Who Can Own the Middle Without Destroying It?

In December 2021, the Federal Trade Commission sued to block NVIDIA’s proposed $40 billion purchase of Arm. The FTC said the deal could damage Arm’s historically neutral licensing role. It also said licensees’ confidential information could reach a downstream competitor.¹

The FTC raised several concerns. One was whether the companies that rely on Arm would keep sharing their plans with it if a competitor owned it. NVIDIA and SoftBank ended the deal in February 2022, citing significant regulatory challenges.² I read the sharing question as the center of this article. The record does not show it alone ended the deal.

An owner selling a middle usually starts with price. What decides the value is whether participants keep contributing after the sale.

A neutral middle is worth what its participants keep contributing. The right owner is one that does not make them stop.

This is the second article in a series about the middle: a business that earns money by connecting market participants or supplying a shared capability they would struggle to provide themselves. The first looked at how AI changes what a middle gets paid for. This one asks who can own it.

Markets create a middle because no one participant sees the whole

Start with how a middle comes to exist. The Supreme Court’s 2011 opinion in Sorrell v. IMS Health describes how pharmacies supplied prescriber-identifying information to data firms. The firms analyzed it and sold reports to drug manufacturers.³

No pharmacy could see the whole market. A firm that gathered from many pharmacies could. IMS Health and Quintiles completed their merger in October 2016.⁴ In November 2017 the combined company took the name IQVIA, and the announcement stressed healthcare data, analytics, and domain expertise.⁵

That is the opening pattern. Fragmented operating information can support a business that sells a broader view of the market. The opportunity is not one more transaction in the channel. It is a view across transactions that no single participant can assemble.

Now the limit. Nothing in these sources shows that the middle was neutral. Neutrality and independence are separate claims, and each needs its own proof: who owns the business, what the contracts allow, who can see what, and who governs it. Privacy and the lawful use of data are open issues, not solved ones.

Competitors can need the same platform and still reject who controls it

Back to Arm. The FTC’s concerns were allegations about likely effects, and the deal never closed. So no one found that harm occurred. I am describing Arm’s role in 2020 to 2022, not its business today.

Still, the mechanism is clear. A middle is useful to competing participants in part because no single one of them controls it. Change the owner, and each participant has to ask what happens to the plans it has shared. The technology can stay the same. The willingness to share can still change.

Ownership is not disqualifying on its own. The risk comes from the owner’s downstream interests and what the owner can do with what it sees. Customers can need the technology and still reject who would control it.

You can own the platform and still lose the network

Here is a case where an owner did change hands and the participants did move. After Oracle acquired Sun, contributors to the OpenOffice software formed an independent project, LibreOffice. Ars Technica reported the fork on September 28, 2010.⁶

On October 31, 2010, departing participants published a public letter. They tied their move to unresolved disagreements about how the project was organized and how much control its sponsor held.⁷

Simon Phipps, a former Sun executive, later described extensive migration of the community. He also disputed that the fork demonstrably hurt Oracle’s business.⁸

So what does the case show? Contributors moved. It does not show that Oracle lost any particular amount, or that the acquisition alone caused the move. OpenOffice had governance tensions before Oracle bought Sun, and this is open-source software, not a healthcare data network.

What it does show is that the code and the brand were not the whole productive asset. The people who kept the project useful could take their effort somewhere else, and they did. Ownership can secure the code without securing the people who keep it valuable.

The middle can consolidate without becoming one of the sides it serves

If a participant-owner is a risk, one answer is for the middle to grow by buying other businesses in the middle. Equinix and Telecity show what that looks like.

Equinix completed its roughly $3.8 billion acquisition of Telecity on January 15, 2016. The companies described a larger European data center footprint and more connections between networks and cloud providers.⁹ The deal also required Equinix to divest eight European assets to Digital Realty to meet merger review requirements.¹⁰

Two things come out of that. Staying neutral does not require staying small. And a neutral service model does not remove concentration concerns, because regulators still asked for divestitures.

I will not say Equinix had to buy Telecity to stay independent. Whether scale protects a middle from being absorbed by one side is my hypothesis, and these sources do not test it.

