Jordi Visser — SpaceX IPO, Bitcoin, and AI Investment Thesis

Source: Should You Invest In SpaceX IPO, Elon Musk, Bitcoin or AI?, Anthony Pompliano (The Pomp Podcast), June 13, 2026.

The Framework: Narrative vs. Bottleneck — Bet the Inputs, Not the Dream

Visser's lens for the SpaceX IPO debate mirrors how he trades AI and Bitcoin: spreadsheet skeptics vs. Elon believers miss that the smoother path to alpha may be physical bottlenecks — silver, copper, indium, energy — rather than paying for a low-probability distribution of orbital dreams at a rich valuation. SpaceX is two businesses (today's revenue put + orbital/cursor call option), but Visser personally builds silver, copper, and energy while the market chases 52-week-high narratives.

Debate	Visser's Read	Where Capital Should Flow
SpaceX IPO rich vs. Elon magic	Valid on both sides; low probability SpaceX earns today's price	Commodity inputs if timeline slips 2×
AI model layer	Commoditization + subsidized tokens + IPO rush = near-term stock problem	Not a bubble in compute demand; bottleneck in seats
Bitcoin / crypto	Below 200-day MA; nobody wants it	Scarcity + proof-of-authenticity when AI slop explodes
AI midcycle	3–6 month slowdown in semis/cooling/infra	Hyperscalers hurt more than component names

"I'll take silver and building a position here in silver and in copper... I have copper stuff. I'm buying energy... Elon has proven a lot, but... it's a low probability event. I'll take silver."

Investment Thesis #1: SpaceX — Fourth Hyperscaler in 45 Days, Orbital Call Option at 5× Cheaper Infrastructure

Argument: SpaceX is two companies: (1) today — Starlink (Visser uses it as backup in Maine storms) plus Colossus making SpaceX the fourth-largest hyperscaler in ~30–45 days, with Anthropic and Google deals after a 122-day first build that normally takes years; (2) future — orbital data centers at ~$5B vs ~$25B terrestrial infrastructure per GW (~$60B all-in with ~$35B chips unchanged). Cursor accelerates xAI coding; Visser heard SpaceX secured ~20% of early Vera Rubin supply — Jensen routes scarce racks to proven builders.

Google's deal embeds a call option on orbital compute if Elon executes. Starship thousands of launches per year by 2028 (≥3/day up-and-back) plus humanoids within five years would re-rate the story. Elon's algorithm — find bottlenecks, subtract friction (sun for energy, no cooling in space) — and launch monopoly create a hardware moat AI software cannot replicate.

Contrarian element: Visser argues terrestrial data centers may be harder than orbital — land, power, plumbers, electricians, regulation, and cooling friction that space eliminates. Sell-side smooth GW charts ignore this.

Trigger: Earnings slightly above expectations; Starship same-day turnaround; orbital rack specs progressing; additional GW resale deals at premium monetization.

Names: SpaceX (PRIVATE — IPO), NVIDIA (NVDA), Alphabet (GOOG).

Investment Thesis #2: Silver and Copper — Orbital + Batteries Need More Metal Than the Spreadsheet Trade

Argument: Visser pitched Micron at $60–$110 in early 2025; it ran to $700–$1,000+ and is still cheap because exponential demand forecasts were wrong. The same error applies to SpaceX orbital narratives. Orbital stations and grid batteries both need far more silver than traditional data centers. An All-In materials interview: copper needed in the next 10–15 years equals all copper mined in the last 10,000 years. You cannot algorithm your way out of mine lead times — innovation speeds software, not ore bodies.

Visser believes silver could 10× from here over Elon's build timeline (assume 2× Elon's stated schedule for humanoids and orbital). Markets ignore silver because it is not working right now — same as Bitcoin below the 200-day moving average. He is actively building silver, copper, and energy positions instead of sizing SpaceX at peak narrative.

"Do you know what both of them need a lot more of than the traditional one? Silver. I want more silver."

Contrarian element: Investors flock to SpaceX IPO headlines and 52-week highs (Lilly last week); Visser buys unloved scarcity while dreams trade at premium multiples.

Trigger: Silver and copper price breaks; AI midcycle slowdown extending physical shortage; battery and orbital capex guides citing material constraints.

Names: Silver (SLV), Freeport-McMoRan (FCX), GE Vernova (GEV) and energy producers.

Investment Thesis #3: Indium Phosphide — China Controls the Hidden Data-Center Mineral

Argument: Indium phosphide is a critical input for Marvell, Coherent, and Lumen photonics — and China owns the supply, having throttled exports like rare earths. You cannot build data centers without it. Visser links this to geopolitics: China needs oil via Venezuela, Iran, Russia — the Strait of Hormuz opens when it must. The Coherent CEO flew to Trump during the Xi visit — input-cost diplomacy in real time.

