By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel largely agreed that SpaceX's high valuation (100x revenue) leaves little margin for error, with the key risk being the potential collapse of Starlink's growth due to regulatory hurdles or market saturation. While some panelists acknowledged SpaceX's potential revenue streams and competitive advantages, the consensus was bearish, with concerns about valuation, lack of durable competitive advantages, and the high capital intensity of Starlink's growth.
Risk: Regulatory and spectrum risk around Starlink, which could cap subscriber growth and cash flow, potentially collapsing the SpaceX thesis.
Opportunity: The potential for SpaceX's launch business to fund Starlink's growth, given the high growth rate and improving unit economics of Starlink.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Big Short legend Steve Eisman says everyone is buying the wrong AI stocks
Shawn Tully
6 min read
Steve Eisman has a simple way to explain why SpaceX is the most absurd stock in America: its revenues are roughly equal to those of the company that makes Froot Loops. The difference? Nobody is valuing Kellogg's at 100x revenue.
If you'll recall, Eisman identified that a housing bubble was building in 2006 and 2007, fueled by an explosion in the issuance of "teaser-rate," sub-prime mortgages, and that a crash was imminent. He famously seized the moment by shorting the home-loan market big time, a move that greatly profited both the trader and his firm FrontPoint Partners, a subsidiary of Morgan Stanley. Michael Lewis made Eisman a Wall Street legend by chronicling his exploits in his 2010 bestseller The Big Short. In the 2015 film version, Steve Carrell played the famously cranky, contrarian (re-named "Mark Baum"), while Marisa Tomei portrayed his wife and co-skeptic, former J.P. Morgan analyst Valerie Feigen ("Cynthia Baum"). Today, Eisman hosts the weekly podcast "The Real Eisman Playbook," a program I highly recommend as much for its rollicking mockery of the group think that dominates the sell-side stock community as its sharp insights on economic trends and knack at nailing the basics that over time, drive outstanding returns to investors.
In a half-hour phone call, Eisman skewered the latest case where he reckons that hype and hysteria are fogging minds—and it's hardly surprising that his new target's none other than the SpaceX phenomenon. "SpaceX has the revenues of Kellogg's, which makes Froot Loops, which I love, but no one is going out of their way to buy Froot Loops," he declares. (The Ferrero Group of Italy bought Kellogg's cereal business; Ferrero's 2025 revenues of $21 billion are indeed close to SpaceX's $19 billion.) "SpaceX stock's being valued at over 100 times revenue, whereas no company of any size has ever had that kind of valuation. By comparison, Palantir is at 50x." Eisman relates that Elon Musk plans to make money from ventures that only exit in the fictional world. "I grew up reading a lot of sci-fi, a ton of it, I've read it all," says Eisman. "Musk and Silicon Valley grew up on it, too. The difference is, Musk and the SpaceX crew take it seriously!"
Eisman notes that a particular source of riches SpaceX hopes to pluck from the planets especially caught his attention, since it's central to the plot of a sci-fi streamer he loves. "In the SpaceX S-1, where they talk about things SpaceX will eventually do, one of them is asteroid mining. Literally, there is really a great, wonderful sci-fi show on Apple TV called 'For All Mankind,' and asteroid mining plays an important part on the show. (Indeed page 71 of the SpaceX S-1 contains the following: "We plan to pursue asteroid mining operations to extract metals and other critical resources from near-Earth and main-belt asteroids, providing abundant raw materials for space-based industries.")
It also puzzles Eisman that Musk is building an enterprise that straddles at least two industries, and may well add another, when the corporate world's moving in the opposite direction. "Why make the company into a massive conglomerate, when conglomerates are totally out of favor? The world is de-conglomerate-izing. People want pure play, and they're in rockets, Starlink and AI." The prospect that SpaceX will buy Tesla, Musk's second largest holding, is especially appalling to Eisman. "Tesla's been a horrible failure for the past several years," he avows. "Every year Musk says we'll have self-driving cars and robotaxis, which he doesn't do, and all we know is that earnings go down year after year. Musk is a cult, so people keep saying 'wait till next year.'"
