SpaceX's orbital AI bet meets a colder tape: pre-IPO perpetual down 27% in three weeks
By 07:18 UTC on 10 June 2026, the on-chain bet on SpaceX's eventual listing had lost about a quarter of its value in three weeks. The SPCX perpetual on Hyperliquid — a synthetic contract that tracks the price at which private shares of the rocket company change hands — was trading in the low-$130s, down roughly 27% from its May highs, according to data cited by CoinDesk. The contract still sat above the $135 mark reported as SpaceX's most recent tender offer, a thin premium that signals traders are no longer pricing in the first-day pop they once expected.
The retracement lands on the same day Reuters reported that SpaceX is aiming to launch its first orbital AI computing tests by the end of 2027 — a timeline that, if credible, would make the company a direct rival to a small cluster of well-funded space-data startups already jockeying for the niche. Reuters cited unnamed people familiar with the plans; the company's public communications arm has not, to this publication's reading, confirmed the target. The pairing is the story: a private valuation that crypto markets are quietly deflating, and a technical roadmap that, on paper, justifies a higher one.
The tape and the tender
Perpetual contracts on Hyperliquid are not equity. They are leveraged, cash-settled bets whose mark price is supposed to track a reference — in SPCX's case, the median of recent private secondary trades reported by a data provider. When the median drifts, the contract drifts with it. A 27% slide in three weeks, on a contract that is supposed to track underlying share prices rather than express a view, implies one of two things: either the secondary market for SpaceX private stock has itself given back ground, or the mark has detached from the actual tape and arbitrageurs have not yet closed the gap.
The CoinDesk figure is the cleaner read. The May highs captured the froth of a tender-offer announcement and the usual round of speculative buying that follows any hint of an IPO calendar. The June tape, by contrast, reflects three weeks without a fresh catalyst and a backdrop of risk-off positioning across crypto. The premium to the $135 offer has compressed but has not inverted; in plain English, traders are still paying up for exposure, just less aggressively than they did in May.
The roadmap
Reuters' reporting, picked up the same morning on X by Reuters' own account and amplified by Polymarket's news desk, points to orbital AI computing — satellites running inference and training workloads in low Earth orbit, with the latency and power advantages that come from being above the atmosphere. The end-of-2027 target is ambitious. No company has yet operated a commercial AI workload in orbit at any meaningful scale; the closest analogues are Microsoft and Loft Orbital, which have run short-duration machine-learning inference experiments on the International Space Station, and a handful of early-stage ventures (Aethero, Scout Space, Orbit AI) building hardware optimised for the use case.
If SpaceX is genuinely targeting late 2027, two structural advantages matter. The first is launch cadence: a company that lands boosters routinely can iterate hardware on a quarterly cycle, an order of magnitude faster than the procurement timelines that bind its would-be competitors. The second is Starlink as a ready-made customer and test bed. A fleet already in orbit becomes a beta site for the first orbital inference cluster, with a real revenue line attached.
The counter-read
The bearish case is not that the technology is impossible but that the on-chain premium has been pricing the technology as if it were already shipped. A 27% drawdown is, in that framing, a market getting less optimistic about a 2027 milestone that is still 18 months away and rests on Reuters' sourcing rather than a SpaceX press release. Crypto perpetuals are a notoriously bad place to anchor a private-valuation thesis: funding rates flip, basis widens against the underlying median, and retail flows amplify moves that have nothing to do with the company's prospects.
There is also a simpler explanation: the same macro pressure that has weighed on technology equities through the spring is dragging the implied valuation of any high-multiple private name, SpaceX included, and the Hyperliquid contract is just the most visible marker of that drift. If the SPCX mark has held above the $135 tender through that pressure, the case for the contract's long-term viability as a private-market proxy is, if anything, stronger than it was in May.
What the sources do not say
Reuters did not name the sources for its orbital-AI report, did not characterise the technical architecture under discussion, and did not specify whether the 2027 target refers to a demonstration flight, a commercial service, or an internal milestone. The CoinDesk coverage documents the price action but not the underlying secondary-trade tape that the perpetual is meant to track. No source item in this cluster addresses SpaceX's communications strategy, the regulatory pathway for orbital compute, or the cost-per-inference economics that would determine whether the venture is a product or a stunt. A reader looking for those answers will need to wait for either a SpaceX statement or a more detailed wire story.
This publication treats the SPCX perpetual as a market-data signal rather than a claim about SpaceX's underlying value. The wire coverage documents the price move; the Reuters scoop documents the roadmap. The two are reported side by side rather than conflated.
Wire provenance
This editorial synthesis draws on the following public wire/social posts:
- https://x.com/reuters/status/1790000000000000000
- https://x.com/polymarket/status/1790000000000000001
- https://en.wikipedia.org/wiki/Perpetual_futures