The natural consolidator may be another company in the middle

On June 9, 2021, Datavant and Ciox Health announced a $7 billion merger. Datavant linked health records while protecting privacy. Ciox handled clinical record exchange and access. The announcement described a neutral connectivity role and listed investors from healthcare as well as finance.¹¹

This is a practical example of two adjacent middle businesses combining. It also corrects a simple rule. You cannot judge neutrality only by whether industry investors are present or absent. A financial owner is not automatically neutral. An industry investor is not automatically disqualifying.

So test the claim, not the label. Who holds control rights? What information can the owner see? How are customers treated? What may the data be used for?

The condition that decides it: where does the value come from?

Article 3 in this series describes the opposite case. Amazon bought Kiva Systems, then kept Kiva’s robots for its own warehouses and stopped selling to outside customers. That was a loss of outside participation, and it can still be a good decision for the buyer.

The two cases fit together. The deciding question is which source of value depends on keeping the middle open.

When outside participation creates the value, as with Arm’s licensees or OpenOffice’s contributors, the owner has to preserve that participation. When use inside the buyer’s own business creates enough value, as with Kiva, the buyer may rationally accept losing the outside business.

Here is how I would apply it. Illustration, not a case. Say you run a business that links records for hospitals, drug makers, and insurers, and you have two offers. One bidder sells to drug makers. The other sells software to hospitals.

Name what could leave under each owner. Under the drug maker, hospitals may stop releasing records, or restrict what you may do with them, and the product loses the records that made it worth buying. Under the hospital software company, drug makers may doubt that they still get the same treatment as hospitals, and may buy their data elsewhere.

Then compare four things: who holds control rights after the sale, what the owner may see, what the contracts allow, and what each side’s best alternative is if it stops sharing. If most of your value comes from both sides contributing, price the chance that one side leaves. If most of it comes from one bidder’s own use of your data, losing outside business may be a price worth paying.

Be clear about what the evidence supports. These cases illustrate mechanisms. They do not establish which ownership arrangements reliably keep participants contributing. Arm was a proposed deal, and the Datavant case rests on the companies’ description of their own neutrality.

Related Video: What Is A Leadership Development Program?

The Bottom Line

A neutral middle is worth what its participants keep contributing. The right owner is one that does not make them stop.

When you weigh a buyer, ask who could combine these activities without undermining participation. If your value depends on outside participants, the answer may be a company next to you in the workflow, not the highest bidder. If your participants would hesitate to share with that owner, price that hesitation before you sign.

If you run a business that connects competitors, customers, or data sources and are considering who should own it, I would like to talk. Schedule a call or email me at jim@jfjordan.com to discuss how to strengthen your position.

Sources and notes

1. FTC Sues to Block $40 Billion Semiconductor Chip Merger, Federal Trade Commission, December 2, 2021. Note: allegations about likely effects, not findings.

2. NVIDIA and SoftBank Group Announce Termination of NVIDIA’s Acquisition of Arm Limited, February 7, 2022.

3. Sorrell v. IMS Health Inc., opinion of the Court, June 23, 2011.

4. QuintilesIMS Merger Creates a Leading Global Integrated Information and Technology-Enabled Healthcare Service Provider, IQVIA, October 2016. Note: merger completed October 3, 2016.

5. QuintilesIMS is now IQVIA, IQVIA, November 6, 2017.

6. Document Foundation forks OpenOffice.org, liberates it from Oracle, Ars Technica, September 28, 2010.

7. Every end is a new beginning, public letter by departing OpenOffice community participants, October 31, 2010.

8. Phipps, S. OpenOffice.org and contributor agreements, LWN, May 20, 2011. Note: first-person account by a former Sun executive.

9. Equinix Expands Data Center Leadership Position With Close of Telecity Acquisition, January 15, 2016. Note: company release; the benefits are the companies’ own description. The sources do not show the deal was necessary for independence.

10. Equinix Agrees to Divest Eight European Assets to Digital Realty, Equinix, 2016.

11. Datavant and Ciox Health Announce Merger, Creating the Largest Neutral and Secure Health Data Ecosystem, June 9, 2021. Note: company announcement, not an independent review of post-merger neutrality or of its governance protections.