Four data center proposals across Virginia and North Carolina, each at a different stage of community response. Click any pin to read its story. Select two to compare side by side.
A structured guide for local communities evaluating a proposed data center. Drawn from case findings across NC and VA, it helps communities decide whether to approve, condition, or pause a proposal.
Data center expansion is critical for U.S. AI competitiveness. But in 2025, two-thirds of proposed projects faced delays or cancellations due to local opposition over electricity costs, water use, and grid reliability (AP News, 2025). This website explores that tension and proposes policy solutions.
It is designed for policymakers, community leaders, and stakeholders trying to understand how to balance accelerating data center deployment with the needs of local communities.
North Carolina and Virginia sit at the epicenter of this conflict. Virginia is an established data center hub, with Northern Virginia hosting some of the world's largest facilities. North Carolina is an emerging frontier, with companies seeking abundant land, water, and electricity at lower costs than Virginia. Both states have experienced the full spectrum of outcomes: approvals, delays, cancellations, and intense community pushback.
In 2025 alone, over 25 planned data center projects in North Carolina and Virginia were cancelled or indefinitely delayed following community opposition. At the same time, electricity costs in parts of Virginia are projected to rise 25% or more by 2030 due to data center growth. The case studies chosen from each state, Tarboro and Balico, show what happens when national urgency collides with local concerns.
Cases and policies in these two states could be extrapolated nationwide to build better policies around data center development.
AI tools (primarily Claude) were used to help code the interactive tools and design the site's layout and styling.
Data centers are the physical infrastructure that stores, processes, and transmits the data underlying everything from cloud storage to artificial intelligence. Technology companies have poured record investment into new construction: U.S. data center construction spending reached approximately $40 billion at a seasonally adjusted annual rate in June 2025, a 30% increase from the previous year, following a 50% surge in 2024.[1]
The primary driver is artificial intelligence. Training and operating large language models requires extraordinary concentrations of computing power, which must be physically housed in data centers. As AI models grow in size and complexity, computing demand increases exponentially rather than linearly, placing pressure on U.S. energy systems and host communities.
In 2023, U.S. data centers consumed approximately 176 TWh of electricity, equivalent to the annual usage of 16.3 million households.[2] Their share of national electricity consumption is projected to grow from 4.4% in 2023 to between 6.7% and 12% by 2028. At the same time, nearly 3,000 new data centers are planned or under construction, roughly doubling existing capacity.[3]
From a federal perspective, data centers are strategic infrastructure. Any bottleneck in this chain (regulatory delays, energy shortages, or local opposition) could be detrimental to U.S. national security. This framing is heavily influenced by competition with China. If the U.S. does not build enough data centers and the energy systems to power them, it could fall behind in the AI race. In 2024, China added approximately 429 GW of new power generation capacity, compared to just 51 GW added in the United States.[4] OpenAI has warned that energy constraints directly threaten U.S. leadership in AI, calling for rapid expansion of domestic energy capacity.
The 2025 executive order “Accelerating Federal Permitting of Data Center Infrastructure” directs agencies to expedite permitting and environmental review for projects deemed critical to national security.[5] The Department of Energy has also proposed expanding the Federal Energy Regulatory Commission's authority to streamline interstate grid access for large data center loads.[12]
By 2028, individual AI training runs could require up to 1 GW of power in a single location, roughly equivalent to the output of a large nuclear reactor. Global AI data center power demand could reach 327 GW by 2030, compared to total global data center capacity of just 88 GW in 2022.[6] RAND Corporation, 2025
At the local level, data centers are not seen as strategic infrastructure but as large industrial projects with significant resource demands. As development spreads beyond established hubs in Northern Virginia and Texas, communities have raised concerns about electricity costs, water usage, noise, land use, and environmental impacts.
By mid-2025, approximately two-thirds of tracked data center projects valued at $98 billion across 11 states had been delayed or stalled due to community and state-level opposition.[7] Several counties in North Carolina have enacted temporary moratoriums on new data center development.[8]
This creates a direct conflict between national priorities for rapid expansion and local resistance driven by infrastructure constraints and perceived unequal distribution of costs and benefits. The intensity of opposition varies by context: established data center hubs often have more experience, institutional capacity, and negotiating leverage, while emerging markets face greater uncertainty and pushback.
The U.S. electrical grid cannot absorb projected data center growth without significant expansion. Current data center capacity is approximately 51 GW, while an additional 44 GW will be required by 2028. However, only about 25 GW is expected to come online within that timeframe, leaving a gap of roughly 19 GW.[9]
The interconnection process is a major bottleneck. In PJM, the largest U.S. grid operator, interconnection timelines now exceed eight years, with similar delays of four to seven years common in Virginia. These delays reflect both the scale of demand and the limitations of aging infrastructure.[9]
Research suggests that relatively small operational adjustments could significantly increase grid capacity. Reducing data center load during peak periods by just 0.25% of the time could allow approximately 76 GW of additional demand to be accommodated without equivalent new generation.[9] Financial Times / Duke University, 2025
These supply constraints translate directly into rising costs for ratepayers. Data centers push up electricity bills through two primary mechanisms. First, utilities often socialize the cost of new infrastructure, such as substations and transmission upgrades, across all ratepayers, even when those investments are driven by a single large customer.[10]
Second, increased demand raises market prices when supply cannot keep pace. Modeling suggests that electricity prices in parts of Virginia could increase by 25% or more by 2030 due in part to data center growth, with national increases around 8%.[11] To keep up, utilities have also extended the operation of fossil fuel plants or added new gas generation, creating tension with companies' renewable energy commitments and further complicating the cost picture for host communities.
