Summary

EchoStar’s $22.7 billion AT&T sale began converting a 15-year trapped-value story into cash while repricing the remaining spectrum upward. Guowei Zhang had valued the 3.45 GHz block at $8 billion and the 600 MHz block at $10 billion; assuming AT&T paid roughly $8 billion for the former, it paid about $15 billion—a 50% premium—for the latter. The surprise was that AT&T paid the high end of expectations without an auction for spectrum it said could take several years to deploy: “The spectrum market has reset.”
Guowei believes $SATS offers better risk-reward in the low-to-mid-$60s than it did near $25 because the transaction materially raised the liquidation floor. His back-of-the-envelope math combines the $23 billion sale with another $11–12 billion of non-AWS-4 spectrum, then subtracts debt and taxes to reach low-to-mid-$50s per share, potentially returned over one to two years. AWS-4, the “crown jewel,” could add $26–27 billion—or roughly $80 per share—supporting “triple digits at least” if realized.
The central risk is that Charlie Ergen uses the proceeds and AWS-4 to chase a $5 billion direct-to-device satellite ambition instead of completing the liquidation. EchoStar could sell or lease AWS-4’s terrestrial rights while retaining its MSS satellite use, but Guowei does not want the full asset “stuck in SATS” behind a late LEO build competing with SpaceX and Amazon. His base case remains continued monetization because EchoStar has abandoned terrestrial mobile and now needs FCC approval for whatever comes next.
AT&T’s embrace of fixed wireless makes the spectrum sale a negative read-through for cable, not merely a positive catalyst for EchoStar. T-Mobile pioneered the category, Verizon is ramping, and AT&T said it was modeling 3.45 GHz for fixed wireless; Andrew estimates the carriers’ fiber and wireless plans require roughly 40 million customers over five years, with cable the obvious source. Cable can buy spectrum and fund a network or keep renting capacity while losing share: “It’s really no good choice for them.”
The DOJ may prefer four national wireless networks, but Guowei sees no viable fourth operator and therefore little practical basis to block the AT&T transaction. Blocking it could force bankruptcy, strand the spectrum in court, and eventually return the same three bidders; a more plausible face-saving remedy would strengthen MVNO terms for cable. The nearer catalyst was EchoStar’s promised Paris-show update, where Andrew expected a D2D partnership and focused more on AWS-4 monetization; Guowei inferred that an AWS-4 deal might already be close because shutting the network weakened EchoStar’s negotiating leverage. More broadly, he thought shareholders would know within two to three months whether the thesis worked.
Broadcaster consolidation may create two scaled affiliate groups, but Andrew Walker and Guowei sharply disagree over what survives the bundle’s decline. Guowei sees irreplaceable local news and content, meaningful deal synergies, and eventual bargaining power against four national networks; Andrew sees affiliates collecting regulatory rents from national sports that could migrate directly to Netflix, Amazon, Paramount, or the networks. Guowei is not long the equities—he is studying credit trades and even a possible post-consolidation short because “there’s going to be a lot of pain.”
The closing lesson was that averaging down can turn a cheap-looking equity into a total loss long before an investor admits the thesis is broken. QVC supplied the specimen: senior debt near 40 cents on the dollar, more debt between it and the equity, yet investors still argued for equity value because of free-cash-flow yield and the Malone halo. Guowei wrote the piece principally as a warning to himself: “You might be right 70% of the time, but that other 30% is—you’re going to lose all your money.”

Deep dive

1. AT&T began releasing value trapped inside EchoStar for 15 years

Guowei’s opening frame was historical: Charlie Ergen accumulated the spectrum over roughly 15 years, while its value remained trapped inside Dish and EchoStar. AT&T’s $22.7 billion purchase is significant because that value has finally “started to come out,” and the sold blocks represent only about one-third of the portfolio.
The transaction covered 3.45 GHz and 600 MHz spectrum. Guowei had marked the former at $8 billion because Dish paid $7.3 billion in the 2022 auction, and he had valued the latter at $10 billion; an inferred $15 billion allocation to 600 MHz would therefore represent a 50% premium.
The speed mattered as much as the price. Guowei expected a longer process, but AT&T bid aggressively enough to avoid an auction: it “wanted to get it first and early,” delivering a price near the top of the market’s $10–16 billion range for the 600 MHz block.
Andrew’s prior concern was that removing Dish—the fourth auction participant—would weaken demand from spectrum-capped Verizon, AT&T, and T-Mobile. Instead, AT&T paid at least cost and probably a premium for recently auctioned 3.45 GHz holdings, undermining the idea that EchoStar’s portfolio lacked buyers once Dish stopped competing.

