For about 15 percent of US households as of 2023, the only internet options are crappy, especially in rural areas. But thanks to the US Broadband Equity Access and Deployment (BEAD) program, which aims to connect everyone in the US to high-quality, mostly fiber-based internet, that’s close to changing. Or at least, it was.

The plan’s lead architect, Evan Feinman, says that before he was forced out by the Trump administration in March, three US states were just one “minor administrative step” away from connecting their first residents under BEAD. In fact, he says, they could have started the process already — if not for US Commerce Secretary Howard Lutnick, a recent Trump appointee. In March, Lutnick announced a “rigorous review” of BEAD, which he claims is too “woke” and filled with “burdensome regulations.” Now the plan may change.

Lutnick’s changes to BEAD could hand a lot of the program’s money over to private interests like Elon Musk’s satellite internet provider, Starlink. And for every person whose home is served by a Starlink connection, their internet will likely be slower, less reliable, and more expensive than what BEAD might have gotten them with fiber.

You can think of BEAD as a modern version of the nationwide US telephone network or electrification projects. It’s been more than three years since the Biden administration established the $42.5 billion program, and so far, it hasn’t actually connected anyone to the internet — a common criticism coming from the political right. But Feinman says that’s by design. 

“At every step of the game, states were screaming that we were going too fast,” he tells The Verge. “When we made allocation, more than a dozen states sent us letters saying, ‘Go slower, go slower, go slower.’” In fact, Feinman says, the first states would already have started construction if it wasn’t for delays by the Trump administration, which has been encouraging states to redo finished phases of the program to make more space for satellite internet. That shift, Feinman said in April, could effectively mean “millions of private capital is in the garbage.”

Until now, BEAD has spent its time laying the groundwork to connect people. The states were given a year and a half to come up with proposals. According to the BEAD progress dashboard, 38 states have either begun or, in the case of West Virginia, finished picking service providers for the proposals. Nevada, Delaware, and Louisiana are just waiting for final approval from the National Telecommunications and Information Administration so they can sign contracts with companies, start laying fiber, and setting up infrastructure.

In his March statement about BEAD’s review, Lutnick accused the Biden administration of “woke mandates, favoritism towards certain technologies, and burdensome regulations.” The Commerce Department “is revamping the BEAD program to take a tech-neutral approach that is rigorously driven by outcomes.” 

He didn’t specify what he hoped the makeup of this “tech-neutral” approach would look like. But Feinman told Financial Times in March that before he left, Lutnick had instructed BEAD’s workers to give more priority to satellite connectivity and “singled out Musk’s provider, Starlink.” Musk, of course, has been a key player in the Trump administration as well as a vocal critic of BEAD. He’s also been accused of trying to enrich himself using his unofficial, yet seemingly very powerful position within the US government. 

The Wall Street Journal reported Lutnick planned to overhaul the program in a way that could funnel as much as $20 billion, or close to half the program’s overall funds, to Musk’s Starlink. When I asked Feinman about this number, he said it’s hard to know how much more money will go to satellite networks, but that the changes Lutnick is proposing “will shift tens of thousands, hundreds of thousands, or millions of locations away from getting a fiber connection and on to the satellite networks.”

As for Lutnick’s “tech-neutral” comment, Brian Mitchell, director of Nevada broadband office OSIT, says, “I think that’s what we did from the start.” Given Nevada’s vast geography and sparse population, he says, “it was never going to be realistic for us to do 100 percent fiber like you might see elsewhere.” Nevada’s final proposal called for fiber internet in 80 percent of its locations, with satellite and fixed wireless internet making up what’s left. 

As an example of what’s on the line, Feinman says West Virginia’s proposal included “a fiber connection for every single West Virginia home and business” with $150 million to spare and ahead of its deadline. That sounds pretty good for a state that says it ranks 50th out of 52 (all the states plus DC and Puerto Rico) in broadband connectivity. 

Now, plans like West Virginia’s could shift many of the proposed fiber connections to satellite instead. Currently, the BEAD program requires states to prioritize fiber over any other connection method for a given location, unless building it out would cost more than a certain cap — a cap each state was allowed to set. But one of Lutnick’s ideas is apparently to issue a single nationwide cap.

