The Textbook Says Innovate Your Way Out. The Rails Say No.

In June 2026, the Sparkassen's own broker cut its base order fee by 81 percent, from 4.99 euros to 95 cents. Nobody cuts 81 percent because they want to. In 2024, Germans had opened more than three million net-new securities accounts, and nine of ten went to a single neobroker. S Broker gained about three thousand, in a market that grew ten percent: the banking family whose name is practically a synonym for careful money in Germany stood at the door of the biggest retail investing wave in decades and watched it walk past. The cut was not strategy, it was capitulation, to years of customers choosing an identical product at a lower price next door. This is not a story about the Sparkassen. It is a story about the rails under them.

The German order fee shows the whole history like sediment: 25 to 50 euros at a branch bank, 4.90 to 9.95 at a direct bank, zero to one euro at a neobroker. That ladder is not a price list, it is a timeline, each rung a generation of brokers built on the same rails with less legacy on top: the branch banks carry buildings and staff, the direct banks carry the nineties, the neobrokers carry an app and a licence. And the pull toward the floor is measurable by date: flatexDEGIRO introduced a zero-commission tier in October 2025, Consorsbank made ETF savings plans free the same month, and then came the Sparkassen cut. Every rung is sliding toward the one below it. The bottom is zero.

The usual telling of this story has a protagonist: Trade Republic, the neobroker that forced everyone down. I don't think that story survives the evidence. The neobrokers were a catalyst, not a cause; the problem they exposed had always existed. Their model won on the only levers the rails leave open: no legacy systems, so a cost base the previous generation could not touch; visible fees near zero, because the revenue had moved behind the scenes into payment for order flow; an app and a brand built for people who would never walk into a branch. Price and brand, pulled harder than any generation before them.

It was the same picture in the United States. Robinhood launched zero-commission trading on the same model in 2015. The big incumbents conceded one round in 2017, cutting commissions of eight and nine dollars to five and six, then held that line for two and a half more years. Zero arrived at the majors only when Interactive Brokers announced free trading in late September 2019, and then the United States ran the whole race in nine days: Charles Schwab cut its commission to zero on October 1, 2019, TD Ameritrade lost a quarter of its value in two days and matched within one, and by October 10 E*TRADE and Fidelity had followed them to zero. Inside five months, $39 billion of forced consolidation had followed. Four years of mounting pressure, nine days of collapse. A cause pushes prices down one by one. A catalyst changes what a price can be, and the structure does the rest.

And the structure has run the same experiment elsewhere, with a completely different cast. The Netherlands ran the race from 2013, and it consumed its own children: BUX, the Dutch neobroker, eventually sold to ABN AMRO for less than its investors had put in. France is running it through BoursoBank, the online arm of a universal bank. Different companies, different decades, different regulators, different investing cultures. Same ending. When an experiment produces the same result under every variation of the players, the players are not the cause. The constant, around the globe, is the electronic infrastructure underneath: rails that cannot carry digital innovation or digital experiences, because they are fundamentally stuck with the analog architecture the industry electrified in the 1970s, creating an innovation ceiling.

Look at what a German broker actually sells. The broker itself is a customer-facing application; every trade behind it passes through half a dozen intermediaries: Xetra makes the reference price, market makers quote around it, Clearstream settles the trade, and back-office providers like dwpbank and Baader Bank run the machinery for much of the market. The biggest have begun pulling steps of that chain in-house, a banking licence here, an execution venue there, but insourcing an intermediary does not abolish it. The step still exists, still needs people and oversight and regulatory capital, and still does exactly what the rails specify. You can buy a wing of the factory. You cannot redesign the assembly line. The product is manufactured entirely upstream, and every broker in the country buys from the same factory. Two apps, two brand colors, one factory.

But the rails do something worse than make products identical. They cap what a product can be. Trading hours, settlement speed, the asset universe, the order types: all of it is fixed upstream, identical for everyone, changeable by no one. A broker can polish its app, and within months every app is polished; it can add the same ETFs, the same savings plans, the same crypto everyone else adds. The ceiling is not a metaphor, it is the technical specification of the rails, and no ambition inside one brokerage moves it.

And the ceiling sits this low for a reason that is easy to miss: the rails are electronic, not digital. An order that executes in microseconds starts a settlement process that takes two days, and for most German stocks ownership still traces back to a printed certificate lying in a vault in Frankfurt. What moves at millisecond speed is an electronic message about that piece of paper. Electronic means the old paper process, run by machines. Digital would mean the asset itself exists as data, and the process rebuilt around that fact. I have written before about why this distinction matters (Natively Digital > Digitized); the short version is that this industry electrified its paperwork and called it transformation.

