Newsletter · · Ashutosh Agarwal

SpaceX and AST Square Off in a 6 Billion Dollar Low Band Spectrum Fight - The Satellite & Space-Comms Race - Week of August 28, 2026

The Satellite & Space-Comms Race for the week of August 22 to 28, 2026. Podcast synthesis on SpaceX and AST SpaceMobile circling Grain Management's nationwide 800 MHz low-band spectrum at a reported 6 billion dollars, a hard calendar running to a November 5 selection and a December 4 FCC filing, SpaceX stock slipping below its 135 dollar IPO price even as Golden Dome work piles up, and Rocket Lab's fourteenth launch of the year.

The Satellite & Space-Comms Race

Week of August 28, 2026: SpaceX and AST Square Off in a 6 Billion Dollar Low Band Spectrum Fight

Podcasts only, trailing 7 days. Every figure and quote below is pulled from a named episode published in the last week.

The one-line version

The fight everyone has been waiting for finally got a price tag. A leaked Bloomberg report says both SpaceX and AST SpaceMobile (ASTS) are circling the same slice of nationwide low-band spectrum, the kind that makes your phone work in a subway or a canyon, and the number floating around is $6 billion. Elon Musk says the report isn't true. The AST community says that's exactly what you'd expect him to say. Either way, the timeline is now real: bidders show their hand in the first week of September, a winner gets picked by November 5, and a plan lands at the FCC by December 4.

TL;DR (15 seconds)

  • The $6B spectrum war is now on a clock. A leaked report has SpaceX and AST both eyeing Grain Management's 800 MHz nationwide low-band spectrum. First offers due first week of September; partner chosen by Nov 5; FCC filing by Dec 4. Musk publicly denied SpaceX is buying, and the AST side reads that skeptically.
  • SpaceX the stock is having a rough month. It slid a fourth straight day to about $132, back below its $135 June IPO price, even as it locks up $8B+ of Golden Dome defense work (roughly a third of the program so far) and pushes a story that it's really an AI and "orbital compute" company.
  • Rocket Lab keeps flying. RKLB launched its 14th mission of 2026 (a radar satellite for Japan's iQPS), though no show this week touched Neutron's schedule or the Iridium deal.

What's new, ranked by what actually matters to a book

1. The $6 billion spectrum war: SpaceX vs. AST, now with a countdown

This is the week's whole ballgame, and the single best breakdown came from the AST SpaceMobile Podcast episode "The $6 Billion Spectrum War With SpaceX" (Aug 21). To be clear on who's talking: this is the host of a dedicated AST fan and investor podcast, an investor voice rather than a company insider, but he walks through the deal mechanics like a banker, so it's worth the detail.

Here's the setup in plain English:

  • The prize. A private-equity firm, Grain Management, controls a chunk of 800 MHz low-band spectrum, about 7 MHz by 7 MHz nationwide. Low-band is the good stuff: it "goes through walls, in your car," and gives phones coverage deep indoors and in dead zones. Grain got it by swapping its 600 MHz licenses (plus roughly $2 billion in cash) to T-Mobile.
  • Why satellites suddenly matter here. When the FCC approved that swap back at the end of June, it did something new: it said a spectrum owner can meet its legal "build-out" requirement by beaming the signal down from satellites, instead of bolting radios onto thousands of expensive cell towers. The host called this the "game changer." AST is already out there testing on this exact spectrum.

It's important to note that AST is the only satellite provider that utilizes low-band spectrum cellular anywhere between 600 megahertz to 900 megahertz. There's no other satellite providers that do that.

That line is the crux of the bull case. SpaceX's direct-to-cell satellites work in mid-band (around 1.9 GHz); Globalstar sits in L- and S-band; Iridium is L-band. On paper, AST is the only constellation flying today that can natively light up this particular low-band spectrum.

