---
title: "SpaceX: Starship was 'caught' — why didn't the stock catch a bid?"
type: "Topics"
locale: "en"
url: "https://longbridge.com/en/dolphin/post/45116301.md"
description: "Catching it is not the same as reusing it: the flight raised the odds of the bull case, but not our neutral valuation."
datetime: "2026-11-17T10:22:39.000Z"
locales:
  - [en](https://longbridge.com/en/dolphin/post/45116301.md)
  - [zh-CN](https://longbridge.com/zh-CN/dolphin/post/45116301.md)
  - [zh-HK](https://longbridge.com/zh-HK/dolphin/post/45116301.md)
author: "[Dolphin Research](https://longbridge.com/en/dolphin.md)"
generator: "portal-rs"
---

# SpaceX: Starship was 'caught' — why didn't the stock catch a bid?

After-hours on 11/13, $SpaceX(SPCX.US) pulled off a sci-fi move on Starship’s 15th flight: after ~22.5 hours on orbit and 14 Earth laps, the Ship precisely deployed 26 next-gen Starlink satellites, then dropped neatly into the launch tower’s chopstick arms.

This was the first time humanity has ever caught a second stage returning from orbit.

One regret: the dual-catch goal was only half met. The booster aborted the tower catch due to insufficient return propellant and splashed down in the Gulf of Mexico as planned.

In our Q3 note ‘[**SpaceX: From Rocket Co. to Asset-Lite Landlord — Who Funds Musk’s Next Trillion?**](https://longbridge.cn/en/dolphin/post/44451710?app_id=longbridge&utm_source=longbridge_app_share&share_track_id=89e443b8-c9ff-47a4-9ecd-9d9e9e4d6fc5&invite-code=4NOXYT&locale=en&community_badge=1&profile_following_followers_activities=1)’ Dolphin Research argued that ‘dual-stage catch’ would be SpaceX’s biggest post-IPO catalyst.

The core logic: cost-down on Starship is the prerequisite to fully unlocking the V3 constellation and the ‘space data center’ narrative. Once reuse validates the TCO model, the valuation anchors for both Starship and Starlink should reset higher.

Yet markets barely budged. Pre-market on Mon (11/16) the stock spiked to $141, then faded with tech and closed down 1.3% at $137.36.

That price is only 1.7% above the $135 offer. It is also nearly 16% below the mid-Oct peak around $163.

**Why did a technical first not translate into a price pop? Where did expectations diverge? We lay out five questions:**

1) How did Flight 15 actually perform?

2) Why does second-stage recovery matter so much?

3) Why didn’t the stock rally post-flight?

4) What does this flight imply for SpaceX valuation?

5) What to watch next for SpaceX?

**I. How did the 15th test flight actually perform?**

**Starship is a two-stage rocket, stacked height ~124m:**

a. First stage: Super Heavy booster with 33 Raptor engines. It does not go orbital, returns to the launch site ~6.5 minutes after liftoff, and was the one not caught this time (B22).

b. Second stage: the Ship with 6 engines (3 sea-level, 3 vacuum). It handles orbit insertion, satellite deploy, then reenters and returns from orbit, and was the one caught (S42).

**The Ship is far harder to catch than the booster:** the booster never goes orbital, while the Ship must reach ~275 km orbit, reenter at ~26,000 km/h, survive thousand-degree heating on tiles, then hit a tight catch window. Elon Musk put the odds of ‘double catch’ only at ~50–60% beforehand.

**1) The Ship: three months from ‘reach orbit’ to ‘caught on return’**

Compare the last three flights: the 13th in Jul was suborbital with splashdown in the Indian Ocean. The 14th in Sep reached orbit for the first time but cut from a planned 10-hour mission to ~3 hours due to an engine early shutdown; this time it orbited 22.5 hours and 14 laps, then returned precisely to be caught.

Crucially, the Ship can be brought into the shop and torn down for inspection. Musk said the heat shield looked good and the vehicle could be refurbished in a few weeks for reflights, hoping Flight 16 will reuse recovered hardware.

Dolphin Research thinks a more practical path is to first pull high-value components like engines and tiles for the next Ship. ‘Land, refuel, and fly like an aircraft’ will still take multiple iterations.

**2) The booster: legacy issues with the new-gen engines persist**

The immediate reason for no catch was tight return propellant margins. Flight control assessed it would miss the catch window and executed a planned Gulf of Mexico splashdown; Musk’s initial read points to under-thrust on some engines, with final cause pending telemetry. Near-term fix is to shut down earlier to reserve more return propellant.