Eisman holds a dim view of the hyper-scalers' future in AI
Eisman points to a part of the S-1 that's effectively a manifesto wagering SpaceX's future on AI. In fact, its chief AI product sports a brand name that Eisman fondly recalls from his teenage sci-fi enchantment. "Robert Heinlein invented the word 'grok' in his novel 'Stranger in a Strange Land' [1961] about a Martian who comes earth," says Eisman. "In the book, 'grok' means 'deep understanding.'" In contrast to its lofty moniker, Grok the product's a lightweight, claims Eisman. "The S-1 says that SpaceX total addressable market is $28.5 trillion, and the irony is that over 90% of that TAM is AI, which is all about Grok," he intones. "Grok is a third-tier product. I've heard reports that even the engineers in Musk's own space division won't use it because it sucks."
Overall, he says, the outlook for the hyperscalers is darkening fast. "We've seen a sea change in their AI story, and not for the better," he declares. "That's because of two vectors. The first is that for the hyperscalers, AI is becoming increasingly capital intensive. Last year, Alphabet spent $80 billion on AI and funded it from cash flow. This year, it will spend $180 to $190 billion and raised $85 billion in stock. Now, they all have to raise funding through stock offerings because the table stakes get bigger and bigger." He adds that post-IPO, SpaceX will need to keep tapping the capital markets since its recurring cash flow doesn't come close to meeting its hunger for AI-driven capex.
"The other vector is that there are no 'moats' in AI," he contends. "Someone moves to ChatGPT then to Gemini then to Claude. Even if AI is the greatest thing since the invention of the printing press, there are no moats to shield the providers. You don't want to be the hyperscalers selling this highly competitive product where you have to cut prices to win customers. You want to be the suppliers selling them the picks and shovels, the chips and networking gear the hyper-scalers buy to make their products. Those products are highly customized and protected." For Eisman, it's far better to be a Nvidia, Arista or Cisco riding the capex boom than a Meta, Oracle, Microsoft or Alphabet battling a field of fellow behemoths in the brutal arena where the enterprise and retail solutions are easily swappable.
Eisman stresses that he's not advising anyone to short SpaceX. "I have no opinion on what will happen to SpaceX," he says. "From a fundamental perspective, it's ridiculous But a lot of things can be ridiculous for a long time." For this dourest of doubters, Musk's claims for the feats ahead can only happen in the SpaceX founder's head, or in the sci-fi fantasies Eisman grew up on.
Four leading AI models discuss this article
"SpaceX’s valuation requires credible, durable cash flow soon; without clear monetization, the 100x revenue multiple risks a sharp re-rating if funding costs rise or cash burn persists."
Eisman’s take on SpaceX trading at 100x revenue is a blunt sanity check that hype can outpace fundamentals. Yet the article understates SpaceX’s potential revenue rails—Starlink monetization, launch cadence, and government/defense programs could deliver cash flow that isn’t captured by a simple top-line multiple. It also treats AI moats as non-existent; data networks, platform ecosystems, and integrated hardware-software stacks can yield durable advantages for selected players. The risk to Eisman’s view is a path to credible, recurring cash flow that the market deems scalable, plus favorable funding conditions that sustain high multiples longer than expected.
If SpaceX can monetize Starlink at scale and win stable government contracts, the upside could defy a simple 100x-revenue critique, and AI platforms may still sustain durable moats through data and ecosystem lock-in.
"The shift from self-funded AI development to equity-dilutive capex cycles signals that the industry is moving from a productivity boom to a capital-intensive utility trap."
Eisman’s critique of SpaceX and hyperscalers focuses on valuation sanity and the lack of defensive moats in software, but he misses the 'platform' argument. While he correctly identifies that hyperscalers (GOOGL, MSFT) face margin compression due to massive capex—Alphabet’s jump from $80B to $190B is a massive red flag—he ignores that the 'picks and shovels' trade (NVDA, ANET) is already priced for perfection. If AI utility fails to translate into enterprise revenue, the hardware suppliers will crash harder than the software providers. Eisman is right to be skeptical of the 'Grok' TAM, but he underestimates the winner-take-most dynamics of cloud infrastructure consolidation.
The 'no moat' argument ignores that AI integration creates high switching costs; once an enterprise embeds LLMs into its workflow, the friction of migrating to a competitor becomes a formidable, non-price-based barrier.
"Eisman conflates SpaceX's speculative ventures with its core launch and Starlink businesses, which have different risk profiles and growth trajectories."
Eisman's critique hinges on valuation multiples (100x revenue) and competitive moats, which are real concerns. But the article conflates three separate bets: SpaceX's core rocket/launch business (mature, profitable, recurring), Starlink (high-growth satellite internet with defensible spectrum), and speculative ventures like asteroid mining. Comparing SpaceX's blended revenue to Kellogg's ignores that Starlink alone is growing 30%+ YoY with improving unit economics. The 'no moats in AI' argument applies to consumer LLMs, not to Grok's niche use case or to the infrastructure layer. Eisman also sidesteps that SpaceX's capex is largely self-funded via launch revenue, unlike hyperscalers burning $180B annually. His strongest point—conglomerate discount—is valid but assumes markets price it in immediately, which they often don't.