  • The leak. A Bloomberg report claimed both SpaceX and AST are interested, and that Grain "hopes to reap about $6 billion." Musk fired back publicly that the report was "not true," described on the podcast as him calling it "an absolute lie." The host's read: this smells like Grain's bankers leaking to create a bidding war, a classic move to make AST "put their best foot forward." His caution is that Musk's denial might only cover part of the story: SpaceX could deny buying outright while still wanting to lease the spectrum, or while using the whole episode to scare the carriers into an MVNO deal.
  • The clock (this is the actionable part). Grain had to name a financial advisor by Aug 7 (done). First-round offers are due the first week of September. Grain must pick its satellite and direct-to-device partner and notify the FCC by November 5. The winning parties then have to file a fully baked satellite plan with the FCC by December 4.
  • The build-out bar. To keep the spectrum, the winner has to cover roughly 70% of 90% of the population within three years, which the host argues is "very doable" by satellite but a serious lift for SpaceX, since it would have to design and launch a fourth satellite platform (a low-band bird it doesn't yet have) inside that window.

Why it moves numbers: For AST, this is the difference between "promising constellation" and "owns or controls a nationwide low-band asset the three big carriers need." The host expects a partnership rather than an outright purchase, likely some cash up front plus a long-term lease and revenue share (he compared it to AST's existing Ligado L-band arrangement), possibly funneled through the AT&T, Verizon and T-Mobile joint venture. He also flagged that AST recently hired a UBS vice chair as a strategic advisor and raised a convertible bond partly to have the cash (a $500 million to $1 billion upfront check) ready for exactly this. For SpaceX, buying it would be a declaration of war on the carriers, since it would "foreclose" any chance of the MVNO deal it's been angling for.

A useful bit of history from the same episode: this playbook has run before. Globalstar's Jay Monroe floated a $10 billion number in early Bloomberg articles to bid up his spectrum, and it worked, eventually getting Amazon to pay up. File the $6B under "opening number," not "clearing price."

2. SpaceX the stock cracks below its IPO price, even as the defense wins pile up

While the private-market bulls celebrate, the newly public SpaceX had an ugly week. On Squawk on the Street (Aug 21), the CNBC desk noted the stock at about $132, "on pace for the fourth straight day of losses… the worst week of its month," and back below the $135 IPO price.

The irony is that the fundamentals underneath it keep getting better on the government side. Same episode:

  • A Wall Street Journal front-pager argued SpaceX "has become the go-to partner for U.S. military and intelligence offices," with the administration targeting rates "around 1,000 launches a year by 2030."
  • SpaceX won a $4.2 billion contract this year to develop hundreds of tracking satellites.
  • On Golden Dome (the U.S. missile-defense build-out), SpaceX "is in line for more than $8 billion… which is about a third of the total that has been budgeted for the project so far." President Gwynne Shotwell was described as "very bullish" on the government wins on the earnings call.
  • The tell for where the market's head is at: one host said the real total addressable market "continues to be vastly around enterprise AI… orbital compute," adding, "We do consider them now a hyperscaler, or at least close to one."

Why it moves numbers: These are all pundit framings (CNBC anchors, plus a secondhand Shotwell reference), not fresh operator disclosure, but for the private-market names in this universe, SpaceX's public tape is the sentiment anchor. A SpaceX that trades below its IPO price cools the "everything space is a rocket ship" mood that has been lifting ASTS and RKLB multiples.

3. SpaceX's AI pivot gets louder, and weirder

Aviation Week's Check 6 ("Not So Dog Days Of Summer," Aug 21) had the most substantive read on SpaceX's earnings and Musk's post-earnings comments to staff. The aerospace journalists laid out the split personality of the business:

  • Starlink makes money. The launch business doesn't, but only "because of the investment in Starship. If you strip the Starship investment out, they would actually be making money too."
  • AI is the cash drag. Capex was "$18 billion in the last quarter" (the panelist hedged "I might be off by a billion or two").
  • The pronouncements, delivered with the usual Musk swagger: $1 trillion in turnover by 2030 ("maybe a year earlier"), and Starship "flying… as soon as next year every day." The panel's verdict: take with "more than a grain of salt."
  • The genuinely surprising part came from what Musk told employees a few days after the call: that in September, AI revenue would overtake all the other revenue of the company; that in Q4 "it would exceed it by a lot"; and that within four to five years, over 90% of the value in SpaceX would be from AI.

Here's all these investors who are invested in SpaceX as the space company. And within basically less than two months of them buying their shares, they're told, well, actually, we're probably going to be an AI company.

