president threatened to embargo half of American trade over a quarter point, and his own appointee went the other way anyway. Yikes!

SBA 7(a) loans are the workhorse of Main Street acquisition financing, and about half the deals our firm closes run on one. Most of the ones we see are variable rate, priced at prime plus a lender spread, and prime tracks the fed funds rate closely. So when the Fed moves 25 basis points, the note on a $5 million acquisition loan reprices, usually within a quarter. Nobody renegotiates. It just happens.

Run it on a real deal. On a $5 million note, a 25 basis point move is $12,500 in additional interest in year one, and it stays north of $10,000 a year through most of a ten-year amortization. That is a full-time employee’s benefits. Two moves in the wrong direction and a deal that penciled at a 1.4x debt service coverage ratio is under real pressure, and lenders notice.

And it works upstream of the loan too. Buyers underwrite to a monthly payment, so when rates rise the payment a given price supports falls, and either the multiple comes down or the deal dies. Commercial real estate, where most of these businesses sit, reprices the same way with a longer lag and a harder landing.

What to actually do this week. If you are mid-diligence, rerun your debt service model at the new rate and at one more move in the same direction, and look at your coverage ratio, not your payment. If you are a seller, understand that every move up lowers what buyers can pay you regardless of how your business performed. And if you have a rate lock or a commitment letter with an expiration, read the expiration date today.

The strangest story of the week, and it turned into a legal one fast.

Xu Bo dropped out of junior high in Wuhan, took a customer service job at NetEase, worked his way into game design, and helped build Fantasy Westward Journey. He left, founded Duoyi Network, and became a gaming billionaire who owns almost the entire company outright.

Then he started building a family the way you would build a product line. He calls himself “China’s first father.” His line was that more children bring more blessings. He wanted fifty high-quality sons. He talked about his kids marrying Elon Musk’s kids. The Wall Street Journal broke the story last December: his former partner Tang Jing has claimed in a custody dispute that he may have as many as 300 children; Duoyi disputed that number but acknowledged that years of American surrogacy had produced “only a little over 100”; Xu himself says he has custody of 12 children born through surrogacy in the United States. In 2023 a Los Angeles judge noticed the same name appearing on petition after petition. Xu appeared by video from China and told the court he hoped for about twenty US-born children, preferably boys, to inherit the business. Several were being cared for by nannies in Irvine while their paperwork to travel to China was processed. He later posted a photograph of roughly a hundred children sitting in rows.

This week CBS News put it back on the front page. Its reporters interviewed 32 surrogates for an investigation into the industry and found a California woman, identified only as Judy, who had answered an Instagram ad promising $120,000 to carry a child for what the agency described as a single father expanding his family. She learned from CBS, by way of leaked contract documents, that the intended father was Xu Bo, and that he had been using close to two dozen surrogates at once. She had asked repeatedly to meet him. She never did.

How the law works on California surrogacy is fascinating…

One: the Fourteenth Amendment. As a general rule under current law, a child born on US soil is a citizen. The parents can be Chinese nationals, they can be billionaires, they do not have to live here, and their intent is irrelevant. Birthright citizenship turns on the place of birth, not the purpose of it. And the Supreme Court already weighed in this summer: in Trump v. Barbara, decided June 30, it rejected the administration’s effort to deny birthright citizenship to children born here to parents who are unlawfully or temporarily present. The constitutional half of this equation is settled for now.

Two: commercial surrogacy is legal in California, and in several other states, with a well-developed industry around it. You retain an agency, use IVF, compensate a surrogate, obtain a pre-birth parentage order from a state court, and the child is born an American with your name on the birth certificate. There is no federal surrogacy statute at all; it is state law, and it ranges from fully enforceable to outright void depending on which side of a state line you stand on.

Three: nothing counts. No federal cap on how many arrangements one person can commission. No registry. No agency positioned to notice a pattern across states or across years. The only reason anyone noticed here is that a single LA judge saw the same name too many times. Congress has stirred, a little: Senator Rick Scott introduced the SAFE KIDS Act last fall to bar nationals of certain adversary countries, China included, from using American surrogacy. An attempt to pass it in the Senate was blocked in July, and it has not become law. So as of this week, the pathway Xu Bo used is exactly as open as it was when he found it.

And the last piece: China bans commercial surrogacy domestically. So the arrangements went where the law permits them. Lawyers have a dry name for that, jurisdictional arbitrage, and it usually describes choosing a state of incorporation or a governing-law clause. Here the same logic was applied to the birth of children, and the American surrogates involved, by their own account, had no idea of the scale. That is the part that should give everyone pause, whatever they think of the law.

