The latest on the wireless decision.

Ars Technica's senior space editor, Eric Berger, published a detailed report this month on what the satellite industry is calling a launch crisis: far more satellites need rides to orbit than there are rockets to carry them. “What we have is an industry in panic,” one analyst told him. The examples are striking — companies are asking Telesat to share the eleven Falcon 9 launches it has booked (the answer is no); AST SpaceMobile says launch availability is now the pacing item for its whole constellation; and Amazon has throttled back production of its Leo satellites because the launches it bought half a decade ago still aren't ready to fly.

Underneath the shortage is one fact this site has documented all year: the Western world effectively has one workhorse rocket, and one company owns it. SpaceX flew 165 of the world's orbital launches last year — about half of everything launched from Earth — and Berger reports it is now winding the Falcon 9 down in favor of Starship, whose satellite dispenser is built around SpaceX's own Starlink v3. SpaceX president Gwynne Shotwell has said the Falcon 9's operational life may end around 2030–2032; Ars reports, citing one source, that SpaceX has told the US government it intends to stop selling commercial launches after 2028. (That last date is a single unnamed source — treat it as reported, not established.)

Read it the honest way: this is not a “Starlink is failing” story. SpaceX owns the rocket, so its own constellation is the one satellite network on Earth not short of launches — and this site's supply-chain review already credits that. The story is about the rescue narrative. The comfortable version of the county's bet says: even if Starlink's prices climb, Amazon Leo or somebody else will show up and competition will hold the line. Berger's reporting says the somebody-else can't get to orbit — Amazon is slowing its satellite factory for lack of rides while an FCC deadline requires all 3,232 of its satellites flying by 2029 — and the one seller of rides is leaving the market. As our Starlink review put it before this news: a rescue that rents its ride from the incumbent arrives on the incumbent's schedule, at the incumbent's price — or not at all.

Meanwhile the network the county already owns sits bolted to mountaintops, with years of spares on US warehouse shelves, needing no launch window, no fairing, and no place in anyone's queue. When the industry that was supposed to discipline Starlink's prices is fighting over seats on Starlink's own rocket, the case for keeping a local yardstick isn't weaker. It's the whole point. Why a backup isn't a foundation →

Some good news, and we'll say it plainly: the District is doubling the speed of most of its rural wireless access points. The top tier on those Cambium radios is roughly doubling — to a high of about 40 Mbps down and 10 Mbps up. After years on the same 2010-era gear, the network the county was on the verge of writing off is getting faster, and that helps real households today. It's worth naming for what it is: progress.

Now the honest part, because that's the standard here. This is a speed bump on the existing Cambium gear, not the modernization the county still needs: 40 down still rides radios past their published service life, and it's still well behind both the 250/50 upgrade and the 600 Mbps a competitor advertises on the same ridges. A stopgap that raises the floor is genuinely good — it just isn't the fix, and no one should mistake it for one.

Here's why it matters for the decision in front of the Board. The case for pausing the upgrade leaned on a story that the wireless network was a fading asset not worth more money. Doubling its speed tells the opposite story: the District is choosing to reinvest in it, because the demand is real and the network works. You don't upgrade something you're letting die. This is the moment to finish the job — the same seven towers, taken from a 40 Mbps stopgap to a modern 250/50, at a price the towns' only other option can't match.

And progress like this doesn't happen in a vacuum. It happens when local providers show up at Board meetings, when more than a hundred neighbors put their names on the record, and when a public asset gets the public attention it deserves. Keep it up. Add your voice →  or  see what the full upgrade delivers →

When we started gathering them, we hoped for a couple of dozen. As of today the Voices page carries 115 first-person statements — and for the first time they come from every one of the 14 towns on the county map, Oroville in the north to Pateros in the south, Wauconda in the east to the Methow in the west. There is no corner of the wireless service territory that isn't speaking.

Of the 115, 82 speak for their own households — retirees and ranchers, remote workers, families stuck on satellite. The other 33 speak for a farm, a business, or one of the county's own internet providers, 14 of them under their business's own name: a lakeside resort and two dental practices, an orchard, a lumber yard, an insulation contractor, a housing nonprofit, and an Aeneas Valley fire district that was pushed onto satellite when its service failed and now wants a local option back for redundancy. Most striking of all, three of the county's local internet providers — CommunityNET, Will Connect, and Highland Internet — have put their own names on the page.

What they describe is the whole shape of this issue in plain language. Households that struggled after the out-of-area buyout and found their footing again with a different local provider — “same tower, same infrastructure, but someone installed it who cared.” Families stuck on Starlink because nothing else reaches them, who say they'd switch “in an instant” if a local option did. Business owners who can't get a static IP or an on-site repair from a satellite. And people who are perfectly happy today but want the option kept anyway — “if they get bad later, we will have no escape and no options.”

Every statement is first-person, given to the campaign by a real Okanogan County resident or business and published only with that person's approval — shown by first name and last initial, or by business name where an owner speaks for their company. Each one is anchored to a place, so you can see them light up town by town on the county map; and because the record has grown large, you can now filter it — by who's speaking (residents, businesses, local providers) or by what they're talking about.

