YWR GP: The Terminal is an iPod.

Why financial information providers need to eat their children

Investment research and the business of providing financial information are in the midst of AI driven disruption (The End of the Seat).

But where are we going with this?

Do 1,000 terminals get replaced with an API used by 1 Claude Agent?

How does the industry price for this?

Where is the Blue Ocean?

Pancras has a solution for the industry.

But it requires eating their own children.

Another PERS View guest post from Pancras Beekenkamp.

---

• The Apple Paradox: A Study in Self-Preservation

• Embedded Infrastructure and the Biological Tax

• The Disruptor Coalition and the Headless Feed

• Claude Cowork and the Decoupling of Intelligence

• The Microsoft Precedent: Eating Your Own Children

• The Bloomberg Gambit: YouTube, Amazon, or Google?

• The Migration to Cognitive Arbitrage

• Solving the Innovator’s Dilemma

One sometimes wonders whether the grandest structural shifts in markets truly arrive with the fanfare of a technological breakthrough, or if they are, in fact, the quiet consequence of a collective refusal. It is perhaps the refusal of an incumbent to dismantle their own fortress that creates the very vacuum a disruptor requires to breathe. We are currently witnessing what is an instructive case study of the digital age: the moment Apple, the master of disruption, allowed itself to be disrupted because it fell in love with the unit economics of a dying paradigm. It is my contention that the financial information industry is now perched upon a similar precipice, caught in the classic tension between the high-margin comforts of the annual seat fee and the technical inevitability of the agentic workforce. To innovate is to risk one’s legacy; yet, the question lingers: is the failure to innovate anything other than a guarantee of obsolescence?

We appear to be moving from an era defined by the scarcity of the golden copy to one defined by the abundance of cognitive arbitrage, a transition that demands a fundamental re-evaluation of what constitutes a data moat. In an age where the cost of processing information is collapsing towards zero, the value of the “screen” as a mediating interface is being interrogated by a new species of digital worker. This note serves as a roadmap through the shifting architecture of financial intelligence. I shall begin by dissecting the historical collapse of the iTunes hegemony—not as a simple failure of vision, but as a crisis of incentive—to understand how the cannibalisation trap functions within a dominant ecosystem. From this post-mortem, I will pivot to the current infrastructure coalition, questioning whether the embedded defences of Bloomberg, LSEG, and S&P Visible Alpha represent a sustainable moat or merely a biological tax that is increasingly at odds with the returns of enterprise AI.

I will then examine the disruptor coalition—the FactSet-Anthropic alliance—to interrogate whether the Model Context Protocol (MCP) and consumption-based pricing are truly the Spotify solution for data, or perhaps a different kind of commodity trap. The arrival of Claude Cowork in January 2026 provides the catalyst for what I term seat destruction, a phenomenon that forces us to consider the digital worker not as a tool, but as a colleague capable of broad-scale integration.

Finally, I will construct the argument for the Bloomberg Gambit, drawing a parallel to Microsoft’s successful cloud pivot to illustrate how a legacy giant might survive by eating its own children. I believe Bloomberg possesses a unique strategic window to transcend its role as a data vendor and become the unassailable, sovereign platform for financial data exchange—a Google-like reference point that integrates the high-end research of the world’s most acute analysts with the independent expertise of Substack and the Superme.ai model. This journey leads us from a world of information arbitrage to the eventual resolution of the dilemma through the creation of a sovereign index of settled, licensed cognition.

The Apple Paradox: A Study in Self-Preservation

To understand the threat to those in the infrastructure coalition like Bloomberg, LSEG, or S&P, one must first deconstruct the collapse of the iTunes hegemony. By 2010, Apple possessed absolute dominance; the iTunes Store commanded 70% of the digital music market, backed by the hardware lock-in of the iPod and a mountain of consumer credit cards. Apple was the apex predator of the music industry, having successfully transitioned the world from physical CDs to digital downloads. They held the embedded infrastructure—the hardware, the payment rails, and the customer relationships. Yet, they lost the future of music to a Swedish startup with a fraction of their capital and zero hardware presence. Why did the master of the ecosystem fail to see the shift in the very ground they stood upon?

Perhaps the hesitation was not a failure of vision, but an inevitable consequence of an incentive structure that rewarded the preservation of high-margin hardware at the expense of an emerging, low-margin utility. Apple’s model was built on the transactional unit: the USD 0.99 download. This high-margin revenue stream was more than just a line item; it was the fuel for a hardware ecosystem where the iPod functioned as a high-margin dongle for the iTunes Store. As Steve Jobs famously argued in 2003, the subscription model of buying music was bankrupt because consumers wanted to own, not rent. He believed that owning a song represented a permanent emotional connection, while a subscription was a fleeting utility.