Why it moves numbers: If you own SpaceX exposure (directly or as a read-through), the thing you bought is quietly being re-labeled. That's a risk and an opportunity: the connectivity and launch businesses may get valued as a smaller slice of an AI story, which cuts both ways for how the market prices the "space" comps around it.

4. Rocket Lab keeps its launch pace red-hot

Astronomy Daily (Aug 21) reported that Rocket Lab launched overnight from New Zealand: an Electron rocket (just 59 feet tall) carrying SUSANO-2, a synthetic-aperture-radar satellite for Tokyo-based iQPS, the ninth iQPS satellite RKLB has launched toward a planned 36-satellite constellation. It went to low Earth orbit about 357 miles up, roughly 50 minutes after liftoff. Crucially for the cadence story: this was RKLB's 14th flight of 2026 and the 93rd Electron overall, "closing in on their own record with plenty of year left."

Why it moves numbers: Steady Electron cadence is the boring, bankable part of the RKLB story, since it keeps the launch backlog converting to revenue. But note the gap: no podcast this week offered operator or analyst commentary on Neutron's first flight, the launch-market economics, or the pending Iridium acquisition. The launch is operational color, not a re-rating catalyst.

The debate: how big is the direct-to-device market, really?

"Direct-to-device" means your ordinary, unmodified phone connecting straight to a satellite, no special hardware and no dish. It's the whole reason ASTS and Starlink's direct-to-cell effort exist. This week's episodes sharpened both sides.

The bull case. Every phone on Earth in a dead zone is a potential customer, and there are billions of them. The carriers can't economically build towers across oceans, deserts, and mountains, so they'll happily pay a satellite partner to erase the "no service" bars. It's a feature they can upsell and a churn-reducer they can't replicate. And the asset is scarce: as the AST host stressed, AST is the only constellation flying today that can use this prime low-band spectrum, the frequency that actually penetrates buildings and cars. Control that, plug it into the three big U.S. carriers, and you have "very steady cash flows over many years," the kind of business the host argued could be worth "10 billion, 12, 13" versus a one-time $6B spectrum sale.

The bear case. Even the AST host, a bull, admitted the physical limit out loud:

7 MHz by 7 MHz is great for coverage, but it's not enough to compete against the carriers.

In other words, direct-to-device from space is a supplemental layer, great for filling gaps and not for replacing your terrestrial data plan. That caps how much anyone will pay for it. Layer on the competitive reality: SpaceX can outbid essentially anyone ("if SpaceX wants to buy it, they can buy it"), and if it commits to low-band, the whole market becomes a capex arms race. And notice where the smart money thinks the real money is: on Squawk, the desk waved away connectivity to say SpaceX's TAM is "vastly around enterprise AI… orbital compute." When even the satellite-connectivity champion's biggest rival is telling investors the value is in AI, that's an implicit vote that the pure connectivity market, while real, is smaller than the hype. The swing factor: whether direct-to-device stays a low-ARPU add-on the carriers control, or becomes a genuine standalone service, which is exactly what the Grain spectrum fight will help decide.

Stocks in play

AST SpaceMobile (ASTS), the name of the week

  • Bull: Only flying constellation that can natively use Grain's nationwide low-band spectrum; already testing on it; incumbent partner with AT&T, Verizon and T-Mobile in the loop; raised a convert and hired a UBS vice chair to have cash and advice ready for a deal. A lease and revenue-share structure could build a business "worth 10 to 13 billion" rather than a one-time spectrum bill.
  • Bear: Could be forced to overpay in a leaked bidding war; 7 by 7 MHz is a coverage layer, not a carrier-killer; SpaceX can outbid it if the fight turns into "a pissing match" (the host's words); the deal could mean fresh dilution.
  • Next catalyst: First-round Grain offers, first week of September; partner selection by November 5; FCC filing by December 4.

SpaceX and Starlink (private)

  • Bull: The government pipeline, with $8B+ Golden Dome, a $4.2B tracking-satellite award, and "go-to partner" status for the Pentagon; Starlink profitable; a credible bid to be the world's launch monopoly (roughly 1,000 launches a year by 2030 as the goal).
  • Bear: Public stock below its $135 IPO price at about $132, four straight down days; $18B per quarter capex is bleeding cash; the AI pivot is re-labeling a "space" investment into something investors didn't sign up for.
  • Next catalyst: Musk's claim that AI revenue overtakes all other revenue in September, a check on whether the pronouncement is real.