US hoards energy; China hoards minerals. The path to GW-scale compute is not smooth — equal news of disappointment and surprise over the next 3–6 months as hoarding gives way to AI midcycle slowdown expectations after the agentic shock.

Contrarian element: Wall Street models smooth GW ramps; Visser focuses on single-mineral choke points invisible in spreadsheet DCFs.

Trigger: Indium export policy shifts; photonics lead-time extensions at COHR/LITE/MRVL; Hormuz / Venezuela oil flow headlines tied to mineral negotiations.

Names: Marvell (MRVL), Coherent (COHR), Lumentum (LITE).

Investment Thesis #4: AI Model Layer — Commoditized, Subsidized, and Racing to IPO Before the Music Stops

Argument: SemiAnalysis found OpenAI and Anthropic subscriptions subsidize tokens below production cost — great revenue, negative unit economics short term. 80–85% of queries could run on cheaper or Chinese open-source models; enterprises now mandate token limits after the "use AI" mandate flipped to "these bills are too high." DeepSeek released a competitive model and cut prices — Chinese AI model stocks sold off in a deflationary race.

Add Bernie Sanders and Trump wanting equity stakes, the Pope speaking against models, and Dario's latest AGI warning — plus token production costs up ~4× YoY while capacity stays capped (1,000 want in, 20 seats). That is why Google raised capital, Meta announced a tens-of-billions raise, SpaceX/Anthropic/OpenAI IPO simultaneously, and OpenAI faces price competition (WSJ). Visser: not an AI demand bubble, but a near-term model-stock problem.

Pricing will tier like airlines — cancer pharma pays premium inference; chat users get routed to Opus 4.7 at $5 vs Fable. Nvidia Nemotron open-source surprised Visser — nearly competitive with Chinese models; edge laptops/phones may host it. Ads and surge pricing follow. Not all model companies win.

Contrarian element: Revenue hockey sticks mask commoditization + capital rush — George Soros "when I see a bubble, I rush in" explains the IPO cluster.

Trigger: OpenAI price cuts; DeepSeek pricing moves; enterprise token-throttle disclosures; SemiAnalysis-style unit-economics audits on subscriptions.

Names: NVIDIA (NVDA) Nemotron/open-source edge play; Anthropic (PRIVATE) as perceived coding/AGI winner.

Investment Thesis #5: Anthropic Over OpenAI — Coding Path to AGI and Enterprise Focus

Argument: Visser guesses Sam Altman may not lead OpenAI within a year — personal incidents, political isolation ("nobody seems to really like him"), and a consumer-Apple strategy vs Anthropic's enterprise/political positioning. Andrej Karpathy at Anthropic; talent migration; Visser used Claude exclusively then rotated back to ChatGPT — but believes Anthropic wins the coding → AGI battle. Coding acceleration with Blackwell and Vera Rubin now arriving makes Dario's "critical point" credible.

OpenAI's uphill fight matters for IPO pricing and Mag 7 hyperscaler trade — Visser sees hyperscalers having more trouble than component names in the current 3–6 month midcycle.

Contrarian element: OpenAI still has product love and user base; Visser is explicit this is guesswork, not certainty — but directionally favors Anthropic for alpha on model-layer winners/losers.

Trigger: OpenAI leadership changes; Anthropic enterprise deal velocity vs OpenAI consumer metrics; coding benchmark share shifts post-Fable/Mythos.

Names: Anthropic (PRIVATE); Alphabet (GOOG), Amazon (AMZN) as Anthropic/cloud beneficiaries.

Investment Thesis #6: Bitcoin and Crypto — Proof-of-Authenticity in an AI Slop World

Argument: Bitcoin sits below the 200-day MA — retail will not buy until narrative returns (52-week-high behavior). Visser's Knicks Finals experience reframes crypto utility: in a world of AI slop, deepfakes, and infinite fake podcasts, experiential truth and ** provable presence** matter. Stablecoins, tokenization, and NFTs are the three stools of crypto network effects (Candy Digital advisor tie-in). Commemorative Game 4 tickets hit eBay at $300–450 overnight — but 24 million fakes can flood any physical collectible without blockchain provenance.

"You don't know unless it's on the blockchain... things will have to be put on the blockchain."

Crypto is not a trade for today’s tape; it is insurance on authenticity when humanoids walk streets in five years and nothing on a screen is trustworthy.

Trigger: BTC reclaim of 200-day MA; stablecoin volume growth; tokenization/RWA pilot acceleration; NFT/experiential proof-of-attendance products at scale.

Names: Bitcoin (BTC); tokenization infrastructure (category).