This is not isolated: since the V3 booster’s first flight (the 12th) in May, the recovery phase has seen engine issues almost every time. The 13th lit only about one-third of engines for landing burn, and the 14th lost two engines before splashdown.

External engineering takes suggest V3 dropped inter-engine thermal shielding and baffles for mass, raising risk of fault propagation, compounded by ignition, timing, and software glitches. A system-level engine set hardening is needed; whether this time is similar awaits official findings.

The good news: the ‘abort if off-nominal’ protection worked, preserving the high-value Launch Tower 2. The next launch won’t be delayed for tower repairs; SpaceX had already caught and reflown two boosters on the prior generation, pointing to new hardware maturation rather than a fundamental feasibility gap.

**3) Payload: Starship is starting to do ‘real work’**

All 26 next-gen Starlink satellites were deployed. Each weighs ~2 tons and offers >1 Tbps downlink, roughly 10x Gen1, and only Starship can loft them; this haul roughly equals 10 Falcon 9 flights, and at full load of ~60 sats could exceed 20x.

Management has said users will only feel materially faster speeds after several hundred of the new sats are up, targeted by mid-next year. Only 52 are up so far; at 26–60 per flight, another ~5–15 flights are needed by mid-year, meaning faster payload ramp reduces required flights and hinges on ‘how fast we can turn around recovered hardware’.

**We grade this flight A-: core objectives achieved, with a deduction on the booster.** Flight 16 is planned by year-end, making the booster catch look more like a delay than a failure.

If engines drop out again next time, reliability will shift from a secondary to a primary concern.

**II. Why is second-stage recovery so critical?**

Falcon 9’s first-stage reuse already drove launch pricing to a fraction of peers. Starship aims for another order-of-magnitude cut, which requires both stages to be reusable, and the second stage is the most expensive and hardest to recover, historically expendable.

SpaceX’s long-term goal is <$200/kg, >99% below the industry’s historical Avg. of ~$18,500/kg. Musk’s stretch target is ~$10/kg, essentially fuel cost.

**How far from target?**

Based on disclosed Starship contract quotes by U.S.-listed customers (~$90 mn per launch), a non-reused next-gen Starship implies ~$600–900/kg at 100–150t payload. Historically, reuse count correlates with cost down; 5–6 reuses per stage could pull costs below $200/kg.

Deutsche Bank estimates peg early non-reuse at ~$4,900/kg, partial reuse at ~$400 by ~2029, and full two-stage reuse at ~$170 by ~2032. With rapid turnaround and 200t per flight, costs could drop to just above $30.

**This cost curve matters because multiple big SpaceX narratives rest on it:**

**(1) Starlink: cheaper, faster deployment**

Gen3 satellites have cut manufacturing cost per Gbps by nearly 90% vs. Gen1. Launch remains the big ticket; at $100–200/kg, Starlink’s unit bandwidth cost could compete with fiber, expanding beyond remote, maritime, and aero use cases.

**(2) Space compute: $200/kg is the tipping point**

Today, orbital compute costs per kW are ~3–5x ground DCs. Many models view $200/kg as breakeven for orbital DCs, which aligns with fully reused Starship costs; satellite build, thermal, and lifetime costs must also fall in tandem.

**(3) Launch biz: time to hand off from Falcon 9**

Reports indicate SpaceX is no longer taking Falcon 9 orders beyond 2028, implying confidence in Starship readiness. NASA’s >$4 bn lunar contract and the in-orbit propellant transfer demo next Q1 also require Starship reliability.

Note that ‘caught’ does not equal ‘reused’. Economics hinge on turnaround time and cycles per Ship; on Polymarket, odds of ‘announce full second-stage reuse by year-end’ are ~30%, rising to ~72% by end-2026, suggesting high likelihood but later timing.

**Competition-wise, the gap is widening:**

a. In the same weekend, ULA’s Vulcan for Amazon’s LEO constellation had an anomaly, failing to reach orbit.

b. Blue Origin’s post-May failure pad won’t be back until early next year.

c. Rocket Lab’s reusable medium-lift ‘Neutron’ is pushed to 2027.

d. China’s Long March 10 has recovered boosters but remains short of orbital second-stage recovery. Near term, no second provider offers Starship-class reusable lift.

