Newsletter · · Ashutosh Agarwal

Japan Files a 136 Satellite Constellation as Europe's Big Four Circle AST - The Satellite & Space-Comms Race - Week of September 11, 2026

The Satellite & Space-Comms Race for the week of September 4 to September 11, 2026: Japan filed a 136-satellite constellation at the ITU for the Rakuten and AST venture, Europe's four largest carriers are reportedly forming a joint venture to counter Starlink, an NTIA official mapped a 2028 target for the first US direct-to-device spectrum auction, and AST jumped 12% on a board member's open-market buy even as satellite BW11 has yet to unfurl.

The Satellite & Space-Comms Race

Week of September 11, 2026: Japan Files a 136 Satellite Constellation as Europe's Big Four Circle AST

If you had to sum up the week in one word, the podcasts hand it to you: sovereignty. Not launch cadence, not subscriber counts, sovereignty. Over a single long weekend, Japan quietly filed for a 136-satellite constellation tied to AST SpaceMobile, Bloomberg reported that Europe's four biggest phone companies are teaming up to build their own satellite service, and the UK put £7.8 billion (GBP) behind a home-grown space plan. The common thread: countries increasingly want a satellite phone network they own and control, and that is turning into a very specific bet against the way Starlink is built.

Meanwhile the stocks themselves had a rough-then-recovering week. AST bled through August with the rest of the high-flying "story" names, then popped 12% when a long-time board member stepped in and bought shares. And SpaceX, now a public company since June, kept living under the cloud of what Wall Street has started calling the "SpaceX slump."

One caveat to carry through the AST sections below: almost all of this week's AST detail comes from two openly bullish independent investors who host the AST SpaceMobile Podcast. They do real homework, but they are cheerleaders, not neutral analysts and not company insiders, and their inferences are flagged as such below.

TL;DR (15-second version)

  • AST SpaceMobile (ASTS) owned the week. Japan's telecom regulator filed a 136-satellite constellation at the global spectrum body for the Rakuten and AST venture; Europe's top four carriers are reportedly forming a joint venture to counter Starlink, and all four already work with AST. The unifying story is national "sovereignty," and it favors AST's design over Starlink's.
  • A US government spectrum official laid out the real D2D roadmap: regulators are studying new airwaves for satellite-to-phone service, aiming to auction a first slice by 2028, with Starlink, Amazon, and AST all named as interested. This is the pipe everything eventually flows through.
  • Quiet week on Globalstar, Rocket Lab, and Iridium, zero dedicated podcast coverage. SpaceX is down about 30% from its June peak, but its post-lockup share dump got absorbed with a shrug.

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

1. Japan files a 136-satellite constellation for the Rakuten and AST venture

The single most concrete development of the week. On September 4, Japan's Ministry of Internal Communications (the rough equivalent of the US FCC) filed with the ITU, the United Nations body that coordinates satellite airwaves worldwide, for a 136-satellite constellation nicknamed "J-Bluebird," on behalf of what is effectively the Rakuten and AST SpaceMobile joint venture (the "JLEO" project). This was surfaced and explained on the AST SpaceMobile Podcast (9/8), whose host is a former decade-long professional investor.

Why it moves numbers: the Japanese government has committed about $1 billion (USD) to the project, with the partners expected to put up a matching $1 billion of their own. The host's key point on the business model, and it's a sharp one, is that a government anchor check de-risks private debt raising, so "it requires very little, if any, equity investment from the partners" and the $1 billion gets multiplied. He also stressed the constellation is "much bigger than Japan": satellites will serve Japan on a sovereign, Japan-controlled basis, but when they fly over other regions they fold into AST's global "federated" network and earn AST additional revenue.

"So when we say they're big in Japan, no, this is even bigger than Japan."

The revenue-split nuance worth writing down: for these sovereign deals, AST's percentage share of the economics may fall (from roughly 50/50 toward, say, 30/70), but in exchange it locks up 100% of a national market rather than a slice. The host's framing: "you would be more than happy to get a hundred percent of a market with 30% of the economics versus a third of the market with 50%." Note the JLEO joint venture itself is not yet formally signed, the ITU filing is a below-the-radar regulatory step, not a press-released deal.