The Ecosystem Map

SpaceX view: Dual structure (Starlink + Colossus today; orbital $5B/GW call); Cursor for coding; ~20% Vera Rubin; Google orbital option
Personal commodity book: Silver (building, 10× thesis), copper, energy — preferred over SpaceX at current probability-weighted price
Pitched winners: Micron (validated), Marvell, Lilly (prior week), energy
Model layer: Anthropic favored over OpenAI; Nemotron surprise; DeepSeek price war
Physical AI: Bezos Prometheus ($12B at ~$40B) — Chimath 80/20 manufacturing; Lilly hiding phase-one data due to AI catch-up risk
Demand signals: Sylvia queries 6 → 15 per week; agentic demand still under-modeled vs human queries
Midcycle call: 3–6 months; semis/cooling/infra sideways; hyperscalers more vulnerable than bottlenecks

Key Risks

SpaceX valuation vs. outcomes: Low probability Elon delivers everything priced in; Visser explicitly chooses commodities over the dream at today's multiple
Model commoditization: DeepSeek pricing, open-source routers, and enterprise token limits compress frontier margins
China mineral weaponization: Indium phosphide and rare-earth-style export controls raise input costs and delay GW builds
IPO/supply glut: SpaceX, Anthropic, OpenAI, Meta, Google raises — capital recycling and price competition
Political/regulatory headwinds: Sanders/Trump equity demands, papal AI skepticism, slowing narratives
AI midcycle drawdown: 3–6 month pause after hoarding phase; semis/cooling names see equal disappointment and surprise
OpenAI leadership: Visser's Altman exit guess may be wrong; product momentum could persist
Bitcoin timing: Below 200-day MA can extend; narrative assets lag while AI earnings dominate attention
Copper/silver timing: Material rallies may lag if innovation partially substitutes (Elon's voltage/copper hacks elsewhere in ecosystem)

Investment Opportunities at a Glance

Tier	Name / Category	Core Thesis	Conviction Signal
1	Silver (SLV)	Orbital + batteries need more silver; Visser building position; 10× potential	"I'll take silver" over SpaceX at low-probability valuation
1	Micron (MU)	Memory for agentic/humanoid world; $60→$1,000 and still cheap pattern	Validated 2025 pitch; agentic memory demand "off the charts"
2	Freeport-McMoRan (FCX)	10-year copper demand = 10,000 years of prior mining	All-In materials interview; cannot mine faster via algorithms
2	GE Vernova (GEV)	Energy hoarding alongside mineral hoarding; grid bottleneck	"I'm buying energy" alongside silver/copper
2	Marvell (MRVL)	Indium phosphide supply chain; photonics/data-center build	Named with Coherent, Lumentum on critical mineral
2	Coherent (COHR)	Indium-dependent photonics; CEO Trump/Xi diplomacy signal	China controls indium phosphide exports
2	Lumentum (LITE)	Same indium/photonics bottleneck set	Data centers "can't build" without mineral
3	SpaceX (PRIVATE)	4th hyperscaler in 45 days; $5B/GW orbital option	Visser structurally bullish, personally prefers commodities at price
3	NVIDIA (NVDA)	Vera Rubin to proven builders; Nemotron open-source edge	20% early Vera Rubin to SpaceX; Nemotron "not that far below" Chinese models
3	Eli Lilly (LLY)	Hiding phase-one data as AI compresses pharma catch-up time	Bezos/Prometheus + physical AI bottleneck theme
3	Bitcoin (BTC)	Scarcity + blockchain proof in deepfake world	Below 200-day MA = crowd absent; activation on narrative return
4	Anthropic (PRIVATE)	Coding → AGI winner vs OpenAI consumer strategy	Karpathy hire; Visser's explicit preference over OpenAI

Monitoring Checklist

Bottom Line

SpaceX is real, rich, and dual-structure — fourth hyperscaler in weeks, orbital infra at 5× discount to land — but Visser would rather own silver, copper, and energy than bet the full Elon distribution at today's price.
The model-layer IPO rush is a warning, not a disproof of AI demand — subsidized tokens, DeepSeek deflation, and token limits mean model stocks face a rough 3–6 months even as compute bottlenecks persist.
Anthropic over OpenAI is Visser's model-layer call — enterprise/coding path to AGI vs consumer strategy; Altman leadership may not survive ~12 months in his view.
Silver is the orbital-and-battery tell — every SpaceX spreadsheet bull should map input metals; Visser sees 10× silver while the crowd buys the IPO headline.
Bitcoin waits for narrative; blockchain wins on authenticity — below the 200-day MA today, but AI slop + deepfakes make on-chain proof of presence non-optional within five years.