California’s BEAD Plan Is Final: What the Satellite Awards Mean for the North Coast
Josh Madrone
On September 17, 2026, the California Public Utilities Commission adopted Resolution T-17920, ratifying the state’s revised Broadband Equity, Access, and Deployment (BEAD) Final Proposal. NTIA had approved the plan on July 17, after a round of “curing” requests that cut costs and reshuffled projects.
BEAD has been talked about for years, so it’s easy to tune out. But this is the version that actually gets built, and the project data has specifics that matter for anyone doing IT, cameras, or access control in rural Humboldt, Del Norte, Trinity, Mendocino, and Siskiyou counties.
The numbers
From Resolution T-17920:
- Total: $1,415,047,585
- Terrestrial broadband: $1,338,487,044 to build to 196,385 locations
- Low earth orbit (LEO) satellite: $76,560,541 for service to 72,937 locations
- Eligible locations: 269,322 statewide, 137,569 unserved and 131,753 underserved
- Projects: 83 recommended for award
NTIA’s curing process reduced funding by $72,148,261 on the terrestrial side and $88,802,581 on the LEO side, and the resolution notes 4,161 terrestrial and 65,081 LEO broadband serviceable locations removed.
For scale: when the CPUC approved the original Final Proposal in December 2025, the program was described as up to $1.86 billion. The approved plan is about $1.42 billion.
Two parties filed comments on the draft resolution: Further Reach Inc. asked for location reclassification, which staff declined, and Tarana Wireless proposed implementation support measures.
Doing the math on the resolution’s numbers, roughly 27% of eligible locations statewide (72,937 of 269,322) are slated for LEO satellite rather than a wired or fixed wireless build.
What the project data shows for our area
The resolution comes with data attachments, including a deployment projects file. It has no county column, but it does list the tribal lands each project intersects, and that’s where the North Coast shows up.
Amazon Kuiper Commercial Services LLC — non-geostationary satellite:
- BEAD funding: $54,943,933.35, with $13,815,958.82 in subgrantee match
- Tribal lands intersected include the Hoopa Valley Reservation, Yurok Reservation, Big Lagoon, Karuk, Round Valley, Manchester-Point Arena Rancheria, Potter Valley, and Cloverdale, among others statewide
Space Exploration Technologies Corp. (SpaceX) — non-geostationary satellite:
- BEAD funding: $21,616,607.83, with $19,798,485.24 in subgrantee match
- Tribal lands intersected include the Karuk Reservation, Quartz Valley Reservation, Round Valley Reservation, and Big Lagoon, among others statewide
Those two awards add up to the $76.56 million LEO total. Both list an estimated performance period running to June 2040.
On the terrestrial side, Siskiyou Telephone appears with two projects: $1,103,173.71 for fiber to the premises plus licensed-by-rule fixed wireless, and $343,967.54 for licensed-by-rule fixed wireless. The data file doesn’t say which communities those serve, so I’m not going to guess.
📝 Note: “Intersects tribal lands” in the data means some BEAD locations assigned to that project fall on those lands. It doesn’t mean every home there is getting satellite, and it doesn’t replace tribal fiber projects already underway. For example, Yurok Telecommunications Corporation’s YTEL fiber-to-the-home project, about 116 miles across Yurok lands in Del Norte and Humboldt counties, received a CEQA Notice of Exemption from the CPUC in May 2026.
What happens next
According to the CPUC, states have six months to enter into subgrant agreements with providers, and that clock starts after the CPUC acts on the revised Final Proposal. The Commission acted on September 17. So the next milestones are signed agreements, then construction and service activation.
For a lot of rural North Coast addresses, the realistic outcome is going to be one of three things: a fiber build (BEAD, CASF, or tribal), fixed wireless, or a LEO satellite subscription. Some sites will have more than one option over time.
Designing security and IT for a site on LEO
This is where I think an integrator’s view is useful. A satellite connection that’s “good enough for Netflix” can still be the wrong design for a camera system or a door controller if you build it like it’s on fiber. What I recommend for rural sites that end up on LEO:
- Record locally, always. The recorder lives on site. The internet connection is for viewing and alerts, not for getting video off the property.
- Budget upload, not download. Remote viewing and cloud backup push data up. Plan around the uplink you actually measure, not the advertised download speed.
