Why this particular six

As a proxy for who wins big tech legal work, the basket is unusually well balanced. Google and Meta have been buying outside counsel at industrial scale for two decades; Google's first data centre in Inzai, Japan and its 2007 DoubleClick purchase sit in the same dataset as OpenAI's reorganisation into a Delaware public benefit corporation last October. Anthropic and OpenAI, by contrast, have existed as serious commercial buyers for roughly four years. SpaceX only started generating capital markets work at scale this year.

So a firm that appears across the whole basket is telling you something a single deep client relationship cannot. It built a bench for the 2010s consumer internet, and then got itself invited into the post-2021 compute economy. Plenty of firms managed one and not the other.

The American Lawyer made a version of this argument in January, using Pirical matter data to rank the firms taking on the highest volume of publicly disclosed tech and AI matters over a decade: Latham & Watkins, Fenwick, Perkins Coie and Quinn Emanuel led on all-time volume, while Cravath, Wilson Sonsini, Paul Weiss, Kellogg Hansen, Winston & Strawn, Gibson Dunn and Keker, Van Nest & Peters had taken on more in the preceding two years (The American Lawyer, 6 January 2026). Bloomberg Law has traced the same pattern from the deal side, naming Wachtell on OpenAI's SoftBank round and Sullivan & Cromwell on the xAI and X merger (Bloomberg Law, April 2025).

Matter volume is one lens. It is also the lens everyone already uses. We wanted three.

List one: alumni now sitting in-house

The oldest observation in legal business development is also the least measured one. Back in 2005, then Wilson Sonsini chief executive John Roos told Stanford Lawyer that the legal departments of a significant percentage of the firm's clients were run by its own former attorneys (Stanford Lawyer). Law360 later found that a large share of the highest-earning general counsel in America trace back to a handful of large firms (Law360, via Kirkland & Ellis). Nobody, though, publishes the map.

Methodology

Ranked by the number of current in-house lawyers and in-house legal directors at the six MANGOS companies who previously worked at the firm. The second column strips out anyone whose only stint at that firm was a summer associateship or internship. Source: Pirical Legal Professionals export, 538 profiles, 3 September 2026.

List two: partners with a strong personal connection

Methodology

Ranked by the number of partners holding a Strong connection, whether career or education, to one of the six companies. Source: Pirical Legal Professionals export, 794 profiles, 3 September 2026.

This is the list that rewards careful reading, because a career connection and an education connection are not the same asset. Goodwin Procter is the case in point: fifth here on 39 partners, and joint-sixtieth on client work with two disclosed relationships. Seyfarth Shaw ranks twelfth on 20 partners and records no MANGOS client work whatsoever, which is roughly what you would expect of a labour and employment shop whose lawyers have circled through big tech's employment functions without the firm winning the panel seat.

Then the opposite failure mode. Quinn Emanuel does not appear on this list at all, despite ranking second on client work below. Davis Polk manages two. Ashurst Perkins Coie manages eight against 45 client relationships. Those are firms whose grip is institutional and litigation-led rather than personal, which is a different sort of durability, and arguably a more fragile one when a relationship partner retires.

The traffic runs both ways now, incidentally. ALM has reported on general counsel in residence programmes at Wilson Sonsini and Mayer Brown, bringing former tech legal chiefs into firms rather than the reverse (Law.com, via Mayer Brown). Bloomberg Law has since covered AI-native firms recruiting out of Cooley, Goodwin and Fenwick, one of them founded by a former OpenAI counsel (Bloomberg Law, July 2026).

List three: partners with disclosed client work

Methodology

Ranked by partner-to-company relationships drawn from publicly mentioned deals and disputes. A partner who has acted for two of the six companies counts twice. Source: Pirical Legal Professionals export, 816 profiles, 3 September 2026.

Three firms post relationships with all six companies: Latham, Kirkland and Sullivan & Cromwell. Sullivan & Cromwell does it on a frontier-lab tilt, with twelve at Google but seven at OpenAI and five at Anthropic, the deepest Anthropic bench of any firm. Cleary's 39 partner relationships at Google is the single largest concentration on any one company by any firm in the data.

Williams & Connolly is the purest specialist on the board: thirteen relationships, all thirteen at Google. Gibson Dunn owns SpaceX with eight, more than double the next firm. Ashurst Perkins Coie, which the American Lawyer ranking picked up under its pre-merger Perkins Coie name, is fifth on volume yet carries almost no Anthropic, OpenAI or SpaceX exposure; it is a consumer-internet practice, not an AI one.

Cross-reference all three lists and only four firms clear the top ten on every measure: Latham & Watkins, Gibson Dunn, Wilson Sonsini and Cooley. Kirkland comes close, held back only by an alumni count that halves once summer associates come out. Morrison & Foerster is second on alumni and seventh on personal connections but seventeenth on disclosed client work, which is a very particular kind of near miss.

What a map like this is for

Panel reviews are won on relationships that predate the RFP. Truth be told, most firms know this and still cannot say, in a pitch, how many of a target's current in-house lawyers trained at their own firm; the information exists in a hundred individual CVs and nowhere in aggregate. The three lists above disagree with each other often enough to be useful, and the disagreements are the point. A firm sitting high on alumni and low on client work has a warm introduction it is not using. A firm high on client work and absent from the connections list has an institutional relationship with no personal redundancy in it.

Both are worth knowing before the incumbent's relationship partner retires.

Data throughout is drawn from Pirical Legal Professionals exports taken on 3 September 2026, covering Am Law 200 firms. Client work reflects publicly mentioned deals and disputes and is not exhaustive.