K2 Space raises $500M Series D at $6.8B to scale satellites

A satellite manufacturer just picked up a $6.8 billion valuation on the pitch that data centers eventually belong in orbit, and that is a story worth pausing on. K2 Space, founded four years ago by brothers Karan and Neel Kunjur, closed a $500 million Series D co-led by Kleiner Perkins and ICONIQ, Bloomberg reported. That is more than double the $3 billion valuation the company carried out of its $250 million Series C in December 2025, and it brings cumulative funding above $1 billion, matched by more than $1 billion in signed commercial and government contracts.

The near-term business is not exotic. K2 builds unusually large, high-power satellites out of a 180,000 square foot factory in Torrance, California, and it says it produces more than 85% of the platform in-house. The order book is anchored in defense and connectivity work: 30 satellites for SES's meoSphere network, participation in the Golden Dome missile-defense program via an Anduril partnership, and a slot on Space Force's Protected Tactical Satcom-Global program. The company's Gravitas satellite is already on orbit, which is the credential Kleiner Perkins's Lucas Oliveira leaned on when he said K2 has "proved it in orbit."

The longer bet is the AI one. K2 says it will introduce a 100kW Giga-class satellite in the second half of 2028, described in the company's release as the foundational platform for deploying massive amounts of compute on orbit. The subtext is familiar. Terrestrial data centers are running into hard limits on power, land, water, and permitting, and a growing chorus of investors are willing to fund the argument that some of that load ends up above the atmosphere.

The honest caveat is that almost none of this is settled. The unit economics of orbital compute against a hyperscaler rack are unproven, the Giga-class platform is still more than two years from flight, and a heavy defense-contract mix carries its own procurement risk. What the reporting also does not give you is how much of the $1 billion contract backlog is firm versus optioned, or how K2 intends to solve thermal dissipation and downlink bandwidth on a 100kW bus.

The forward-looking read is that capital is now flowing to the picks-and-shovels layer of orbital infrastructure, not only to launch and imagery. If K2 hits its 100-satellites-per-year target, the near-term beneficiaries are the defense primes and connectivity operators buying today; the longer-term winner, if the thesis holds, is whichever hyperscaler is willing to bankroll the first pilot compute rack in space.