A sceptic might suggest: was this a genuine philosophical conviction, or was it motivated reasoning designed to protect the unit economics of the past? To launch a streaming service would have meant actively cannibalising a lucrative download business—trading a customer who spent USD 200 a year on individual tracks for one who spent USD 120 a year on all-you-can-eat access. Furthermore, one must consider the internal pressures of record label negotiations. Apple had spent years convincing labels that the USD 0.99 track was the only way to save them from piracy. To pivot to streaming meant telling those same labels that their price floor was about to vanish. It meant telling the board that the hardware margins of the iPod, which relied on proprietary library lock-in, were no longer the primary moat. Apple chose the comfort of the unit; Spotify chose the sovereignty of the index.

Apple’s error, if we are to call it that, was to mistake the medium for the message. They believed they were in the business of selling digital files, when they were actually in the business of managing attention. By insisting on the download, Apple forced the user to act as a librarian—managing storage, syncing devices, and making discrete purchasing decisions for every artistic unit. This created a high-friction experience that was manageable only as long as no alternative existed. Apple missed the fundamental shift in consumer psychology: the migration from ownership to access. Spotify viewed music as a utility to be streamed. Whilst Jobs believed that playlists and libraries were the product, he failed to see that in a world of digital abundance, the value shifts from the item itself to the index of items. The discovery of music—the algorithm that finds your next favourite song—became more valuable than the file containing the song itself.

Spotify became the iPod killer—not because it was a better piece of hardware, but because it made the hardware irrelevant. It was a low-friction technological wedge that turned the iPod into a redundant box of stored files. The iPod killer wasn’t a device; it was a protocol. It was the shift from local storage to cloud-based stream. While Apple guarded its walled garden, Spotify built a bridge to the entire world, commoditising the very thing Apple sought to preciousise. By the time Apple capitulated and launched Apple Music in 2015, they were forced into a defensive fast follower position, burdened by the legacy UI of a previous era. They had preserved the download for five years too long, and in doing so, they lost the platform. They were no longer the ones defining the experience; they were merely trying to match it. It is a choice between guarding the unit of the past and owning the platform of the future; a choice that the current infrastructure coalition is now being forced to make.

The Embedded Infrastructure and the Biological Tax

We now see an identical architecture of displacement forming in the financial data markets. For forty years, the industry has been ruled by the infrastructure coalition. Their bedrock is the annual seat fee, a pricing architecture predicated on a 1:1 ratio between a human operator and a screen. This model is the financial equivalent of the USD 0.99 download. It assumes that value is inextricably tied to human cognitive bandwidth—the ability of a trader to navigate an interface and process information via the optic nerve. It is a world of information arbitrage, where the vendor’s role is to deliver the golden copy of data to a human biological brain. But is the human brain still the primary consumer?

The biological tax is the premium paid for the friction of human mediation. Like the iTunes download, the annual seat fee is a unit of scarcity linked to a biological constraint. LSEG (Refinitiv) and niche providers like S&P Visible Alpha are currently attempting to contain the agentic disruption through walled gardens and non-display restrictions. These are not merely technical features; they are legal fences designed to protect the unit economics of the human analyst. LSEG’s strategy, for instance, requires a user to hold both a Workspace licence and a Microsoft Copilot licence to access AI tools. As I wrote previously in The Agentic Dilemma, this is the equivalent of Apple saying: “You can stream music, but only if you pay for the human licence first.”

One wonders whether non-display restrictions are a defensible moat or a regulatory target in waiting. These terms essentially prohibit an algorithm from seeing or reasoning with the data unless a human eye is present. It is a literal tax on machine intelligence. In an era where McKinsey suggests that AI could reduce an asset manager’s cost base by 25% to 40%, these restrictions act as a drag on the very productivity gains the industry requires. The incumbent’s trap is the belief that their embedded infrastructure—the thousands of terminals installed in bank basements—is an unassailable advantage. Apple thought the same of the iPod. But embedded infrastructure is a weight as much as a weapon; it creates a sentimental attachment to the hardware interface at the exact moment the market is moving to a headless protocol.

These platform players are betting that the terminal is a mandatory heads-up display for human intelligence. But as Azeem Azhar argues, any model that relies on biological limits is fundamentally fragile in an exponential age. If the biological seat is the past, then the unit of value must shift from the screen to the stream. Yet, the question lingers: can the incumbents actually afford to let go? If the vast majority of your revenue depends on the biological tax, can you truly embrace a headless future without inducing a corporate heart attack? To dismantle the fortress while the siege is underway requires a degree of strategic courage that is rarely found in public company boardrooms. It is a transition already being engineered by those with no legacy to protect, who view the terminal not as a fortress, but as a bottleneck.

There is one listed financial information provider making the pivot and starting to get it right.

The Disruptor Coalition and the Headless Feed