2. Europe's big four reportedly form a JV to counter Starlink, and they all work with AST

Per a Bloomberg report discussed on the same AST SpaceMobile Podcast (9/8), Europe's four largest mobile carriers, Vodafone, Deutsche Telekom, Orange, and Telefónica, are in talks to form a joint venture to chase Europe's 2 GHz "S-band" satellite spectrum, explicitly to counter Starlink. The tell: all four already work with AST (Vodafone, Orange, and Telefónica are established partners; Deutsche Telekom was disclosed as a testing and integration partner in early August, the step that, in the host's words, "hint, hint" precedes a definitive agreement).

Why it moves numbers: the spectrum in play is 30x30 MHz of 2 GHz S-band currently licensed to EchoStar (SATS) and Viasat, whose licenses expire in May 2027 and then get reallocated. The expected structure is three 10x10 MHz blocks: one for Europe's Iris² consortium, and one reserved for a majority-European-owned player, which the host argues points to the Vodafone and AST "Satellite Connect Europe" venture. His non-consensus call: he thinks there's a real chance AST's venture and a broader European venture each land an independent 10x10 block, "which I don't think the market's pricing in." Treat that as the host's speculation, not confirmed.

This is the same movie playing in the US (AT&T, Verizon and T-Mobile all working with AST rather than Starlink) and, per the host's due diligence, possibly in Brazil (a carrier there received a 10x10 MHz S-band allocation; AST already works with the top Brazilian carriers).

3. A US spectrum official maps the direct-to-device (D2D) auction timeline

Genuinely useful signal from outside the fan bubble. At a Hudson Institute event (Hudson Institute Events Podcast, 9/9), Tricia Paoletta, Principal Deputy Assistant Secretary at NTIA, the arm of the Commerce Department that manages federal airwaves, walked through the government's spectrum agenda.

The relevant parts for this universe:

  • Congress last year ordered NTIA and the FCC to free up 800 MHz of spectrum for high-power 5G and 6G use. NTIA must identify 500 MHz of federal spectrum within five years (by July 2030), including at least 200 MHz within two years (by July 2027).
  • Crucially, NTIA added the 1675 to 1695 MHz band to study specifically for direct-to-device satellite-to-phone service. It has already flagged the first 5 MHz to the FCC and is "very interested in moving and trying to get an auction in that band by the end of the president's term, so by 2028."
  • She named the field directly: "Starlink, owned by SpaceX, is already operating direct-to-device in some of these nearby bands. We've got plans from Amazon's satellite company, Amazon Leo, to do the same. We have another provider, AST SpaceMobile … that wants to do that."

Why it moves numbers: this is the pipe. A confirmed 2028-target US auction for dedicated D2D spectrum is the regulatory backbone under every bull case here, and a reminder that Amazon Leo is a third serious entrant, not just an AST versus Starlink fight. Watch the World Radio Conference (Shanghai, mid-October to mid-November) where the ITU takes up related bands.

4. A board member's open-market buy marks a line in the sand, stock up 12%

After a brutal August in which AST fell with the whole basket of high-momentum "story" stocks, board member Adriana Cisneros, a long-time director, early (Series B) investor, and personal ally of founder Abel Avellan, bought a large chunk of stock on the open market (not a pre-scheduled plan). Both hosts on the AST SpaceMobile Podcast (9/8) read it as a deliberate confidence signal that has historically "happened to mark a bottom." The stock jumped 12% on heavy volume the day it landed.

Why it matters: insider buying by someone that close to the company, during a drawdown, is a real sentiment tell, though, as the host fairly noted, insiders can buy precisely because nothing so material it's off-limits is imminent. A signal, not a fact about the fundamentals.

5. Bluebird deployment hiccup, and no September launch

The honest fly in the ointment. On the hyperscaler episode (9/8), the host confirmed that satellite BW11 is in its correct orbit but has not yet unfurled or deployed. His best guess (an inference, from a conversation with a contact) is heightened solar activity, you don't unfurl a large antenna during a solar storm, with optimal spacing and a possible software fix as alternatives, and outright mechanical failure viewed as unlikely. AST filed an FCC letter confirming the satellite is otherwise nominal.

