Musk abandoned his own ‘solar electric economy’ to burn gas for an AI chatbot no one uses

The article argues xAI is burning more than 6 million tons of greenhouse gases annually through 62 unpermitted methane gas turbines while Grok’s downloads have fallen 60% and monthly active users dropped to 12.2 million. It also says xAI is leasing Colossus compute to Anthropic for $1.25 billion per month, a deal worth over $40 billion through 2029, amid legal pressure from the EPA, NAACP, and Earthjustice. The broader implication is that Musk’s new space-solar thesis and the reported $1.75 trillion to $2 trillion SpaceX IPO valuation are being used to reframe the company’s energy narrative.

Analysis

The market implication is not “Musk hypocrisy” as a headline; it is capital allocation distortion inside the Musk ecosystem. SpaceX is trying to re-rate itself from a launch vendor into a full-stack energy/infrastructure platform, and that narrative needs an addressable market large enough to support a trillion-plus valuation. If investors buy the story, the indirect winners are the picks-and-shovels: high-end chip suppliers, launch/space component vendors, and grid-adjacent power equipment names that can sell into AI buildout without depending on Grok’s actual product quality.

The more important second-order effect is that the article weakens the credibility of xAI as a standalone demand engine. If Grok demand is soft and compute is being monetized externally, xAI’s capex becomes a liability rather than a growth asset; that matters because the market tends to finance AI infrastructure on utilization assumptions, not just installed capacity. If utilization disappoints, the overhang shifts from “can they build it?” to “can they fill it?”, which is where multiple compression starts.

For TSLA, the bearish read is not about current vehicle demand so much as governance and strategic muddle. Tesla’s optionality premium has always depended on the market believing Musk can convert moonshots into durable cash flows; this article argues the opposite, that adjacent ventures are siphoning credibility and capital while contributing nothing to Tesla’s core industrial moat. The risk is that a cleaner separation of assets or a hype cycle around SpaceX can temporarily mask this, so the short is best expressed with defined downside risk rather than a naked directional bet.

The contrarian point: the space-solar concept may be ridiculous on 2026 economics, but it is still a plausible narrative asset if launch costs fall and orbital compute has policy tailwinds. That means the idea can drive stock performance well before it becomes commercially real. The key timing variable is the IPO window; if SpaceX files and markets well, the story can extend for quarters even if the operational logic remains weak.

Key Decisions for Investors

• Short TSLA on rallies for a 1-3 month horizon; thesis is governance/strategy dilution and reputational drag, with downside limited by already-depressed sentiment but upside for the short if SpaceX/AI narrative cools.
• Buy NVDA or a NVDA/TSLA pair long-short into any AI-infrastructure pullback; if xAI utilization is overstated, the best risk-adjusted long is the upstream chip supplier still monetizing real capex across multiple customers.
• Initiate a tactical short on any listed space/launch proxy or related high-beta supplier into SpaceX IPO hype; use a 4-12 week window because valuation enthusiasm typically front-runs fundamentals by one to two quarters.
• Consider long utility/grid equipment names versus TSLA as a cleaner ‘energy transition’ expression; the article reinforces that terrestrial electrification still wins on economics, while Musk’s narrative shift is mostly capital-market theater.
• Avoid shorting NVDA solely on xAI softness; the article argues demand is being redistributed, not destroyed, so the better expression is a relative short on xAI-adjacent hype rather than a sector-wide AI bearish bet.