The Terminal Isn't a Data Product. It's a Tax Wall Street Can't Stop Paying.

Every other media company on earth is fighting the same war: ad rates collapsing, subscriptions churning, readers who pay nothing and complain anyway. Bloomberg sits the war out. It runs a newsroom of thousands, a global TV network, a wire service, and a magazine — and almost none of it has to make money. The bill is paid by a black keyboard with two amber-lit screens that costs $31,980 a year per seat as of January 1, 2025, a 6.5% hike the company tied to 'weighted global inflation over the prior two-year period.'¹ That's not a subscription. It's a tax — and Wall Street keeps paying it.

The official story is tidy: Bloomberg is a data company that uses its profits to fund great journalism. Strike that. Bloomberg is a switching-cost monopoly that uses journalism to protect the data company. The news isn't the mission the terminal supports. It's the moat the terminal pays for.

Why the terminal is the company, and everything else is the brochure

Start with the math nobody can quite verify, because Bloomberg LP is private and files no public financials. The most-cited estimate puts 2024 revenue near $15 billion — and a long-running disclosure pegged terminal sales at more than 85% of the company's annual revenue.⁵ That 85% figure traces to a 2011 disclosure and has been recycled ever since with no fresh primary number behind it, so treat it as a load-bearing estimate, not gospel. But even at the conservative end, the terminal is thought to throw off something like $10–13 billion of the total. Whatever the exact split, the shape is unmistakable: the terminal isn't a leg of the business. It's the body. Everything else is a limb.

Here's the part that breaks the 'data product' framing. You can get most of the raw numbers somewhere else for a fraction of the price. What you cannot get anywhere else is the other 350,000 people. Bloomberg's own product page says the terminal connects 'more than 350,000' decision-makers⁷ — and they don't just read on it, they message and trade on it. The terminal is where the market talks to itself. Leaving doesn't cost you a data feed; it costs you your counterparties. A junior analyst can drop a competitor's feed in an afternoon. A trading desk cannot leave the room where every other desk is standing. That's the whole engine: not the data, the room.

Built on someone else's $30 million, then bought back at a fortune

The founding tells you the model was always about lock-in. In 1981, after Phibro acquired Salomon Brothers and paid Michael Bloomberg $10 million for his partnership equity — a buyout, not the 'severance package' of legend — he and three colleagues built a machine to price bonds.² The Market Master terminal launched in December 1982, and Merrill Lynch became the first customer, ordering its first batch under a five-year non-compete that kept Bloomberg from selling to Merrill's rivals.³ Two years later, in 1984, Merrill put in $30 million for a 30% stake.³ Then watch what happened to that stake as the network compounded: Bloomberg bought back a third of it in 1996 for $200 million, and Bloomberg Inc. purchased Merrill's remaining shares in 2008 for a reported $4.43 billion.⁴ The same equity that cost $30 million to seed cost billions to retrieve. That's not data appreciating. That's a switching-cost moat compounding.

So what is all that journalism actually for?

If the terminal is the body, the newsroom is brand insurance with a circulation. The news appears first inside the terminal — speed and exclusivity that make a $32K seat feel earned, not extorted. The TV network, the magazine, the website: those put the Bloomberg name in front of the same finance audience that signs the terminal checks, and they do it as a marketing line item disguised as journalism. Bloomberg Media is even growing on its own terms — paid online subscriptions reached 740,000 in October 2024 before the company tightened its counting to active sign-ins (625,000 in February 2025, 660,000-plus by early 2026), with media revenue up 7% in the first half of 2025.⁸ Healthy. Also irrelevant to whether the lights stay on. The media division could vanish and the terminal business would barely notice; the terminal could wobble and the media business would have no oxygen at all.

The honest objection: isn't this just an expensive product people choose to buy?

The fair counter is that 'switching-cost monopoly' is too dramatic for what is, after all, a tool people willingly renew. Refinitiv exists. FactSet exists. Cheaper feeds exist, and plenty of cost-conscious firms do cut terminals in lean years. So calling it a monopoly overstates it. Fair — but notice what the price does. Bloomberg raised the seat 6.5% in January 2025 and indexed the rationale to inflation,¹ which is the move of a company that knows demand barely flinches at price. A genuinely competitive product can't lift price into a customer base that openly resents the bill. The alternatives are real at the margins — for the analyst, the back office, the firm trimming costs. They are not real for the trading desk that needs to be in the same chat as everyone it deals with. That's the tell. Competition exists for the data. It does not exist for the room. And the room is what's priced at $32,000.

Strip away the newsroom, the TV anchors, the glossy magazine, and you're left with the actual company: a near-$32,000-a-year tollgate on the one room in finance everyone has to stand in. The price has compounded in one direction for four decades because the cost of leaving compounds faster. Bloomberg didn't build a media empire and bolt on a terminal to pay for it. It built a switching-cost machine, then dressed it in journalism so the world would admire the wall instead of resent the toll. The genius was never the data. It was making sure that to quit the bill, you'd have to quit the business.