2. AT&T’s fixed-wireless pivot resets the competitive map

Guowei found the 600 MHz purchase “a little bit strange”: AT&T said it would take several years to deploy, yet still paid a major premium. That led him to conclude that “the spectrum market has reset” at a higher level, though he remained uncertain about bidding dynamics for the unsold blocks.
Andrew offered the bear case—“death, taxes, and AT&T wildly overpaying for telecom assets”—and asked whether the sucker had already left the auction. Guowei pushed back that Verizon has hardly been disciplined, citing Straight Path and C-band, while AT&T itself said the $23 billion transaction would be earnings-accretive.
AT&T said it was modeling 3.45 GHz for fixed wireless, marking a clear change from its earlier resistance to the product. Guowei immediately read that as negative for cable: T-Mobile already pioneered fixed wireless, Verizon is expanding it, and “now you got a big player coming in” to take additional broadband share.

3. Cable faces an expensive build-or-rent dilemma

Andrew estimated that the carriers’ announced fiber and fixed-wireless plans imply roughly 40 million customers are needed over five years: “Where’s it going to come from? Cable.” He described a market where cable faces a fiber competitor plus one or more wireless overlays, rather than its former local monopoly or simple duopoly. Cable is already about 60% overbuilt by fiber and could reach roughly 80%, while fixed wireless may attack about 5% of each market.
Andrew suggested Comcast and Charter could buy EchoStar’s remaining spectrum, combine it with CBRS, offload dense traffic, and retain an MVNO for rural coverage. Guowei questioned whether that works without low-band coverage: 600 MHz is now gone, and a “blotchy network” would not support a competitive mobile product.
Even if leasing low-band solves coverage, owning the spectrum could cost cable $30–40 billion before towers and network investment. Guowei understood why cable shareholders would reject that capital allocation; the alternative is continued dependence on carrier networks while fixed wireless and mobile “eat cable’s lunch.”
EchoStar’s network shutdown was the clearest evidence that Guowei inferred cable had declined the opportunity. After perhaps $8–10 billion of network spending, EchoStar will incur further costs removing radios and dealing with tower leases—“a complete waste of capital.” Andrew captured the inversion: the supposedly valuable new network proved worthless, yet the stock tripled because the spectrum was more valuable than expected.

4. The liquidation floor now supports the stock before AWS-4 contributes

At roughly $62–63 during the discussion, Guowei thought $SATS had “fantastic” risk-reward. The $23 billion already sold, plus $11–12 billion of remaining spectrum outside AWS-4, could produce low-to-mid-$50s per share after debt and taxes—substantial downside protection if liquidation proceeds over one to two years.
AWS-4, the 2 GHz “crown jewel,” sits on top of that floor. Guowei estimated it at $26–27 billion, equivalent to roughly another $80 per share, which is why he believes the equity is “worth triple digits at least” even after its rapid rise from approximately $25 a month earlier.
The rights have two components: terrestrial use and MSS satellite use. An efficient outcome could sell or lease the terrestrial portion to a mobile operator while EchoStar retains the satellite rights for D2D; Guowei would prefer selling the spectrum so cash comes back to shareholders rather than leaving it leased by Ergen.
Guowei noted that much of the sold value would pay down holding-company notes, transporting cash up to the holdco; Andrew said that structure protects the downside. The transaction also removed much of the prior fear that Ergen would continue funding a nationwide terrestrial build, while the premium sale price raised marks across the remaining portfolio.

5. A late D2D moonshot is the thesis’s largest unresolved risk

EchoStar had discussed potentially spending approximately $5 billion on a LEO direct-to-device constellation serving global network operators. The pitch is wholesale connectivity in uncovered areas, potentially saving carriers the capital cost of remote towers, with AWS-4’s satellite authorization supplying the necessary spectrum.
Guowei’s concern is execution: EchoStar knows geostationary satellites, not a LEO constellation containing hundreds or thousands of satellites, appears at least five years behind SpaceX and Amazon, and—most importantly—has not shown him “the people” capable of operating such a system.
The economics are equally daunting. Holding $26–27 billion of spectrum would require roughly $2 billion of unlevered net income at an assumed roughly 8% return, and Andrew argued the company would need visibility toward perhaps $5 billion of operating income. Guowei could see retention being justified only if a large government or defense contract aligned the project with an administration objective.

