Weekly Wrap: SpaceX IPO vs spectrum reality

Firstly, it’s important to understand how SpaceX wants to go public. SpaceX is offering less than 5% of shares to the public, well below the typical 15–20% for a major IPO. It’s a decision that creates artificial scarcity: limited supply creates price pressure before a single trade is made.

SpaceX’s ambition to beam AI computing from orbit at scale would require downlink capacity far beyond that provided by current satellite technology

Of those scarce shares, 30% are reserved for ordinary retail investors rather than the usual 10%. Financial analyst George Pearkes, featured in a recent documentary on the IPO, read this as a signal, not a gesture of generosity. He argued that when professional investors won’t buy at a given price, you sell to retail investors instead.

Another important element of this IPO is a new Nasdaq fast-track rule. Nasdaq changed its rules this spring to allow very large companies into the Nasdaq-100 index within just 15 trading days of going public, bypassing the standard “seasoning” period that exists to let markets rationally price new stocks. According to Reuters, this was a condition laid down by Musk for SpaceX’s listing on the exchange. Nasdaq, which earns listing fees and data revenues from big companies on its list, obliged.

The rule change means that big Index funds tracking the Nasdaq-100 will be forced to buy an estimated $7 billion of SpaceX stock in a single day. Not because they believe in it, but because the rules require it.

Early investors will likely cash out during this price spike, while the buyers – retirement savers, through their index funds – have no say in the matter.

The Nasdaq rule change did not happen in a vacuum, and it will not end with SpaceX. According to Former SEC official Ben Schiffrin, $30 trillion in assets is directed by index providers operating under a process that is, in his words, “effectively self-regulated”. No regulator approved Nasdaq’s rule change. The SEC has no jurisdiction over how an index chooses to include stocks. Nasdaq ran a comment period, and ordinary investors objected, but the rule changed anyway.

Bloomberg reported that the S&P 500 is considering its own version of the same change: shortening the seasoning period and potentially allowing large companies to enter without ever having shown a profit. The reason is not mysterious. OpenAI and Anthropic, both expected to go public, are neither profitable nor proven at the scale their valuations imply.

All of this is legal. Whether it is ethical is another question entirely.

In terms of promises to public investors, the SpaceX S-1 SEC filing tells a revealing story. SpaceX claims a total addressable market of $28.5 trillion, exceeding US GDP, of which $26.5 trillion is attributed to AI.

However, the S-1 also shows that the company lost nearly $5 billion in 2025. Its AI division lost $6 billion in operations. The one genuinely profitable business is Starlink, which generated over $11 billion in revenue last year. Everything else in the filing, the space-based AI data centres, global mobile coverage, and data centres in orbit, is based on long-term vision, not current revenues.

This vision collides with a wall that the S-1 barely acknowledges: spectrum.

Every byte of data from every AI satellite has to travel back to Earth via radio waves. Radio spectrum is finite, internationally regulated under ITUThe International Telecommunication U… treaty law, and already becoming congested as Amazon, OneWeb, Telesat and Chinese state-backed constellations compete for the same orbital slots and frequency bands.

SpaceX’s ambition to beam AI computing from orbit at scale would require downlink capacity far beyond what current satellite technology provides – a physics and engineering challenge the S-1 brushes over.

“For decades, the primary constraint of space operations has not been the ability to collect data, but the capacity to return it to Earth,” says an article in SatNews. “As modern sensors generate petabytes of high-resolution imagery and signals intelligence, they have collided with the ‘downlink bottleneck’, a physical limit imposed by oversubscribed radio frequency (RF) bands and limited ground stations.”

For global mobile coverage, on the other hand, the challenge is regulatory rather than physical: every country controls its own spectrum. Becoming a genuine global mobile operator means negotiating with every national regulator on Earth, in bands that overlap with existing terrestrial networks. That is a decade of diplomacy, and not at all comparable to a product launch.

Just a week ago, the European Commission proposed to limit Starlink and other non-EU operators’ access to MSSMobile-satellite service (MSS) describes a… (mobile satellite service) spectrum, a decision that could result in Starlink missing out on offering services to a significant market.

The S-1 filing does mention spectrum scarcity, but it does not explain how SpaceX will overcome that scarcity.

Law professor Ann Lipton put it plainly in The Verge: shareholders have three rights, to vote, to sell, and to sue. With Musk holding 80–85% of voting power, an arbitration clause curtailing litigation, and index fund inclusion making shares hard to offload, all three are diminished.

“There’s no getting off the train,” she said.

The least you can do is know what train you’re on.

Here’s what else PolicyTracker covered this week:

• WiFi industry representatives have criticised the UK and EU strategies for the upper 6 GHz band
• Iridium takes spectrum fight to US regulator as Europe moves to restrict its L-band downlinks
• European spectrum regulators meeting this week will decide whether a draft report on satellite direct-to-device will move to public consultation
• The US NTIA has announced a renewed push to identify potential 6G spectrum
• The EBU is lobbying to protect the car radio in the DNA