← All posts / Industry

From OpenAI's No. 2 to the Boardroom: Fidji Simo Joins Nscale Ahead of Its US IPO

Two months after stepping down as OpenAI's AGI Deployment chief, Fidji Simo has joined Nscale's board — bringing consumer-scale product instincts to a $45B-valued AI infrastructure play racing toward a US listing.

From OpenAI's No. 2 to the Boardroom: Fidji Simo Joins Nscale Ahead of Its US IPO

Two months after stepping away from OpenAI’s day-to-day leadership, Fidji Simo has picked her next seat — and it is not where most people guessed. The former CEO of AGI Deployment at OpenAI, once described as the company’s second-in-command, has joined the board of directors of Nscale, the London-based AI data center and cloud infrastructure startup. The appointment, announced September 11 and effective immediately, lands just as Nscale finalizes preparations for a US initial public offering that could raise as much as $3 billion.

The appointment

Nscale Global Holdings confirmed that Simo joins a board that already reads like a masterclass in consumer-tech and media governance: Sheryl Sandberg, Susan Decker, and former Meta president Nick Clegg all hold seats. Bloomberg, which first reported the news, framed the hire as the data center startup “bolstering” its governance bench ahead of its listing. Quartz noted the company is eyeing an IPO at a valuation of roughly $45 billion.

For Simo, the move caps a remarkable two-month transition. On July 9, she announced she would step down from her full-time role at OpenAI — where she oversaw product and business as CEO of AGI Deployment, responsible for ChatGPT and the company’s application portfolio — after a medical leave proved “longer and harder” than expected. She remained a part-time advisor to the company. At the time, few would have predicted her next formal role would be in the capital-intensive world of GPU data centers rather than at another frontier lab or consumer application company.

Why Nscale wanted her

On the surface, a product executive joining a data center operator looks like an odd fit. Look closer and the logic is sharp.

Nscale sells AI compute — raw GPU capacity, managed Kubernetes, and inference infrastructure to enterprises and AI labs. Founded only in 2024, it has grown at a velocity that mirrors the demand curve for AI training and inference. In March 2026 it closed a $2 billion Series C led by Aker ASA at a $14.6 billion valuation, with Goldman Sachs and J.P. Morgan acting as joint placement agents. By September 4, Reuters reported the company was in talks to raise approximately $3.5 billion in pre-IPO financing. One week later, its newest board member was announced.

The through-line: Nscale’s customers are exactly the kind of product organizations Simo has spent her career building and selling to. Before OpenAI, she was CEO of Instacart, where she took the company public in 2023 during one of the most hostile IPO windows in recent memory. Before that, she spent nearly a decade at Meta leading the Facebook app — arguably the largest consumer product on Earth, serving billions of users. She knows what scaled consumer AI products need from infrastructure, because at OpenAI she was the one signing enormous compute contracts to serve hundreds of millions of ChatGPT users.

An IPO prospectus needs a growth narrative, and Simo’s presence helps Nscale tell a story that goes beyond megawatts and GPU counts: that the company understands the workloads, the buyers, and the product velocity of the AI application layer it serves.

The IPO context

Nscale is joining a crowded queue. 2026 has been the year the “neocloud” sector went to public markets, with AI infrastructure providers racing to list while investor appetite for compute exposure remains hot. Nscale is working with Goldman Sachs and JPMorgan Chase on a US listing that reports suggest could come as soon as this fall, targeting a raise of up to $3 billion at a valuation near $45–50 billion — a figure that would represent roughly a 3x jump from its March private mark.

The fundamentals investors will scrutinize: Nscale has reportedly secured around $51 billion in total contracted revenue and is developing a pipeline approaching 10 gigawatts of data center capacity across the UK, Europe, and North America. Those contracted-revenue figures are the sector’s favorite metric — long-dated, take-or-pay-style commitments from AI labs and enterprises that make capacity economics look more like an infrastructure utility than a startup.

But the risks are equally structural. The neocloud model leans on enormous debt and vendor financing, chips depreciate fast, and the entire edifice assumes sustained growth in AI inference demand. Competitors — from CoreWeave to hyperscalers themselves — are building comparable capacity. An IPO at $45 billion will demand that Nscale convince public-market investors its contracted backlog converts into durable cash flow.

What it signals about talent flows

The appointment is also a data point in one of 2026’s most-watched meta-trends: where senior AI talent goes after the frontier labs. Simo’s trajectory — from Meta’s consumer empire, to Instacart’s IPO, to OpenAI’s deployment org, to the board of an infrastructure company — traces the industry’s own shift in perceived value. A year ago, the most prestigious move for a departing OpenAI executive would have been to another model lab. Today, the compute layer itself has become the strategic high ground, and board seats at companies that own it are among the most coveted perches in tech.

Sandberg, Clegg, and Decker on one board already signaled Nscale’s ambition to be seen as an institutional-grade company rather than a venture bet. Adding the executive who ran ChatGPT’s business completes the picture: governance credibility from consumer tech, and domain credibility from the largest AI application ever deployed.

What to watch

The next milestone is the IPO filing itself. If Nscale prices at the rumored $45 billion level, it will instantly become one of the most valuable infrastructure listings of the year and a bellwether for public-market appetite toward AI compute. Watch for the S-1’s disclosure of revenue concentration — how much of that $51 billion backlog depends on a handful of AI lab customers — and for any disclosure about debt structures tied to GPU purchases.

As for Simo: a board seat is rarely the end of the story. Her portfolio now spans advising OpenAI and governing one of its potential European compute suppliers. If Nscale’s listing succeeds, don’t be surprised if her next move is a chairmanship — or a return to an operating role, this time on the infrastructure side of the AI economy she helped build on the application side.

Sources are listed in the post metadata and rendered below.