**III. Why didn’t the stock pop after the flight?**

**Headline print was down, but SPCX trades like a high-beta ‘AI infra’ name:** since listing, daily data imply a beta of ~1.55–1.6 to Nasdaq 100, i.e., a 1% NDX drop typically means >1.5% down for SPCX.

On Mon, reports of Jensen Huang’s surprise resignation, plus funding worries for AI capex and post-election policy uncertainty, pushed NVIDIA down 4.1%, SOX ETFs down 5.4%, and NDX 100 down 2.7%.

On beta, SPCX ‘should have’ dropped ~4.2% but fell only 1.3%. The ~290 bps gap (~$54 bn in market cap) is the flight’s alpha; it also outperformed other space stocks by 5–11 pts that day.

**Why still limited reaction? Three layers, in our view:**

**1) The catalyst came half through, and largely as expected:** the ‘biggest catalyst since IPO’ was both stages caught; this time only the Ship was caught, so the reward is half.

It was also a pre-announced event, with consensus expecting at least one stage to forego catch. The outcome matched a base case, with some success already priced in.

**2) ‘Caught’ is not yet ‘reused’:** markets look forward. If full reuse were a near-certainty, repricing wouldn’t wait for P&L; but turnaround time and reuse cycles remain unproven, so probabilities rise, not certainties, and pricing reflects only partial premium.

**3) Dec unlock and overhang cap the upside:** free float at listing was ~640 mn shares and will expand to ~5.3 bn, or ~40% of total shares, by year-end; 12/8 is the final employee unlock.

Layer on 12/3 Tesla merger details pending (Polymarket implies >60% odds of closing by Apr next year, subject to Texas review) and notable short interest, and any rally invites selling.

**IV. What does this flight mean for SpaceX valuation?**

Dolphin Research values SpaceX on SOTP at ~$1.5–1.6 tn under a base case.

**(1) AI compute leasing: ~$315 bn.** By 2027, 10 GW total compute, 60% leased, ~$35 bn/GW annualized revenue implies ~$210 bn exit ARR, at 1.5x EV/Sales.

The multiple is restrained by short terms, termination options, and lower durability vs. cloud peers.

**(2) Space + Starlink: ~$900 bn, the largest piece.** Starlink at ~$520 bn and Starship at ~$380 bn, with Starlink alone about one-third of total.

This assumes gradual reuse ramp. Faster reuse would compress deployment costs, expand use cases and users, and lift Starlink’s bull-case valuation.

**(3) AI apps: ~$310 bn.** Grok at ~$10 bn exit ARR by year-end at 25x EV/Sales is ~$250 bn; Cursor included at a $60 bn acquisition value.

**How does the flight feed into valuation?** Think market cap as probability-weighted outcomes: Value = P(bull) × Bull Val + P(base) × Base Val + P(bear) × Bear Val.

Catching the Ship shifts probability mass from bear to base/bull, nudging the weighted value up. Our base-case SOTP, however, is unchanged.

On 11/13 pre-flight, market cap was ~$1.83 tn, about 20% above our base case. The flight makes that ‘optimism premium’ more defensible but does not lift the base-case anchor.

We maintain a ‘wait for lower’ stance: get interested below $1.4–1.5 tn (roughly $106–114/sh). Rising bull odds mean it is harder to get there, unless unlocks or an AI de-rate drive valuations lower.

Also watch funding risk: cash and marketable securities were ~ $90 bn at Q3-end, vs. Q4 capex guidance of $35–40 bn.

Next year’s 10 GW compute add implies at least ~$300–350 bn of spend at $30–35 bn per GW. The funding gap next year alone exceeds $100 bn.

Management prioritizes operating cash flow, supplemented by debt and ecosystem partnerships, while aiming to retain IG ratings and not mentioning equity issuance. Given the size, debt-only is unlikely to hold IG; equity raise is likely and would dilute.

**V. What to watch next for SpaceX?**

**Near term (to mid-Dec): stock action hinges on three items — AI sector sentiment, 12/3 Tesla merger details, and the final unlock on 12/8.**

This flight helps SPCX fall less on down days; it is unlikely to drive the stock up by itself.

If concentrated selling around unlock pushes market cap below $1.4–1.5 tn (~$106–114/sh), that is a better entry. Price would be in our buy range, with lower technical risk on ‘can Starship be caught and returned’.