If you've lived any of this, the next voice on the page is yours. It takes two minutes, and it's what the commissioners actually read. Add your voice →  or  read the whole record →

Most of the argument on this site is numbers. Our newest page is the machine itself: an interactive, hand-drawn 3D climb of a real tower site — Jackass Butte — in fourteen stations, from the dirt to the rooftop it serves. Drag to look around, and climb.

It starts where every network actually starts: the ground somebody had to lease. Jackass Butte belongs to a local family, leased for $5,500 a year — the District owns none of its nine mountaintop sites. From there you climb through the cabinet where fiber light becomes radio, up the caged ladder past the bolted joints, to the sector array, the GPS antenna that keeps every radio in lockstep (and runs on Einstein — the satellite clocks need a ~38-microsecond-a-day relativity correction), and the span of open air the District's own field test crossed at 900+ Mbps down.

Six stations go deeper than any brochure would. A beamforming simulation computes the actual interference physics of an antenna array — watch one antenna's ripples organize into a narrow beam aimed at a single rooftop, with the textbook radiation pattern drawn over it. Then the guide keeps going into what makes the proposed gear genuinely different: the "beam" dissolving into frequency-specific beamlets, each slice of the channel aimed on its own; a cabin with no line of sight served off the side of a barn — the reflection recombined into a working link, drawn straight from the vendor's published barn-bounce field test (and labeled as the vendor's test); a live oscilloscope card showing burst interference being subtracted out of the received signal; and the 6 GHz band's protected microwave links, with the radio stitching the leftover fragments into one service. And the last stop is the local installer: the District builds and runs the network, but the person on your ladder is a neighbor from one of the county's local providers — often the owner — who checks the speeds against what you were promised before the truck leaves.

Every figure in it traces to the record: the lease, today's 3–20 Mbps tiers, the 250/50 upgrade plan, the field test, even the grounding code on the subscriber wall. The one thing that isn't a portrait is the steel: the tower is drawn as a simplified pole to keep the radios readable, not as a likeness of any one structure — real mountaintop sites vary, and often carry several towers of different builds. The guide says so on the way up. Begin the climb → (You'll also find it under Explore in 3D in the menu, beside the network map.)

The numbers on this site tell you what's happening. Our new animated map shows you the shape of it — in three minutes, on Okanogan County's real mountains and valleys. Press play and watch the whole arc: Enloe Dam and the out-of-town power company (1906–1929), the county buying its own grid for $2,314,240.33 in 1945, the 82-mile fiber backbone drawing itself up the valley in 2002, the 2010 federal wireless build blooming across ~80% of the service territory — and then 2026, when the awards start going back.

It's built from the real record: USDA's own award polygon for the $30.2M grant the District is moving to hand back, the District's actual Conconully fiber engineering routes from its state application, the PUD's own published tower map, federal terrain and grid data. A running ledger tracks what a century of public investment built next to what 2026 is unwinding. Anything schematic is labeled, and the two places where sources disagree are footnoted right on the page.

It's honest, it's sourced — and frankly, it's fun to watch. Scrub the timeline yourself, or let it play. Watch it here →

On June 30, 2026, SpaceX cut Starlink service across the Memphis metro to half price — roughly $27.50–65/mo instead of $55–130, with no hardware cost. The reason wasn't a breakthrough in what satellite internet costs to deliver. It was a community-relations move: SpaceX has been fighting backlash — including pollution and noise complaints — over the enormous xAI/Colossus data centers it runs in the area, and cheap internet for the “neighbors” helps buy goodwill for projects it wants accepted.

Read that the right way. This is not a sign that Starlink could simply choose to be cheaper for everyone, and it's no reason to cheer. It's the opposite lesson: Starlink's price isn't tied to what the service costs, or to any public-service commitment. It's a lever the company pulls whenever moving it serves the company — downward in Memphis this month, because a discount there advances a data-center fight worth far more to SpaceX than the forgone subscriptions.

When the customer is the one who's stuck, the lever goes the other way. In the same stretch, Reuters reported that SpaceX pushed the Pentagon from about $5,000 to $25,000 a month per drone connection — a fivefold jump — after the military had built critical weapons around Starlink and had no comparable alternative; the Pentagon reportedly paid. (SpaceX and the Pentagon dispute the report.) And rural Western customers with the fewest alternatives already sit at Starlink's top residential tiers plus regional congestion surcharges. Cheap where Starlink wants something from you; expensive where you have nowhere else to go.

None of this makes SpaceX a villain — it's simply how a single, distant, investor-owned provider is built to behave: its price answers to its owner and shareholders, not to a rural county's budget. That's exactly the difference the PUD's wireless network protects. The public network's rates are set in open meetings by locally elected commissioners, on the public record, with several local providers competing on top of it. It doesn't stop anyone from choosing Starlink — it keeps Starlink from becoming the only choice. That's the line between being a customer and being dependent.

