To The Moon

As 2026 unfolds, SpaceX, OpenAI, and Anthropic are poised for massive public debuts—potentially the biggest IPO wave in tech history. These companies promise transformative technologies: reusable rockets and Starlink broadband from SpaceX, and frontier AI models from OpenAI and Anthropic. But they face enormous capital demands for scaling compute clusters, energy infrastructure, data centers, and global operations. The central questions are how they will scale efficiently, monetize their products sustainably, and achieve genuine profitability—or whether they will need to dramatically raise subscription prices. History offers a warning: venture capital-fueled price suppression can mask underlying economics until the music stops.

Scaling Challenges: Capital Intensity on Steroids

All three companies operate in extraordinarily capital-hungry domains. SpaceX must fund thousands of Starship launches, constellation expansions for Starlink, and ambitious projects like Mars missions. OpenAI and Anthropic require ever-larger GPU clusters, specialized chips, massive power contracts, and cooling infrastructure. Training and inference for next-generation models already consume gigawatts of electricity—equivalent to powering small cities—and demand is exploding.Post-IPO, public markets will provide access to capital, but at a cost: quarterly scrutiny, earnings pressure, and less tolerance for multi-year losses. These firms will likely pursue a mix of strategies:

Vertical integration and efficiency gains: SpaceX already manufactures most components in-house and reuses rockets aggressively. AI labs are investing in custom silicon, optimized training techniques, and inference efficiency to reduce per-token costs.

Partnerships and cloud deals: OpenAI and Anthropic have major cloud and investment ties (Microsoft, Amazon/Google). They will expand these into broader revenue-sharing or dedicated capacity agreements.

Government and enterprise contracts: SpaceX benefits from NASA and DoD deals. AI companies are courting big enterprise licensing, sovereign AI projects, and regulated industry verticals (healthcare, finance, defense).

Scaling successfully will hinge on unit economics improving faster than capacity grows. If inference costs don't fall dramatically through algorithmic breakthroughs and hardware advances, the marginal cost of serving millions of users could remain punishingly high.Monetization Paths: Beyond

Monetization Paths: Beyond Subscriptions

Current revenue models lean heavily on subscriptions (ChatGPT Plus, Claude Pro, Starlink kits + monthly fees) and usage-based API pricing. These have grown rapidly but may not suffice alone at public-company scale.Tiered consumer and pro subscriptions: Expect aggressive upselling to higher tiers with better performance, priority access, and specialized features. Voice, video, and multimodal capabilities could justify premium pricing.

Enterprise and API dominance: This is where the real money likely lies—high-volume API access, fine-tuning, private deployments, and SLAs for businesses. Pricing here can be usage-based with volume discounts, but margins depend on cost control.

Hardware and ecosystem lock-in: SpaceX bundles Starlink with terminals and future mobility offerings. AI firms could expand into developer tools, agents, search, coding assistants, or even consumer devices.

New frontiers: Advertising (subsidized free tiers), data licensing, synthetic data sales, or AI-powered services in robotics, autonomous systems, and scientific discovery.

The risk is "feature creep" without corresponding willingness-to-pay. Users have shown price sensitivity; many stick to free tiers or rotate between models.

Profitability: The $100 Billion QuestionWill they ever be consistently profitable? Optimists point to network effects, technological moats, and deflationary cost curves in computing. Pessimists highlight the "picks and shovels" reality: even if AI creates trillions in downstream value, the infrastructure providers may capture only a fraction while bearing huge fixed costs.SpaceX's Starlink has clearer path to profitability as user numbers grow and launch costs decline—fixed satellite costs are amortized over more subscribers. AI companies face a tougher slog. Training runs get more expensive with each generation, and inference must be extraordinarily cheap to support broad adoption. Without major efficiency breakthroughs or energy cost reductions, sustaining losses on usage could persist.Subscription prices could indeed need to rise, but sky-rocketing them risks user churn and slower adoption. A more likely path is price discrimination: keeping entry-level access affordable while charging significantly more for power users, enterprises, and high-reliability applications. Hybrid models—VC-style subsidized consumer growth funded by enterprise margins—may continue, but public markets will demand a timeline to breakeven and positive free cash flow.

The Uber Cautionary Tale: When the Subsidies StopUber offers a stark parallel. For years, venture capital subsidized rider fares and driver payouts to achieve hyper-growth and market dominance. Prices were artificially low, masking poor unit economics in many markets. Drivers received bonuses and incentives; riders enjoyed cheap rides. The strategy worked for network effects and valuation, but when public markets demanded profitability, the reality hit: fares rose, driver pay was squeezed, and growth slowed. Uber eventually reached adjusted profitability through operational discipline, pricing power in certain segments, and diversification—but only after massive cumulative losses and years of skepticism.

The AI and space companies risk a similar trap. Massive funding rounds have allowed aggressive investment in R&D and capacity while keeping consumer-facing prices relatively accessible. Post-IPO, with public shareholders focused on margins and cash flow, the subsidies must end. Cost discipline becomes paramount: ruthless efficiency in data centers, supply chain control, talent allocation, and go-to-market. Companies that fail to improve gross margins or achieve positive contribution per user will face valuation compression, activist pressure, or forced pivots.SpaceX may have an advantage here due to its vertically integrated hardware model and proven execution under Elon Musk. Pure AI software/services players could struggle more if commoditization accelerates and open-source alternatives erode pricing power.OutlookThese companies can achieve profitability, but it will require excellence in execution rather than endless capital infusions. Scaling will succeed for those who drive down costs fastest while building defensible products customers happily pay more for over time. Subscriptions may rise selectively, but broad price hikes alone won't solve structural issues—better economics must come from technology and operations.The Uber lesson is clear: artificial deflation via subsidies buys time for dominance but doesn't create sustainable businesses. When funding discipline returns—as it always does in public markets—only the truly efficient and differentiated will thrive. For SpaceX, OpenAI, and Anthropic, the IPO bell may signal the end of the subsidized era and the true test of their economic models.