According to Feinman, if Lutnick’s “one-size-fits-all cap” is low enough, satellite internet companies will always win out. In Nevada, for instance, that could see the ratio of fiber-to-satellite flipping from around 80 percent fiber to about 70 percent satellite internet, according to Feinman.

Not long after Lutnick announced his BEAD review, Feinman emailed his team to say he was leaving and warned that Lutnick’s actions could mean “stranding all or part of rural America with worse internet so that we can make the world’s richest man even richer.” Feinman says that he had offered to continue running BEAD, but that the administration declined to renew his tenure.

Although Lutnick hasn’t detailed specific plans for BEAD, he’s already at work changing the program. Near the end of March, West Virginia Gov. Patrick Morrisey announced he’d met with Lutnick and got his state a 90-day extension that Feinman says West Virginia will use to reopen service provider bidding, endangering the state’s pending fiber agreements. The Commerce Department has since issued a blanket invitation to all states to seek such extensions. That likely means at least some states will turn to slower, pricier internet for their citizens. 

There are states where Starlink makes sense. Mitchell says that his state “knew that satellite was going to play a big part in connecting Nevadans, and that’s the result that we delivered.” Satellite connections make up 10 percent of the state’s final proposal. That’s a “great result for Nevadans,” he says, adding that providers are prepared to sign contracts and start building out connections “as soon as the paperwork comes in from NIST,” or the National Institute of Standards and Technology. The first Nevadans could be connected as soon as this summer, according to Mitchell. (Neither NIST nor the NTIA responded when The Verge asked about the status of those approvals.)

Continued delays could be costly. Garry Gomes, CEO of Sky Fiber in Nevada, urged Lutnick earlier this month to push Nevada’s BEAD program forward in a letter to the NTIA forwarded to The Verge by OSIT. He said Sky Fiber — Nevada’s biggest awardee under BEAD — has “already invested over $360,000 in equipment, engineering, and staffing” and that its teams are ready to “immediately” start construction on BEAD deployment.

“Yet the project remains stalled,” Gomes wrote. “Delays not only risk higher costs and lost labor but also erode public confidence in what is arguably the most ambitious and promising broadband investment in our nation’s history.”

Would more Starlink involvement be such a bad thing? Sure, it doesn’t come close to touching fiber, but a service only has to offer over 100Mbps down and 20Mbps up, with no less than 100 milliseconds of latency, to qualify as “reliable broadband internet” under BEAD. Feinman says Starlink, while a “really really good technology,” offers service that’s “barely nosing over” that requirement. This was something that Ookla echoed in December with its median findings from Starlink users in Maine, showing that median users saw 116.77Mbps down and 18.17Mbps up. But, hey, a D minus is still a passing grade, right?

But while there’s room in BEAD for satellite internet, it’s not a replacement for hard-wired, ground-based connections. Even Musk himself has said that, having called satellite a “nice complement” to tech like fiber and 5G back in 2021.

Performance-wise, Starlink pales in comparison to fiber, which increasingly offers as much as 5 gigabits per second, down and up. That’s more than 43 times faster than Ookla’s reported median speed for Starlink. It’s 20 times faster than 249Mbps, the best-case-scenario throughput the company shows for my midwestern address. And it’s hardly worth comparing the upload speeds between the two options.

Starlink service isn’t just undeniably slower than fiber; it tends to be more expensive, too. Its fixed service is $80–$120 per month and comes with pricey upfront equipment fees. (Starlink’s site advertises $30–$50 monthly, but when I tried to sign up for service while reporting this out, the price jumped to $120 a month with a $349 equipment fee, for a “total due today” of $376.57.) In some areas, Starlink’s website says its equipment is free if you sign a 12-month contract. Meanwhile, AT&T offers a gigabit fiber plan in my city of Milwaukee for $80 per month, with a $150 equipment fee.

Even if all things were equal, Feinman says the company can’t scale to the needs of the program, as it “doesn’t have the capacity to serve that many locations.” What’s more, it’s costly to maintain Starlink’s service, which requires rocket launches into space to replace dying or outmoded satellites. Subterranean fiber, on the other hand, can last decades before needing to be replaced. 

