Stoke Space: Building the World's First Fully Reusable Rocket

Executive overview

Every rocket mission today throws away its upper stage — a multimillion-dollar machine that burns up at 17,000 mph and 2,700°F on reentry. Stoke Space is solving this. Their Nova rocket pairs a high-efficiency first stage with an Andromeda upper stage that uses a liquid-hydrogen heat exchanger and 24 thrusters to survive reentry and land intact.

Full reusability changes the economics of space: launch frequency can scale without scaling factories. Andy Lapsa and Tom Feldman left Blue Origin in 2019, built a prototype engine in a shipping container in Tom's backyard, went through YC, and have since raised ~$990M.

Full reusability of the upper stage is the last unsolved problem in commercial rocketry — and it unlocks an entirely new space economy.

The reusability problem

• Only ~150 commercial launches happen per year; most slots consumed by Starlink
• Industry can now reuse first stages, enabling scale from ~10–20 to 150 launches/year
• Second stages still discarded on every mission — multimillion-dollar hardware, single use
• Cost and limited availability block new verticals and customers

Nova and Andromeda architecture

• Nova is a two-stage orbital rocket with a high-efficiency first stage engine — one of the highest fuel-efficiency engines ever produced
• First stage returns and lands downrange or at the launch site
• Andromeda (upper stage) closes up after payload release and reenters atmosphere
• Custom heat shield uses cold liquid hydrogen in a heat exchanger to absorb reentry heat
• 24 small thrusters control descent angle and landing

From backyard to factory

• Tom and Andy met at Blue Origin as jet propulsion engineers; left to start Stoke in 2019
• First months: Andy building pitch decks in the basement, Tom welding test hardware in the garage
• Prototype engine tested in a shipping container on Tom's property
• COVID hit just as they began fundraising; raised first check ~6 months in
• Now operating from a 168,000 sq ft factory in Kent, Washington, building for ~7 vehicles/year
• Launch site under construction at historic Complex 14, Cape Canaveral

Iteration speed as competitive advantage

• All parts manufactured in-house to keep iteration cycles short
• Test-stand-to-change-to-retest cycle reduced from ~1 month to 1–2 days
• Multiple test articles built in parallel so a failure doesn't halt progress
• Moses Lake structural test facility for cryogenic qualification

BoltLine: internal operations software

• Disposable rockets require no fleet tracking; reusable rockets require continuous maintenance data
• Questions like part service life, preventative maintenance triggers, and failure history must be answered at scale
• Stoke built their own tool — BoltLine — to log, abstract, and route operational data across functions
• Software gap between garage-stage and FAA-regulated human spaceflight is a common failure point in the industry

Fundraising lessons

• Raised $990M to date; operating at high capital efficiency relative to peers
• Early environment: no network, no "rich uncles," COVID froze the market
• YC provided network, fundraising language, and momentum for the round
• Most VCs were SaaS-focused; had to find angels closer to hard tech
• Core advice: get good at hearing no — conviction must survive long rejection streaks