Two recent board actions show the District investing in the core of its network even as it pauses the wireless layer that reaches rural homes. On May 18, 2026, the Board unanimously approved pursuing state funding for a "Carrier Ethernet Backbone Upgrade" (a grant application to the Washington Department of Commerce Public Works Board). And the June 22 agenda took up a motion to "purchase network hardware to support a 400G backbone" — a 400-gigabit core. (The outcome of the June 22 vote awaits the District's posted minutes.)

Why a backbone upgrade is a big deal. The backbone is the high-capacity "middle mile" that carries all of the county's traffic to and from the wider internet. Lifting the core to 400 gigabits and modern Carrier Ethernet is a large jump in capacity — and more backbone capacity generally means a lower cost to move each megabit and removes the core as a ceiling on speed. In plain terms, it's the District laying the foundation for faster, cheaper broadband across the county. That's genuinely good news, and worth recognizing.

But a backbone only reaches a home through the last mile. The core doesn't connect a single rural house on its own — it gets there over the access layer: the towers and radios of the wireless network. For most of the county, that wireless network is the last mile. A 400G backbone makes gigabit-class rural service possible — the District's own Tarana field tests already clocked 243–946 Mbps — but only if the towers are modernized to carry it. On the aging Cambium gear the District has declined to replace, a rural home still tops out near 20 Mbps no matter how fast the backbone behind it runs.

So the same District building a faster, cheaper core is declining the roughly $1.2 million it would take to upgrade the towers that carry that capacity the final miles to rural homes. It is widening the highway while leaving the rural on-ramps to age out. The backbone investment doesn't weaken the case for the wireless upgrade — it strengthens it. The capacity is coming; the open question is whether the rural majority will have any way to reach it.

On the same June 22 agenda, the Board took up Resolution 1863 — "Rescind RUS ReConnect Award WA 1707-A73": a move to give back the roughly $30.2 million USDA ReConnect grant the District accepted in 2023 to build fiber to the northeastern communities — Chesaw, Molson, Havillah, Nine Mile, and Siwash Creek. (As with the other June 22 items, the recorded vote awaits the District's posted minutes.)

That grant — about 245 miles of fiber to roughly 745 homes and 21 businesses — was awarded in 2023, but by the end of 2024 the District had spent $0 of it. Now it's moving to return the award.

Here's the pattern that should worry rural customers. The District accepted tens of millions to run fiber to dispersed rural homes, couldn't get it out the door, and is handing it back — while the wireless network that already serves those same rural areas today is the one being left to age out on unsupported equipment. Fiber to the far corners is the part the District can't execute; wireless is the part it won't maintain. The real risk for the rural majority is ending up with neither.

At the Okanogan County PUD's June 22 meeting, several local retail internet providers — the companies that deliver service over the District's wireless network — and a county ratepayer addressed the Board, urging it to reconsider the pause on the wireless upgrade. What followed was an open, constructive discussion. The recording below is the record — watch it and judge for yourself. (We summarize the gist here rather than quote it: the room audio in the original recording is rough. We've since produced a cleaned-up audio version — noise-reduced and leveled for clarity, content untouched — which you can listen to or download below the video.)

There was real common ground. The discussion surfaced genuine concern — shared across the room — about the county losing its local providers and ending up dependent on a single out-of-state satellite, and about how hard and costly such a network is to rebuild once it's gone. That is the heart of this campaign's case, too.

What's holding up the wireless upgrade — fairly stated. The upgrade is not in the District's 2026 budget, and management asked for a clearer business case given the subscriber decline — reasonable questions from stewards of public money. At the same time, the District is investing about $4.3 million in fiber to Conconully (367 homes) and upgrading its core backbone. We support the Conconully build — it's a worthwhile project for an unserved community that needs it. Our question is simply how the District can also fund the ~$1.2 million wireless upgrade — seven towers serving 878 customers today, on a network of ~2,700 — which is affordable from the network's own surplus.

The encouraging part: the Board agreed a workshop is the right next step — sitting down with the providers to work through performance targets, a phased rollout, and the business case. That is exactly what the ratepayers asked for. Now it needs to follow through.

It isn't only the ReConnect money. In March, the Methow Valley News reported that the Okanogan County Electric Cooperative and the PUD jointly lost a $19.7 million federal BEAD grant (plus a $6.6 million state match) after the federal government changed the program's rules after applications were in, moving to a "tech-neutral approach." As one local broadband manager put it, BEAD "went from buzzword to kind of buzz saw… the federal government changed the rules of the road post-application, and we were not successful."

That's now two fiber awards in retreat — the $19.7 million BEAD grant and the ~$30.2 million ReConnect grant the Board is moving to hand back.

But "tech-neutral" doesn't mean the money disappears — it means it moves. The same shift that cost this fiber grant is why Washington's own BEAD plan steered nearly half its funding into fixed wireless, with the single largest award going to a fixed-wireless provider (see The Case). The funding winds are blowing toward exactly the kind of network the District already owns — the rural wireless system it has now chosen to stop modernizing.