Eisman's 2006 housing call was prescient because he identified unsustainable debt dynamics and fraud; SpaceX's 100x multiple could compress hard if launch demand softens or Starlink subscriber growth disappoints. But unlike subprime mortgages, there's no hidden leverage or systemic risk here—it's a valuation call on a real, profitable business, which is far harder to time.
"Rising AI capex without moats will pressure hyperscaler margins faster than infrastructure suppliers' revenue growth."
Eisman's core warning on hyperscalers' exploding capex (Alphabet jumping from $80B to $180-190B) without AI moats is underappreciated, as easy model switching will force price competition and erode returns on the $28.5T TAM SpaceX claims. Suppliers like NVDA, ANET, and CSCO capture the durable spend on customized silicon and networking while hyperscalers (MSFT, GOOGL, META, ORCL) absorb the risk. The article underplays how quickly this dynamic could compress margins once training spend peaks and inference pricing falls. SpaceX's pre-IPO valuation at 100x revenue highlights the same hype risk but remains inaccessible to public investors.
Hyperscalers can embed models into proprietary data ecosystems and enterprise contracts, creating effective lock-in that pure model switching costs overlook, allowing them to monetize AI at scale despite commoditized inference.
"Starlink’s growth hinges on regulatory/spectrum approvals; a major setback could cap subscriber growth and undermine SpaceX's valuation beyond any revenue multiple."
Responding to Claude: The 'no moats in AI' line misses one big moat risk for SpaceX: regulatory and spectrum risk around Starlink. Even with 30% YoY growth, growth hinges on favorable licensing in dozens of jurisdictions, spectrum pricing, and orbital approvals; a setback in any major market could cap subscriber growth and cash flow. If Starlink stalls, the SpaceX thesis collapses more than a simple multiple debate, given Starlink’s cash use and capex needs.
"SpaceX's self-funding model is fragile and relies on perfect execution of launch cadence to support the capital-intensive Starlink rollout."
Claude, you’re conflating SpaceX’s launch profitability with Starlink’s capital intensity. While launch is a cash cow, Starlink requires constant, massive investment in satellite replenishment and ground infrastructure. If SpaceX’s launch cadence hits a regulatory or technical wall, that 'self-funded' model evaporates, forcing them to tap capital markets at a time when the 100x revenue multiple would be under extreme scrutiny. The moat isn't just spectrum; it’s the ability to sustain capex during a downturn.
"SpaceX's valuation assumes regulatory tailwinds persist; a single major jurisdiction blocking Starlink deployment cascades into both revenue and capex pressure simultaneously."
Gemini and ChatGPT both flag Starlink's capex intensity, but they're missing the asymmetry: launch revenue funds Starlink's growth *today*, but if regulatory headwinds hit (India, EU spectrum auctions, orbital debris rules), SpaceX loses both revenue streams simultaneously. The real risk isn't sequential—it's correlated collapse. Eisman's 100x multiple critique becomes prescient not because of valuation math, but because it leaves zero margin for policy error.
"Launch revenue is buffered from Starlink-specific regulatory risks by separate government frameworks."
Claude's correlated-collapse scenario overstates linkage: launch contracts with NASA, DoD, and commercial satellite operators fall under distinct FAA and ITAR regimes that spectrum auctions or orbital-debris rules for Starlink do not automatically trigger. Even if India or EU licensing stalls Starlink, reusable Falcon cadence can continue, preserving self-funding capacity and blunting the 100x multiple pressure until a broader policy shock hits both businesses.
The panel largely agreed that SpaceX's high valuation (100x revenue) leaves little margin for error, with the key risk being the potential collapse of Starlink's growth due to regulatory hurdles or market saturation. While some panelists acknowledged SpaceX's potential revenue streams and competitive advantages, the consensus was bearish, with concerns about valuation, lack of durable competitive advantages, and the high capital intensity of Starlink's growth.
The potential for SpaceX's launch business to fund Starlink's growth, given the high growth rate and improving unit economics of Starlink.
Regulatory and spectrum risk around Starlink, which could cap subscriber growth and cash flow, potentially collapsing the SpaceX thesis.