Separately, the other host said flatly there will be no September launch; he expects an October launch, possibly two, with satellites BW14 to BW16 in work (two reportedly finished, one with an issue). Bottom line for a book: deployment and launch timing keep slipping, and that is the near-term risk the bulls are asking you to look past.

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

This is the question for the whole group. Here's the steel-man on both sides, using what was actually said.

The bull case (the AST hosts, this week):

  • Sovereignty is a structural tailwind. Countries were spooked by watching Elon Musk switch Starlink on and off in the Russia and Ukraine war, and even by the Trump and Musk feud over launch. They now want networks they control. AST's "bent-pipe" design, where traffic originates and lands inside the country and the local carrier keeps the customer, fits that. Starlink's design routes traffic through its own global network and "owns the customer," which the hosts argue is "diametrically opposed to what the MNOs and countries want."
  • The addressable market isn't static, it can 10x. One host's line worth keeping: "the TAM can be just as volatile as the company itself." Beyond phones, they point to IoT, drones, robots, autonomous vehicles, even connected appliances, potentially outnumbering human subscribers several-to-one, all needing 100% coverage with no dead zones.
  • The "hyperscaler" framing. In one launch AST put up three satellites with double the total surface area of the entire old Iridium constellation, which cost roughly $3 billion (USD) 25 years ago; the hosts claim AST has effectively launched "the equivalent of four Iridium constellations" in a couple of months. Their analogy: buying AST now is like buying Amazon, Google or Meta 15 years ago.
  • Product edge over Starlink. They argue Starlink's roughly 650 direct-to-cell satellites are a "minimally viable product", line-of-sight, text-only, no indoor coverage, while AST is going after true broadband that works indoors.

The bear case (steel-manned):

  • It's still pre-revenue with slipping timelines. BW11 hasn't unfurled, there's no September launch, and prior launch attempts have slipped repeatedly. The bull case is entirely about future execution.
  • A lot of this week's "news" is inference, not signed deals. The Japan JV isn't formalized. The European JV is a Bloomberg report, not a contract. The Anduril tie-up (below) is speculation. The "each gets a 10x10 block" call is explicitly the host's own.
  • Starlink's scale and vertical integration are real. Roughly 650 D2C satellites already serving T-Mobile, thousands of Starlink birds overall, a near-monopoly on launch, and a balance sheet the pre-revenue names can only dream of.
  • The sovereignty story is unproven economically. Selling satellites to governments at a lower revenue share is a different, lower-margin business than the bulls' 50/50 carrier model, and no one has put hard numbers on it yet.

The most rigorous dedicated D2D bear last week was MoffettNathanson's Julie Zhu (physics limits, wholesale-only economics). No equivalent dedicated bear episode aired this week, so the bear column above is assembled from execution risk and the hosts' own acknowledged caveats rather than a fresh skeptic.

Stocks in play

AST SpaceMobile (ASTS), market cap cited at roughly $25 billion (USD).

  • Bull: Sovereignty tailwind across Japan, Europe, UK, Brazil; 60+ carrier partnerships; only player targeting true broadband to an unmodified phone; insider buy plus "hyperscaler" launch cadence.
  • Bear: Pre-revenue; BW11 not deployed; no September launch; much of the week's catalysts are unconfirmed or inferred; sovereign-sales economics unproven.
  • Next catalyst: October launch (and whether BW11 finally unfurls); any formal T-Mobile, Deutsche Telekom, or JLEO agreement; European S-band reallocation (licenses expire May 2027).
  • On valuation, via the hosts relaying Berenberg's sell-side note (secondhand): base target $92; adding T-Mobile lifts 2030 EBITDA by roughly $1.1 billion and the target to $121 at a 15x EBITDA multiple (worth roughly plus $29 a share). One host argued the stock "more likely than not will trade at 30 or 40 times EBITDA" given growth, his opinion, not Berenberg's.

SpaceX and Starlink (now public since June)

  • Bull: Largest ISP in the world via Starlink, near-monopoly on launch, roughly $2 trillion (USD) market cap and the sixth-largest company in the world.
  • Bear: Down roughly 30% from its roughly $200 a share June peak to the $140 to $150 range; billions of shares still to unlock; its D2D design alienates the very carriers and governments now leaning toward sovereignty.
  • Next catalyst: Further lockup expirations and index rebalancing (see Read-throughs).