Before the Trump administration started futzing with BEAD, Feinman says the program enjoyed a great deal of bipartisan support.

“This is not what anybody outside of a very small circle of Trump administration folks wanted the program to become,” he says. “This is not what Senate Republicans wanted it to become. It’s not what any member of the Democratic coalition on Capitol Hill wanted. It’s not what Republican governors wanted. It’s not what the industry wanted.”

Mitchell echoes that, saying that both Republicans and Democrats in his state have “been very supportive” and adding that “nobody is campaigning on slower, more expensive internet for their constituents.” He also says he doesn’t think that has changed as the administration changed hands. “All of our local officials and local governments are excited for who was awarded,” Mitchell says, “and are ready to start working with them to issue the necessary permits so they can move forward with deployment.” Instead, they’re waiting for Lutnick’s review.

A bipartisan group of 115 state legislators from 28 states signed a letter to Lutnick in April, stating that while they welcome some changes to BEAD, they urge him to make them optional. “At this late stage, major changes would undermine our work and delay deployment by years,” they wrote.

As for what’s next, Feinman says he’s working hard to make noise about what’s being done with the program in an effort to preserve its core mission of connecting everyone to high-speed broadband internet. His departure letter in March said he thought the BEAD program would still mostly work even without so-called “woke” requirements, which include things like fair and safe labor practices as well as outreach to historically underrepresented and marginalized groups. He also said “shovels could already be in the ground … in half the country by summer” without Lutnick’s proposed changes and that if the administration let the program move forward, “it would be a huge political win for the Trump team.” 

“Instead,” Feinman says, “they’re working very hard to significantly delay a program they said was moving too slowly.” 

Discord is beta testing “Discord Orbs,” a new in-app currency that can be redeemed for digital items, Discord has announced. Users earn Orbs by completing the app’s promotional Quests, Discord’s initiative that rewards participants who interact with advertiser content like videos or stream specific games to friends.

In addition to spending Orbs on regular items on the Discord Shop, users can exchange the digital tokens for Orb exclusives like special badges or 3-day passes to try out Discord’s subscription service, Discord Nitro. Discord says Orbs are rolling out globally to a “small number” of its users to start before a wider rollout. If you’re part of the beta test for Orbs, you will get a notification like the one below. 

Before this, publishers or brands that offered Quests had to provide their own rewards — things like avatar decorations or in-game bonuses. They can still do that if they want, Discord spokesperson Bradley Sheets tells The Verge in an email; awarding Orbs is simply an alternative option.

The Browser Company has said repeatedly that it’s not getting rid of the Arc browser as it moves onto its new AI-centric Dia browser. But what the company also not going to do is develop new features for it. A new blog post from CEO Josh Miller explains why, and what happens next.

The Arc browser was a big rethink of what browsers should be like, and it has dedicated users, including yours truly. But a lot of the reasons for ceasing Arc’s development that Miller gives in the blog — like that it’s too complicated to go mainstream, that it was slow and unstable at times (true!), or that The Browser Company wants to recenter the experience on AI — he also gave back in October. 

Why not just roll Dia into Arc? One big thing Miller mentions is security. Arc has had at least one big security issue: a security researcher discovered a vulnerability last year that The Browser Company quickly patched, but which let attackers insert arbitrary code into a users’ browser session just by knowing their user ID. According to Miller, The Browser Company has now grown its security engineering team from one person to five. This focus is particularly important, he writes, as AI agents — AI systems that carry out tasks autonomously — become more prevalent. 

As for what this all means for Arc and its users, Miller still insists that the browser won’t go away. Arc will still get security and bug fixes, and will be tweaked as the Chromium code it’s based on is updated. But he also says The Browser Company isn’t going to open-source or sell Arc, because in addition to Chromium, it’s built on a custom infrastructure that also underpins Dia. He says the company would like to open the browser up someday, but not until “it no longer puts our team or shareholders at risk.”

The Browser Company didn’t immediately respond when The Verge asked whether that same bigger security team is also working to shore up the security of Arc itself. We will update as we learn more.