From Fiber to Fabric Only Works if BEAD Fixes Its Delivery Model
Commentary on Eric Bathras: From Fiber to Fabric, the Next Evolution of BEAD
By Bhavin Gandecha
Eric Bathras has written a thoughtful piece arguing that the next stage of the federal broadband effort should move beyond single purpose fiber builds and toward what he calls a digital fabric. His argument is that the remaining BEAD funds should not be locked into another round of narrowly defined deployment, but used instead to support shared rights of way, integrated planning platforms, and corridor based strategies that serve multiple needs at once. He makes the case from Maryland, where his team has built a real working version of this idea through their Single View platform and their Digital Infrastructure Group. I agree with almost everything in the piece. The question I want to add is a quieter one. Even if NTIA agrees, and even if Congress agrees, will the delivery model itself actually let this happen.
This is the part of the conversation that I think we keep underestimating. The vision of a digital fabric is sound. The financial logic is sound. The operational benefits are clear, and Maryland has shown that the idea works in practice when the right governance is in place. But the BEAD program, as it has been structured so far, is built around a delivery model that is very good at building one thing in one place for one purpose. Asking it to suddenly deliver a multi tenant, multi agency, shared infrastructure outcome is not a small adjustment. It is a different kind of work, and the program will need to look quite different on the ground if it is going to support it.
The Delivery Model Is the Real Constraint
In every infrastructure programme I have worked on, the moment of truth is not the policy memo. It is the day someone has to pick up a shovel, sign a permit, agree a route with a utility, and stand behind a delivery date. BEAD has shown over the last three years that even with strong intent, even with significant funding, even with detailed proposals, the work of actually building things runs into the same constraints every time. Permitting timelines. Pole attachment disputes. Labour shortages. Make ready costs. Compliance overhead. Subgrantee selection processes that take months and sometimes years. None of these problems are solved by the policy moving from a fiber model to a fabric model. They are problems of execution, and they get harder, not easier, when the delivery scope becomes more ambitious.
If a state is given permission to spend its remaining BEAD funds on integrated corridors that carry fibre, support tower sharing, enable intelligent transportation, and prepare for AI driven services, that state is now coordinating across at least four agencies, several utilities, multiple commercial partners, and a federal compliance regime that was not written with this in mind. The risk is not that the idea fails because it is wrong. The risk is that it stalls, quietly, in the layers of approvals and interfaces it now has to clear. We have seen this pattern before. BTOP taught us that ambitious programs without de-risked delivery end up under-delivering relative to their promise.
What Maryland Actually Solved
It is worth being honest about what is special in the Maryland example. The work Bathras describes did not succeed only because somebody believed in shared infrastructure. It succeeded because there was a governance structure that made shared infrastructure possible to deliver. The Digital Infrastructure Group is a small set of named representatives from named agencies with a clear mandate written by executive order. The Single View platform is real software, populated with real data from agencies that have agreed to contribute their information into a common map. There is a defined process for how a fibre route is proposed, evaluated against multiple use cases, and then approved as a joint build.
In other words, Maryland did the hard institutional work first, and then the technology and the funding followed. If BEAD is going to enable this approach more broadly, the federal guidance has to acknowledge that this institutional work is the precondition, not an optional extra. A state that does not have its own version of DIG, its own version of Single View, and its own executive level coordination function is not going to produce a digital fabric outcome just because the funding category was expanded. It will produce another set of single purpose builds with slightly more flexible cover sheets.
What Federal Policy Should Actually Enable
If we take the fabric idea seriously, the right federal move is not just to permit broader uses of remaining funds. The right move is to permit, and where possible to fund, the delivery capability that makes those broader uses real. That means supporting state level integrated planning platforms. It means funding the governance structures that pull agencies into shared decision making. It means recognising that a corridor planner, a geospatial data analyst, and a programme manager who can speak both broadband and transportation are now part of the critical workforce, not back office overhead.
It also means rethinking how progress is measured. A program that scores success purely by passings connected, or by miles of fibre laid, will quietly punish states that take the longer integrated route. If three states finish single purpose builds in twenty four months while a fourth state takes thirty months to finish a corridor that also carries transportation, public safety, and tower backhaul, the headline numbers will make the fourth state look slow. The fabric framing only survives if the measurement framework treats multi purpose value as the win it actually is.
The AI Layer Is Coming Faster Than the Policy
Bathras also points out, correctly, that the next decade of infrastructure will need to support AI driven applications, data centre interconnection, and a far more demanding pattern of use than today's broadband planning assumes. This is where I think the case for a fabric approach becomes urgent rather than merely interesting. A state that builds out single purpose last mile fibre over the next three years, and then discovers that it needs middle mile capacity, edge facility connections, and AI tolerant routing on top, will end up rebuilding parts of what it just finished. That is the worst kind of outcome. It is expensive, it is slow, and it is politically damaging because it looks like waste even when it is the predictable result of an underspecified plan.
Designing the network now for the AI era is not an act of speculation. The traffic patterns are already visible. The data centre footprint is already growing. The public sector workloads that will run on top of this infrastructure, from emergency response to permitting to citizen services, are already being prototyped. The question is whether the delivery side of BEAD can move at the same pace as the strategic conversation. Right now, my honest view is that it cannot. Not without changes.
What I Would Like to See Next
First, NTIA should issue not just permission but a template. A short, clear set of conditions under which a state can apply remaining funds to integrated infrastructure, with a small number of required deliverables that prove the institutional readiness is in place. A geospatial inventory. A documented multi agency governance body. A defined process for evaluating multi purpose builds. States that have this can move quickly. States that do not have it can use a portion of the funds to build it, with a clear deadline.
Second, Congress should be honest about the timeline implications. A fabric approach is the right approach, but it is not faster than a single purpose build in the first year. It is faster over five years, and dramatically faster over ten. Lawmakers who want the headline of households connected this year will not get a useful answer from a fabric program until later. The political incentive needs to be set up so that states are rewarded for doing the harder, longer, more durable work rather than the quicker, narrower one.
Third, the program needs to take delivery capacity seriously as a national constraint. The skilled workforce required to build the kind of infrastructure Bathras describes does not exist at scale today. Some of the remaining funds, if they are to be spent on anything other than direct deployment, should go into building that workforce. Apprenticeships. Cross training between transportation and broadband engineering. Programme management capability inside state broadband offices. Without this, every plan will outrun the people available to deliver it.
Closing Thought
Eric Bathras has done the field a service by naming the next step clearly. The digital fabric framing captures something real, and the work in Maryland shows that it can be more than a slogan. My contribution to the conversation is simply this. The framing only delivers if the program around it changes too. We do not need a louder vision. We need a delivery model that can actually carry the vision into the ground. Until we have that, fiber to fabric will read as a smart phrase rather than a built outcome, and the country will spend the next several years explaining why an idea everyone agreed with did not quite happen.
Maryland has shown what readiness looks like. The next move is for federal policy, state governance, and the broader delivery community to meet that readiness halfway, in writing, in funding, and in how we measure what counts as success.