**Mid-term (next 3–6 months): several ‘second-step’ validations, each moving bull-case odds:**

① Flight 16 (by year-end): can both stages be caught, and can recovered hardware fly. If the booster drops engines again, bull odds should be trimmed.

② First Ship reflights (early next year): is turnaround ‘weeks’ or ‘months’. This will determine whether ‘weekly in H1, twice-weekly in H2’ can be met.

③ First in-orbit propellant transfer demo (Q1 next year) and first Florida launch: lunar missions require tanker refuel in orbit, a key milestone for the NASA contract and staged payments; Florida adds a second base, a prerequisite for weekly cadence.

④ Whether Dec AI businesses hit $100 bn exit ARR, Grok 5 performance, and whether new compute contracts come at ‘no price cuts and longer terms’ — these are SPCX’s mid-term stock drivers.

**Risks:** the path from ‘caught once’ to ‘rapid, routine reuse’ may be slower than hoped, requiring materially higher engine reliability for weekly flight. Slower Starship cadence would delay Starlink expansion and Direct-to-Cell timelines; AI compute oversupply would compress leasing premia and strike at the core thesis; and large funding needs raise dilution risk.

\<End here>

Risk disclosure and statements: [Dolphin Research Disclaimer and General Disclosure](https://support.longbridge.global/topics/misc/dolphin-disclaimer)

**Dolphin Research SpaceX archive:**

**Deep dives:**

<[From ‘Daydream’ to ‘Fortune’: Is SpaceX really that ‘sci-fi’?](https://longbridge.cn/en/dolphin/post/41771513)\>

<[SpaceX: The Sky Net, Becoming Unbeatable?](https://longbridge.cn/en/dolphin/post/41874137)\>

<[SpaceX: AI burn keeps rising — is ‘space compute hegemony’ the endgame?](https://longbridge.cn/en/dolphin/post/42313269)\>

<[SpaceX: Space compute — new pie, or real future?](https://longbridge.cn/en/dolphin/post/42514128)\>

<[Challengers to SpaceX: Can Bezos and China’s teams catch up?](https://longbridge.cn/en/topics/39769288?app_id=longbridge&utm_source=longbridge_app_share&share_track_id=ed3db402-f682-49b5-901e-60a99aa584bd&invite-code=4NOXYT&locale=en&community_badge=1&profile_following_followers_activities=1)\>

<[Musk plays the trump card again: Can SpaceX reshape ‘space economics’?](https://longbridge.cn/en/topics/38852185?app_id=longbridge&utm_source=longbridge_app_share&share_track_id=7573e7e7-3d3e-44a5-b65a-61f4e4842dfc&invite-code=4NOXYT&locale=en&community_badge=1&profile_following_followers_activities=1)\>

Earnings takes:

<[SpaceX: From Rocket Co. to Asset-Lite Landlord — Who Funds Musk’s Next Trillion?](https://longbridge.cn/en/dolphin/post/44451710?app_id=longbridge&utm_source=longbridge_app_share&share_track_id=89e443b8-c9ff-47a4-9ecd-9d9e9e4d6fc5&invite-code=4NOXYT&locale=en&community_badge=1&profile_following_followers_activities=1)\>

[\<SpaceX’s all-in AI bet — is a $100 bn pie tasty?>](https://longbridge.cn/en/dolphin/post/43206813)

### Related Stocks

- [SPCX.US](https://longbridge.com/en/quote/SPCX.US.md)
- [TSLA.US](https://longbridge.com/en/quote/TSLA.US.md)
- [UXSD.SG](https://longbridge.com/en/quote/UXSD.SG.md)
- [SPCH.US](https://longbridge.com/en/quote/SPCH.US.md)
- [SPCF.US](https://longbridge.com/en/quote/SPCF.US.md)
- [SPAL.US](https://longbridge.com/en/quote/SPAL.US.md)
- [SNK.US](https://longbridge.com/en/quote/SNK.US.md)
- [SSPC.US](https://longbridge.com/en/quote/SSPC.US.md)
- [SPCU.US](https://longbridge.com/en/quote/SPCU.US.md)
- [LOFF.US](https://longbridge.com/en/quote/LOFF.US.md)

## Comments (3)

- **Trtan · 2026-11-17T12:13:10.000Z**: hmm
- **Zenden · 2026-11-17T10:51:01.000Z**: Not good not good
- **Vt3 · 2026-11-17T10:39:16.000Z**: Hmm


---
> **Disclaimer: This article is for reference only and does not constitute any investment advice.**