6. The FCC has leverage, while the DOJ has few workable alternatives

Because EchoStar is shutting its terrestrial network, AWS-4 sits in regulatory limbo: spectrum licenses carry buildout conditions, and future uses require FCC approval. Guowei therefore framed the question as “what does the FCC want to do with the spectrum,” not simply what Ergen wants.
Andrew identified DOJ opposition as the second major risk because antitrust officials historically wanted four national wireless competitors. Guowei became less worried after considering the alternatives: nobody besides cable appears willing to build a fourth network, and he inferred that cable’s lack of interest helped lead to the shutdown.
Blocking the sale could push EchoStar into bankruptcy, strand spectrum in litigation for years, and eventually produce the same three carrier bidders. Guowei said any intervention could instead focus on “stronger MVNO agreements” that preserve some competitive pressure; Andrew suggested cable might receive unusually favorable AT&T access.
EchoStar had repeatedly promised news at the Paris show, which Guowei thought was in the week of September 14, though Andrew only tentatively agreed. Andrew expected a D2D partnership there but cared more about AWS-4 monetization. Announcing the network shutdown before securing that spectrum’s future would otherwise destroy negotiating leverage, so Guowei inferred that a transaction might already be close.

7. Broadcaster consolidation creates scale without resolving terminal risk

Guowei expects the next 12–24 months to combine today’s public broadcasters and smaller operators into two large station groups, each owning two of the four major affiliates in local markets. Nexstar’s $6.2 billion enterprise-value acquisition of Tegna followed reports of a $25–30-per-share Sinclair proposal, while Gray also appeared to need a partner.
Nexstar traded near six times EBITDA and could remain around that multiple after paying Tegna’s premium once synergies are included. Guowei’s central asset is local scale: local news and content are “not replicable” by large technology platforms, giving consolidated affiliates something the national networks cannot manufacture centrally.
Andrew’s pushback was structural: local newspapers, sports, and digital outlets do not monetize like affiliates that receive a cut of expensive bundles carrying national sports. As viewing migrates to Netflix, Amazon, Paramount, and direct streaming, he worries networks will ask why they should keep sharing economics with what he called “a leech on the back.”
Guowei argued the opposite: national networks produce substitutable shows and merely bid for sports, while affiliates own differentiated local content and can bid for sports themselves, including through assets such as the CW. He is nevertheless not long the equities; consolidation may create credit trades or a later short, and he could imagine bankruptcy preceding a smaller local-content business valued at 10–15 times EBITDA.

8. QVC shows how a value thesis becomes a bag-holding reflex

Guowei’s essay began with QVC investors defending the equity even while senior debt traded near 40 cents on the dollar and multiple claims sat between that debt and common stock. The fascination was psychological: balance-sheet evidence could not dislodge attachment to free-cash-flow yield, John Malone, or an old thesis.
He implicated himself rather than mocking others: “I have a lot of experience with bag-holding.” Value investors know they should reassess after a decline, yet the gambling instinct says the lower price confirms the opportunity—“don’t touch the fire, but you still stick your hand in.”
Andrew described the same habit as “sucking my thumb”: a stock falls 20%, fear appears to create a bargain, investors double down repeatedly, and an eventual 80% loss turns great portfolio results into bad ones. Hearing someone justify a ride from $100 to $15 with the original thesis is precisely the warning sign.
Their Malone discussion widened the lesson from individuals to cycles. Andrew questioned whether falling rates and leveraged cable assets helped create the legend before old-media investments faltered; after Guowei said Malone had done “basically no right” from 2010 onward, Andrew cautioned that 2010 misses the SiriusXM bankruptcy grand slam and pointed instead to 2013 for later Liberty Global, LILAC, and Discovery outcomes. Guowei’s broader conclusion was that the internet and mobile “carried out” an entire generation, and AI may do the same again: “It’s more of the time than anybody being so smart or anybody being so dumb.”

The Little Tech Agenda for AI ElevenLabs CEO/Co-Founder, Mati Staniszewski:The Untold Story of Europe’s Fastest Growing AI Startup