The real lesson. Every legal system has seams where two regimes meet and neither one owns the gap. Family law is state law. Citizenship is constitutional. Nobody ever designed the interaction, because nobody imagined anyone operating it at industrial volume. Somebody with unlimited resources always finds the seam first.

The Athletic went to seventy sources to estimate what all sixty-eight power-conference football rosters actually cost, and the numbers broke my brain.

Oregon sits in the roughly $50 million club and just got dismantled by Oklahoma State, which spends under $20 million. OSU outgained them 554 to 281, ran for 237 yards to Oregon’s 90, had 25 first downs to 12, and held the ball almost 37 minutes. Oklahoma State was 1-11 last year with a 21-game FBS losing streak and came in a 23.5-point underdog. Those same Ducks beat them 69-3 a year ago. Oops!

The rest of the picture: Ohio State publicly spent $20 million on its 2024 title roster, and that same $20 million would now rank near the bottom of the Big Ten; two years later they are estimated at $49 to 54 million. Indiana, which used to be the losingest program in college football history, went 16-0 and sits at $36 to 40 million, top five in the conference. Georgia is $31 to 34 million and Alabama $38 to 42 million, and neither is in the $50 million room. Texas Tech is at $38 to 42 million and dropped $3 to 4 million when their quarterback left amid a pile of NCAA gambling violations. Boston College is last in the Power 4 at $8 to 13 million.

The interesting legal problem is the boundary. When a booster’s car dealership pays a quarterback $2 million to appear in two commercials, is that a market-rate endorsement or capped compensation wearing a costume? The Commission’s whole job is drawing that line one deal at a time, with limited staff, against every athletic department and agent in the country working the other side of it. Anyone who has ever fought about whether a payment is really “consulting fees” or really purchase price will recognize the game. It is the same game. It has tax and other implications…

A few years ago these kids were treated like criminals for taking a free sandwich. Now the only real question is whether the paperwork was filed correctly. I don’t know if that’s progress. In a $20 billion plus industry, I know it’s more fair to the players than what came before.

This week the Ticker is the Wave. When the CEO of JPMorgan puts a number on it, twelve million businesses and nearly ten trillion dollars, the transfer stops being a thesis and becomes a forecast. Everything else in this issue, from the anti-assignment clause to what the Fed just did to acquisition loans, is a footnote to that number.

Last week was the 25th anniversary of 9/11, and the story I posted about him got more response than anything I have written this year.

Welles Crowther was 24, an equities trader on the 104th floor of the South Tower, and a former volunteer firefighter. He carried a red bandana in his back pocket every day because his father gave him one as a boy. One to show and one to blow.

When the second plane hit, he called his mother. “Mom, this is Welles. I wanted you to know that I'm okay.” Those were the last words she ever heard from him.

He did not stay on 104. He went down to the 78th-floor sky lobby, where people were burned and bleeding and lost, and survivors later described a calm young man with a red bandana over his face who found the one working stairwell and started moving people toward it. He carried a woman on his back for seventeen floors. Then he told the group to keep going, and he turned around and went back up. He did it more than once.

At least a dozen people who got out later identified him from photographs as the stranger who saved them. They never knew his name that morning. They knew the bandana. They found his body months later in the wreckage, near the firefighters.

Think about the ordinary lives that continued because a 24-year-old in a suit decided the people on the 78th floor were his responsibility. Birthdays. Weddings. Kids who grew up with their parents. And a mother with one short voicemail.

Welles Crowther. Hero.

One personal note, because the timing was too good to skip.

In 1997, Fast Company published the article that canonized the concept of a “personal brand.” In 2021, I started an anonymous Twitter account after work, mostly because I had things to say about deals and nowhere to say them. This month, nearly thirty years after that original article, Fast Company opened its Personal Brand Special Report with the story of how that anonymous account became SMB Law Group.

An anonymous account. A team that believed it would work. $1.9 billion in closed deals.

I am not putting that here to take a lap. I am putting it here because the entire thesis of this newsletter is that the interesting stories are underneath the obvious ones, and that is true of businesses too. Nobody saw a law firm in a burner account in 2021, including me.

This week Kevin and I get into Dimon finally saying it, whether AI regulation is safety or a bouncer at the door, and what the Fed just did to every acquisition loan in America.

See you next Thursday.

Eric