Anduril (private), a partnership rumor to watch, not a fact

  • The hyperscaler-episode host flagged that AST is recruiting a "head of radar" and Anduril is hiring for the same role, that Palmer Luckey is an AST investor, and that an AST job posting references "a dedicated minimally modified derivative of ASTS commercial satellite constellation." His read: a possible dedicated defense or "BlackBird" constellation of 90 to 500 satellites. Clearly labeled speculation, but a high-value one if it firms up, because a 500-satellite defense shell is a different model than a 90-satellite one.

Read-throughs

  • Carrier partners (VZ, T, TMUS): No dedicated episode, but this was a live conference week, Verizon and T-Mobile at Citi and Goldman on Wednesday, AT&T at Citi and BofA on Thursday, where analysts were expected to press them on satellite strategy. The structural read stays the same: the US big three are aligned with AST, not Starlink. A Vodafone executive interview quoted on the hyperscaler episode captured the operator tone: "What's becoming more interesting is the involvement of satellite… Vodafone has got a deal with AST… whole new packages and propositions being developed for these new verticals."
  • EchoStar / Hughes (SATS) and Viasat: No dedicated episode, but both surfaced as the incumbent holders of Europe's 2 GHz S-band licenses expiring May 2027, that is, the spectrum the new European JV wants. Separately, one host cited a claim of a talent exodus at Viasat, "12 VP+ level execs gone since June", as a contrast to AST's hiring. (Attribution is a third-party tweet relayed on the podcast; treat as unverified.)
  • Iridium (IRDM): Quiet, only referenced as the historical yardstick for AST's launch scale.
  • Amazon Leo: The under-discussed third D2D entrant, named by the NTIA official as pursuing the same satellite-to-phone spectrum.
  • SpaceX private-market valuation as sentiment anchor: The clearest tell this week. On The Compound and Friends (9/8), the panel noted SpaceX's post-IPO share unlock, roughly 900 million shares, was "absorbed" by the market with barely a ripple, and that an index rebalance could raise its NASDAQ weighting from roughly 1.25% toward 1.5%, adding buying pressure. On Rich Habits (9/11), the hosts coined the frame directly: Wall Street now fears the "SpaceX slump," down roughly 30% from a roughly $200 June peak, as the cautionary template for the coming Anthropic mega-IPO. And on The Rollup (9/9), investor Dan Held laid out a maximalist thesis, SpaceX to $10 trillion in three years and $100 trillion in 20 years, a useful marker of just how wide the sentiment range on this name still is. Notably, short-seller David Einhorn (Greenlight) publicly called SpaceX a bubble this week; one AST host read that skepticism itself as a contrarian bottoming signal.

What changed vs last week

  • The story rotated from Berenberg to sovereignty. Last week was dominated by Berenberg's first big-bank initiation ($92 base, $242 bull, $23 bear). This week the Berenberg framework was still discussed, but the new news was Japan's ITU filing, Europe's carrier JV, the UK's £7.8 billion plan, and the NTIA spectrum roadmap, a shift from "here's a valuation" to "here's why governments are lining up behind AST."
  • A cleaner read on Bluebird status. Last week's snapshot logged 13 Bluebirds "in orbit" after the early-August batch; this week clarified that being in orbit is not the same as being deployed, BW11 is up but hasn't unfurled, and set expectations for an October (not September) launch.
  • SpaceX: the slump continued but the feared unlock was a non-event. Last week the story was SpaceX reclaiming roughly $149 and its IPO price; this week the roughly 900-million-share unlock got absorbed, yet the stock still sits roughly 30% below its June peak.
  • Still quiet, now for multiple weeks: Rocket Lab (RKLB), a second straight silent week (no Neutron date, no Iridium-deal update). Globalstar (GSAT), multi-week silence; no Apple emergency-SOS chatter. Iridium (IRDM) and standalone EchoStar (SATS), no dedicated coverage.
  • New name on the board: the Anduril partnership rumor and the possibility of a dedicated 90 to 500 satellite defense constellation, brand-new this week, and pure speculation for now.