Now the rule underneath the race can be stated exactly. When every competitor is limited to the same capabilities, and for brokers the capability set is the shared electronic infrastructure and the value chain running on it, differentiation collapses to the two things the rails do not standardize: price and brand. That is the entire strategic menu. Brand slows the slide; the Sparkassen held a 4.99-euro price for years after the neobrokers reached one, because trust is worth a premium. But brand cannot stop the slide, because the product is provably identical, and every customer can watch the identical product get cheaper next door. In June 2026, the banking group with more retail customers than any other in Germany cut 81 percent. Price decides. Brand only delays. And the price of an identical product has one destination, marginal cost, which on electronic rails is close to zero.

The race has one more lesson, and it may be the most important: reaching zero ends nothing. A brokerage fee used to be direct, a visible line the customer could compare. At zero the cost does not disappear. It turns indirect: order flow, spreads, interest on idle cash, currency margins. The cost is not hidden, just harder to find: it no longer arrives as a fee, it arrives as reduced investment performance. At its peak, payment for order flow made up a third of German neobroker revenue.

Then the EU ban arrived, and the response is the most instructive part of the whole story. Recovering that revenue through the visible price was never an option, not with the floor at zero and every customer able to walk next door, so the neobrokers moved themselves into the position that used to pay them. Germany's biggest became its own market maker; its one-euro default now executes against its own book, and choosing an outside exchange costs two. The second biggest became a market maker on an exchange it helped launch, then repriced every external venue to 1.99 euros while its own kept the old 99 cents. The economics moved one layer deeper, into spreads they now capture themselves. The visible price barely moved; the plumbing underneath was rebuilt around it. The US wrote this chapter earlier: industry payment for order flow dipped after 2021, then climbed to a record $4.4 billion in 2025, and Schwab, which already earned most of its revenue from interest on customer cash before the cut, leaned on it harder afterward.

The audited numbers at home show the same migration. At flatexDEGIRO, the largest listed German online broker, commission income fell from four fifths of revenue to under sixty percent between 2021 and 2024, while interest income went from a seventh of revenue to more than a third. And that new mainstay is the one line no broker controls: interest on idle cash is priced by the central bank, and the ECB has already cut its deposit rate from four percent back to two. When rates fall, that revenue falls with them, and there is no visible fee left to raise against it.

None of this is bad faith, and none of it is an escape. At the floor, indirect revenue is the only revenue left, and every one of these moves is proof of the pressure, not relief from it: price and cost remain the only levers the field allows. The bottom of a race to the bottom is real. It is just no longer a fee the customer can see; it is opaque cost.

Who survives down there? Only scale. The neobroker that took nine of ten new German accounts earns a profit at one-euro trades, on a cost base nobody with branches can reach. It is not the only one scaling: Scalable Capital passed fifty billion euros in client assets this spring, and flatexDEGIRO custodies more than a hundred billion. But nine of ten new accounts is a different order of pull. Other markets show how this chapter has ended elsewhere: Schwab absorbed its largest listed rival, and the Dutch market consolidated into a handful of balance sheets. Which leaves the most interesting open question in German brokerage: with the floor reached and one player pulling away, what do the dozens of other brokers do from here?

That resolves the strange question the German market poses. Why do dozens of brokers, with real engineering budgets and real ambition, not innovate their way out of a commodity trap? Because they cannot, not on these rails. German brokerage is a textbook red ocean, the crowded, margin-bleeding market Kim and Mauborgne wrote Blue Ocean Strategy about in 2005. But their escape route, creating uncontested market space through value innovation, assumes a company can change what it sells. Every broker here is optimizing inside a box whose walls none of them built and none of them can move. The race to the bottom is not a strategy anyone chose. It is what competition looks like when the product is fixed and only the price and the logo are free.

So I no longer watch fee announcements; they have become predictable to the cent, because the floor is zero and the direction is known. What I watch for is the first broker competing on something the shared rails cannot do today. Settlement in seconds rather than days. Securities that finally stop behaving like paper: the certificates are long gone, but the machinery built to move them survived into the electronic age almost unchanged. An investment strategy built around a single investor rather than one product sold to a million identical accounts. A Sunday market with the liquidity of a Tuesday morning. Any of that would mean somebody found a wall that moves. Until then, every market runs this same race to the same bottom, and Germany is simply the one running it now. The next cut will not be news. Ninety-five cents was never a price. It was a confession.