[FINANCIAL] Space Infrastructure ·
Orbital Data Centers · DeepTech / VC —
March 30, 2026 — Series A COMPLETED
Starcloud Raises $170M Series A at $1.1B
Valuation — Y Combinator's Fastest-Ever Unicorn Is Building AI Data Centers in
Space
The boldest infrastructure bet of Q1 2026: Redmond, WA-based
Starcloud — founded just 26 months ago — closes a $170 million Series A led by
Benchmark and EQT, reaching unicorn status 17 months after its YC demo day and
becoming the fastest Y Combinator graduate to hit $1 billion in history. Its
thesis: as terrestrial data center construction is gridlocked by power
constraints and permitting delays, orbit offers unlimited solar energy and zero
zoning issues.
DEAL SNAPSHOT
|
ROUND AMOUNT $170 Million — Series A
(in two tranches) |
COMPLETED March 30, 2026 |
|
POST-MONEY VALUATION $1.1 Billion — Fastest
YC Unicorn (17 months post-demo day) |
TOTAL FUNDING ~$200 Million |
|
LEAD INVESTOR Benchmark + EQT (co-led
both tranches) |
MILESTONES First NVIDIA H100 in
orbit / First AI model trained in space (Nov 2025) |
THE PARTIES
|
LEAD INVESTORS Benchmark + EQT + Macquarie
Capital + NFX + Multiple Others Benchmark
(Menlo Park): world-renowned early-stage VC, known for highest-returning fund
in venture history; GP Chetan Puttagunta (six-time Midas lister) joins
Starcloud board. EQT (Stockholm): world's second-largest PE firm with $100B+
AUM, owns 70+ data centers globally — strategic data center expertise
alongside VC conviction. Both firms co-led an initial tranche and an
extension round. Additional investors: Macquarie Capital ($500B AUM, world's
largest infrastructure fund), NFX, Nebular, Y Combinator, Adjacent, 776
Ventures, Fuse Ventures, Manhattan West, Monolith Power Systems. Angel
investors: Gen. Stephen Wilson, former Boeing CEO Dennis Muilenburg, former
Starbucks CEO Kevin Johnson. |
COMPANY / RECIPIENT Starcloud, Inc. (formerly
Lumen Orbit) Redmond,
WA-based space infrastructure company. CEO and co-founder Philip Johnston
(ex-McKinsey). Co-founders: Adi Oltean (ex-SpaceX Starlink, Microsoft Azure)
and Ezra Feilden (ex-Airbus Defence and Space). Founded January 2024 (El
Segundo, CA), relocated to Redmond. Mission: build the world's first
commercial orbital data center network — satellites equipped with GPUs, solar
panels, radiation shielding, and cooling systems — that operate as AI compute
infrastructure in low Earth orbit. Current satellite: Starcloud-1 (Nov 2025
launch on SpaceX; carries NVIDIA H100, first AI model trained in space).
Next: Starcloud-2 (2026, 100x more power, NVIDIA Blackwell B200); Starcloud-3
(200kW, 3-ton, Starship-compatible). FCC application for up to 88,000
satellite constellation filed. |
DEAL OVERVIEW
Starcloud
(formerly Lumen Orbit) closed its $170 million Series A on March 30, 2026, in
two tranches co-led by Benchmark and EQT. The round values the company at $1.1
billion — achieving unicorn status 17 months after Starcloud's Y Combinator
demo day, making it the fastest company in Y Combinator's 20-year history to
hit the billion-dollar mark. Total capital raised since founding in January
2024 now stands at approximately $200 million. Critically, Starcloud launched
its first satellite — Starcloud-1 — in November 2025 on a SpaceX rideshare
mission with just $3 million in pre-seed funding, designing, building, and
launching within 21 months: a pace of execution without precedent in aerospace.
Starcloud-1
achieved three historic industry firsts: (1) first NVIDIA H100 GPU deployed in
orbit; (2) first AI model (LLM) trained in space; (3) first inference run of
Google's Gemini model in orbit. The $170 million will fund: development of
Starcloud-3 (a 200-kilowatt, 3-ton spacecraft designed to launch from SpaceX's
Starship — with a single Starship launch capable of carrying approximately 50
Starcloud-3 satellites, equivalent to ~10 megawatts of compute); a dedicated
3,000-square-metre manufacturing facility in Woodinville, WA; and procurement
of future launch contracts. Commercial workloads begin with Starcloud-2, with
early customers including Crusoe Inc. and partnerships with AWS, Google Cloud,
and NVIDIA.
"The AI revolution is colliding with the physical limits of our
terrestrial energy grid. We are quickly running out of places to build new
energy projects for data centers on Earth. By moving AI compute to space, we
unlock access to unlimited solar power and completely remove the energy
bottleneck."
— Philip
Johnston, CEO & Co-Founder, Starcloud
THE TERRESTRIAL DATA CENTER CRISIS CREATING
THE ORBITAL OPPORTUNITY
The bottleneck
confronting AI infrastructure development is no longer GPU supply — it is power
and permitting. New data centers in the US now face 3-5 year permitting
timelines, utility interconnection queues exceeding 5 years in many states, and
"not in my backyard" opposition from local communities resistant to
the noise, traffic, water consumption, and visual impact of large data center
campuses. Meanwhile, AI training workloads are doubling in compute requirements
every 6-12 months, creating an urgency that terrestrial construction timelines
cannot match.
Starcloud's
orbital solution bypasses both constraints: space offers unlimited solar energy
(orbital solar panels generate approximately 5x more power than terrestrial
panels due to constant sunlight with no atmospheric attenuation), and zero
permitting requirements. The economics are currently challenging (launch costs
mean orbital compute is more expensive per kilowatt-hour than terrestrial for
most workloads), but Starcloud's CEO expects cost parity within 3-5 years as
SpaceX's Starship begins commercial launches at ~$500/kg to orbit.
INVESTORS
Co-Leads
(both tranches): Benchmark (GP:
Chetan Puttagunta, joins board) + EQT
Infrastructure
/ Strategic Investors: Macquarie
Capital + Monolith Power Systems
VC
Participants: NFX + Nebular + Y
Combinator + Adjacent + 776 Ventures + Fuse Ventures + Manhattan West
Angel
Investors: Gen. Stephen Wilson +
Dennis Muilenburg (former Boeing CEO) + Kevin Johnson (former Starbucks CEO)
Customers
(Early): Crusoe Inc. + AWS + Google
Cloud + NVIDIA (partnerships)