Rocket Lab (RKLB)

  • Bull: Relentless cadence, with 14 launches in 2026 and the 93rd Electron overall, plus a deep, diversified customer base (the iQPS constellation build-out).
  • Bear: The catalysts that matter (Neutron first flight, launch-market economics, the Iridium deal) got no podcast commentary this week, so there is no fresh confirmation of the timeline.
  • Next catalyst: A confirmed Neutron first-flight date.

Read-throughs

Carrier partners (VZ, T, TMUS). The carriers are central to the Grain story. Per the AST host: AT&T likely can't buy the spectrum itself (it would trip the FCC's spectrum-screen limits), Verizon could in theory but has "the same issues," and T-Mobile just divested this very spectrum. That leaves a cooperative path, the three-carrier JV as the "vessel" that holds the spectrum with AST contributing the satellites, as the most likely structure. The carriers' incentive, bluntly stated, is to keep low-band out of SpaceX's hands.

Incumbents. Globalstar (GSAT), Iridium (IRDM), and EchoStar/Hughes (SATS) drew no dedicated coverage. The only mentions were as reference points inside the AST episode: Globalstar's Jay Monroe as the template for floating a big spectrum number to attract a buyer (Amazon), and Iridium as an example of a past sale process where a disciplined bidder walked rather than overpay.

Launch and component suppliers. The one supplier data point: Catalyst Space Technologies, which built NASA's $30 million "Link" satellite-rescue spacecraft (launched July 3 on a Northrop Grumman Pegasus XL). The mission failed on August 19 when Link began spinning uncontrollably, so NASA's SWIFT telescope will now make an uncontrolled re-entry before year-end. A reminder that in-orbit servicing is still an unproven business.

SpaceX private-market valuation as sentiment anchor. The most-repeated theme across the pundit shows this week was SpaceX-as-benchmark. Anthropic is reportedly prepping an IPO to "match or beat SpaceX's record-setting $75 billion public debut" ($86.2B including over-allotments), floating a roughly $2 trillion valuation, per the Elon Musk Podcast (Aug 21) and The Rundown (Aug 21). The useful contrast for space investors: SpaceX owns its physical infrastructure (rockets, launch pads, manufacturing), whereas an AI lab rents its compute, a distinction the Elon Musk Podcast used to argue SpaceX's valuation rests on harder assets. Meanwhile, on Watchdog on Wall Street (Aug 22), RIA Chris Markowski took a victory lap for refusing to buy at $220 to $225 a share, needling a bullish Jim Cramer call from the IPO: "38 days after he said that, the stock was down to $120 to $125." His bottom line: he loves the company long-term, just not the entry price.

What changed vs. last week

Last week's issue led with AST clearing the FCC for test authority on Grain's 800 MHz spectrum, and flagged that Grain must name a direct-to-device operator by around November. This week that story escalated hard: it's no longer a quiet test-and-wait, it's a named, leaked, $6-billion contest with SpaceX, a public Musk denial, and a concrete auction calendar (first offers early September, selection November 5, FCC filing December 4). The "Grain must name an operator by around November" note from last week is now pinned to an exact date: November 5.

  • AST advisory bench: Last week we noted AST hired Ozzy Ramos (ex-Lehman, Barclays and UBS vice chair) as an M&A signal. This week the AST host tied that same UBS-vice-chair hire directly to the Grain negotiation timing, so the M&A signal now has an obvious target.
  • SpaceX stock: Last week the story was a roughly 40% post-lockup rally. This week it reversed, with four straight down days, back below the $135 IPO price at about $132. The sentiment anchor flipped from tailwind to headwind.
  • SpaceX AI pivot: Reinforced and made more concrete. Last week it was Musk claiming AI would be 99% of value in about 5 years; this week the framing tightened to "AI revenue overtakes all other revenue in September" and "more than 90% of value in 4 to 5 years," with a repeated $1T-by-2030 turnover goal.
  • Golden Dome: Last week it was a general procurement magnet; this week it got a SpaceX-specific number, $8B+, roughly a third of the budgeted total.