By Jeffrey Dastin and Ross Kerber
SAN FRANCISCO, Oct 27 (Reuters) – Few public companies claim to have the future of humanity on the line while making mundane choices about board composition.
Not so with Anthropic. Having warned that its artificial intelligence could one day cure cancer or kill up to 8 billion people, Anthropic has set up a complex plan for who can govern the company as its technology rapidly progresses, its Monday IPO filing showed.
Its expected $1.5 trillion valuation would make Anthropic the world’s most highly valued public benefit corporation, a setup under Delaware law that means the AI lab must balance investor interests with promoting a wider good.
More than fine print, Anthropic’s setup may have big implications for investors and a public increasingly adopting the San Francisco-based company’s AI. The details show how Anthropic has insulated itself from investor pressure so it can pursue what it says is best for humanity.
Anthropic gave 50.1% voting power over certain matters to its seven co-founders, who, according to the filing, are “distinctly equipped to be stewards of our mission,” Reuters previously reported. Though an independent trust will control a majority of board seats, the co-founders’ supervoting power will extend to elections of the remaining directors and shareholder proposals generally.
The Monday filing revealed still more intricacies to overseeing a business that Anthropic expects will outlive the tight-knit crew that created it.
Once most of its co-founders are out of the picture and their “Class F” supervoting stock is gone, Anthropic’s board will divide into three similarly sized groups, serving three-year terms.
Staggering directors’ reelection like in the US Senate, Anthropic has embraced a structure once popular as a way to stem hostile takeovers. The company has long voiced concern about the risks of AI falling into authoritarian or nefarious hands.
That structure also goes against some corporate-governance tides. In this century, major US companies have largely moved to electing all directors annually, making companies more responsive to shareholder demands on areas like executive pay or strategy.
Anthropic declined to comment.
‘LONG-TERM FLEXIBILITY’
Beyond its board, Anthropic has another governing body with its own set of rules: its long-term benefit trust.
Currently chaired by global health leader Neil “Buddy” Shah, the trust controls a special share class called “Class T” stock that will let it elect a majority of Anthropic’s board. Anthropic has contemplated what happens if that stock goes away, too.
The goal of such eventualities is to “preserve long-term flexibility if circumstances change,” its filing said. Even if someone not bound by Anthropic’s mission to benefit humanity takes hold of its Class T stock, Anthropic’s setup would require directors to follow “provisions relating to our public benefit purpose.”
Ultimately, the trust’s power over the board makes Anthropic a “controlled company” under Nasdaq rules, exempting it from following some typical requirements like having an independent nominating and corporate governance committee. Anthropic is using that exemption to put Daniela Amodei, its president and sister of CEO Dario Amodei, on that committee.
THE LEADERS IN CHARGE
Anthropic set up a safety and risk committee on its board as well, currently made up of Novartis CEO Vas Narasimhan and former Trump White House official Chris Liddell. Billionaire Netflix co-founder Reed Hastings is Anthropic’s current lead independent director.
Still, Anthropic’s seven co-founders — the Amodei siblings among them — remain at the core of governance for the foreseeable future. This partly is through the existence of an entity called “Founder LLC,” which, after the trust, will have the biggest say in board elections.
(Reporting by Jeffrey Dastin in San Francisco and Ross Kerber in Boston; Editing by Peter Henderson)