- Use sub-streams for remote viewing. Most VMS platforms can serve a lower-bitrate stream to phones. That’s the difference between a usable remote view and a frozen one.
- Assume you can’t port forward. Many satellite and fixed wireless services put customers behind carrier-grade NAT (CGNAT). Pick systems that connect outbound to the vendor’s cloud or use a VPN initiated from the site. That is the safer design even on fiber.
- Plan for outages. Access control should keep working locally when the internet is down. Door schedules and credentials need to live on the controller, not only in the cloud.
Is this connection behind CGNAT?
This quick check traces the first few hops and looks for addresses in the shared address space CGNAT uses (100.64.0.0/10, from RFC 6598). It needs traceroute installed. It’s a heuristic, not proof, so use at your own risk.
#!/usr/bin/env python3
"""Rough CGNAT check: look for 100.64.0.0/10 addresses in the first few hops."""
import ipaddress
import re
import subprocess
CGNAT = ipaddress.ip_network("100.64.0.0/10")
def first_hops(target: str = "1.1.1.1", max_hops: int = 5) -> list:
out = subprocess.run(
["traceroute", "-n", "-m", str(max_hops), "-w", "2", target],
capture_output=True, text=True, check=False,
).stdout
hops = []
for line in out.splitlines()[1:]:
match = re.search(r"\b(\d{1,3}(?:\.\d{1,3}){3})\b", line)
if match:
hops.append(ipaddress.ip_address(match.group(1)))
return hops
if __name__ == "__main__":
hops = first_hops()
for hop in hops:
tag = " <-- CGNAT range" if hop in CGNAT else ""
print(f"{hop}{tag}")
if any(hop in CGNAT for hop in hops):
print("Likely behind carrier-grade NAT: inbound port forwarding won't work.")
else:
print("No 100.64.0.0/10 hops seen. Still confirm with the provider.")
Will the uplink carry remote viewing?
A back-of-the-envelope calculator for how many camera sub-streams a measured uplink can carry, with headroom left for everything else on site:
#!/usr/bin/env python3
"""How many camera sub-streams fit in a measured uplink?"""
def max_streams(uplink_mbps: float, substream_kbps: int, headroom: float = 0.4) -> int:
"""headroom = fraction of uplink reserved for everything else (0.4 = 40%)."""
usable_kbps = uplink_mbps * 1000 * (1 - headroom)
return int(usable_kbps // substream_kbps)
if __name__ == "__main__":
uplink = float(input("Measured uplink (Mbps): "))
sub = int(input("Sub-stream bitrate per camera (kbps): "))
cams = int(input("Cameras someone may view at once: "))
fit = max_streams(uplink, sub)
print(f"About {fit} simultaneous sub-streams fit with 40% headroom.")
if cams > fit:
print("Lower the sub-stream bitrate, or plan for fewer cameras in a live view.")
Measure the uplink at a few different times of day before you trust the number.
Why I’m watching this
We work across Humboldt County, and a lot of the sites that need cameras and access control the most, including ranches, gates, remote shops, and community buildings, are exactly the addresses BEAD is aimed at. The final plan tells us which technology many of those addresses are likely to get, and for a meaningful share of them that’s satellite.
That’s not bad news. It just means the security design has to fit the connection. Local recording, outbound-only remote access, sensible bitrates, and controllers that keep working offline are good practice on any link. On LEO, they’re the difference between a system that works and one that generates service calls.
Sources
- Resolution T-17920 (BEAD Final Proposal ratification) — California Public Utilities Commission, adopted September 17, 2026
- California BEAD Final Proposal — resolution and data attachments — California Public Utilities Commission, September 2026
- Appendix B2 – Deployment Projects (CSV) — California Public Utilities Commission, September 2026
- NTIA approves California’s final BEAD proposal — Fierce Network, July 21, 2026
- California Receives NTIA Approval for BEAD Final Proposal — Benton Institute for Broadband & Society, July 20, 2026
- CPUC Finalizes BEAD Plan to Secure Up to Billions for Statewide Broadband Expansion — California Public Utilities Commission, December 18, 2025
- YTEL FTTH Yurok Telecommunications Corporation (T-17907) — CEQAnet, May 14, 2026
