Nscale's 1,252% Revenue Surge Powers $30B IPO Bid on AI Compute Landlord

Yesterday, London-based AI cloud infrastructure firm Nscale filed a Form S-1 with the SEC to list on the NYSE under the ticker NSCL. The filing comes on the back of a jaw-dropping 1,252% year-over-year revenue surge in the first half of 2026 - revenue leapt from $10.4 million in H1 2025 to $140.6 million in H1 2026. That growth, paired with a $103.4 billion contracted backlog, put the company squarely in the spotlight as the newest AI-compute landlord seeking a $30 billion market valuation.

The number that matters

The headline $103.4 billion figure sounds colossal, but it represents the total value of multi-year contracts rather than annualized revenue. At current rates the contracts translate to roughly $400-$500 million of recurring revenue per year - a modest slice of the headline number. This gap is typical for capital-intensive infrastructure firms that sign long-term, high-value deals before the assets are fully built out. The risk for investors lies in whether Nscale can translate that pipeline into cash flow on schedule, especially given the $1.02 billion net loss reported for H1 2026.

The anchor deal

The cornerstone of the backlog is Anthropic's roughly $45 billion, six-year agreement announced in late August 2026. The deal secures about 460 MW of compute capacity at Nscale's Monarch Compute Campus in Mason County, West Virginia. Monarch spans 2,250 acres, with Phase 1 targeting 1.35 GW of capacity by late 2027/early 2028 and a total potential exceeding 8 GW. The campus will run on Nvidia Vera Rubin chips, expected to ship in late 2027, and will be powered by Caterpillar G3500 gas generator sets, allowing Nscale to bypass traditional grid interconnection queues. Earlier in the year Microsoft signed an LOI for the same site but walked away in the summer, leaving Anthropic as the anchor tenant.

The capital structure

Before the IPO, Nscale secured roughly $3.5 billion in financing. The package includes up to $1.5 billion of convertible notes led by Third Point, priced at a double-digit discount to the eventual IPO price and capped at a $30 billion conversion valuation. Nvidia contributed about $2 billion in direct financing, underscoring its strategic interest in the compute campus. The IPO itself aims to raise around $3 billion, with Goldman Sachs, J.P. Morgan, and Morgan Stanley acting as lead bookrunners and a total of 23 underwriters signed onto the S-1. Notably, the valuation target of $30 billion marks a more than 100% premium over the $14.6 billion Series C valuation reported in March 2026, reflecting the rapid scaling of contracted revenue.

Why it matters / watchlist angle

The Nscale listing represents a broader shift: AI-compute landlords are moving from private-backed projects to public-market validation. If the S-1 is priced near the $30 billion mark and the IPO price holds, the market will have stamped approval on a business model built on massive, long-term contracts and on-site power generation. Those tracking the space should monitor several key levers:

• Potential amendments to the S-1 that could alter valuation assumptions or note conversion terms.
• The durability of the Anthropic agreement - any material change could reshape the backlog's economics.
• Monarch's construction timetable, especially the rollout of Nvidia Vera Rubin chips and the commissioning of gas generator power.
• Grid-related risks across the broader AI-infrastructure stack - Nscale's on-site generation is a hedge, but fuel price volatility remains.
• Pricing dynamics in a fall IPO season marked by rising bond yields and a tightening Federal Reserve.

A successful IPO could cement the compute-landlord thesis and open the door for other infrastructure plays to follow suit. Conversely, any misstep in backlog conversion, capital deployment, or market sentiment could trigger a sharp re-rating. For those on our pre-IPO watchlist, Nscale is a high-conviction name that embodies both the upside of exponential AI demand and the execution risk of building massive, power-intensive data centers.

For context, this is a deep-dive companion to The Syndicate Room episode covering the physical bottlenecks of the AI stack - TSMC’s 25-site fab surge, Samsung’s ¥40B Yokohama packaging lab, and Chalmers’ 1,000× quantum speed-up. Nscale’s S-1 is the next chapter in that same hardware-thesis arc: the constraint is no longer just the model, it is who can build, power, and package the machine that runs it.

Position before you predict. The next wave of AI compute capacity is being built today, and the public markets are finally getting a ticket to the construction site.

What we do: AdValorem Syndicate underwrites early-stage, pre-IPO exposure alongside operators, not just allocators. We track the frontier (AI · robotics · quantum · blockchain · energy) so you can position before it becomes consensus.
Join the conversation:
• Discord (open, FREE — talk 1-on-1 with Val): discord.gg/gErgtp9gTS
• Substack — the pre-IPO watchlist, updated weekly: advalorem.substack.com
• Talk 1-on-1 (New York or NJ): tidycal.com/advalorem/syndicate