Federal, state, and local policy currently pulls in different directions on data center development. Federal action has broadly accelerated permitting and grid interconnection to match the pace of AI investment, while North Carolina and Virginia communities are increasingly conditioning, limiting, or rejecting projects through moratoriums, zoning denials, and rate-allocation challenges. The table below summarizes the major policies shaping this landscape and the direction each one pushes.
| Policy / Action | Direction | Status / Notes |
|---|---|---|
| Federal | ||
| EO: Accelerating Federal Permitting of Data Center Infrastructure Directs agencies to expedite permitting and environmental review for projects deemed critical to national security. | Supports | Active - issued July 2025 |
| DOE / FERC interconnection streamlining DOE has directed FERC to expand authority to accelerate interstate grid access for large data center loads. | Supports | Proposed / in development |
| Federal energy efficiency guidance for data centers Guidance on PUE and water use targets exists but is voluntary and not tied to enforcement. | Limits | Voluntary, unenforced |
| North Carolina | ||
| County-level moratoriums on new data center permits Several NC counties have enacted temporary moratoriums while impacts are studied. | Limits | Active in multiple counties (2025-26) |
| Local zoning denials / Special Use Permit precedent Tarboro denial (Sept 2025) established a site-specific basis for rejecting projects in industrially-zoned land adjacent to residential areas. | Limits | Case-by-case, under appeal |
| Virginia | ||
| JLARC review of data center impacts Joint Legislative Audit and Review Commission findings on rate impacts, grid strain, and local fiscal tradeoffs from hyperscale growth. | Conditions | Informing state policy debate |
| SCC rate-allocation rulings / Dominion ratemaking State Corporation Commission and Dominion proceedings on whether hyperscale-driven infrastructure costs should be recovered from large-load customers rather than socialized. | Conditions | Ongoing |
| Local zoning actions (Pittsylvania, Prince William, Loudoun) Project withdrawals, conditional approvals, and rezoning debates shaped by organized community opposition. | Limits | Active across multiple counties |
| Other / Comparative | ||
| ERCOT “connect and manage” approach (TX) State-level model that allows large loads to interconnect more quickly under managed conditions. | Supports | Cited as model in federal proposals |
| State Public Utility Commission challenges to cost allocation Several state PUCs are reviewing whether infrastructure costs driven by single large customers should be socialized across ratepayers. | Conditions | Ongoing in multiple states |
| State-level data center policies (e.g., Maine) Emerging state frameworks that condition or limit new data center development. | Limits | Reference / comparative |
Two data center proposals, one in North Carolina, one in Virginia. Both were denied or withdrawn following community opposition, making them the clearest illustrations of what drives backlash and what communities can do about it.
Energy Storage Solutions, LLC · Edgecombe County, NC · Permit denied September 2025
Energy Storage Solutions (ESS) proposed a 300 MW data center and energy storage system on approximately 52 acres of Heavy Industrial-zoned land on the south side of Anaconda Road in Tarboro, NC. The project also involved construction of a dedicated substation and solar generation capacity. The developer marketed the facility as “net zero” and “EMP hardened,” though these claims were contested during hearings.
Tarboro is a small town of roughly 10,000 in Edgecombe County - one of North Carolina's most economically distressed counties, with high unemployment and a legacy industrial base. The site was zoned Heavy Industrial, suggesting it should in principle be suitable for large-scale industrial development. However, the parcel was adjacent to residential properties, creating an immediate conflict between the zoning designation and the lived experience of nearby homeowners.
Notably, the same developer received approval for a much larger $19.2B, 900 MW campus just miles away in the same county at Kingsboro - demonstrating that the Tarboro denial was not a blanket rejection of data center development in the region, but a site-specific and process-specific outcome.
Promises: 500 employees and $11M per year in tax revenue. The developer also claimed net-zero operations and cited the EMP hardening as a national security benefit.
Concerns raised by adjacent property owners with legal standing: Air quality impacts from the natural gas infrastructure, noise pollution from cooling systems, light pollution, water demand, detrimental effect on adjacent residential property values, and incompatibility with the character of the surrounding area. Nine adjacent property owners established legal standing to participate as parties in the evidentiary hearing.
The Council voted 7-1 that the application was not compliant with findings of fact on four grounds: that the development would be detrimental to adjacent properties; that it would endanger public health or safety; that it would injure the value of adjoining property; and that it would not be in harmony with the area in which it is located. The permit was subsequently denied 7-1.
The procedural record is also notable. The Mayor recused himself from the hearing, and nine adjacent landowners established legal standing as parties, an unusually robust participation structure for a local zoning proceeding, suggesting the community was organized and legally prepared well before the hearing date.
The Town of Tarboro denied the Special Use Permit in September 2025. Energy Storage Solutions appealed the decision to the Edgecombe County Superior Court.[15] The appeal is pending as of April 2026. The developer has not publicly withdrawn the project, distinguishing this outcome from Balico in Virginia.
The Tarboro case illustrates that adjacency to residential areas is a decisive liability even in Industrial-zoned locations. The developer's net-zero and EMP-hardened claims did not overcome concrete proximity concerns from neighbors who could demonstrate special damages. The organized legal participation of nine adjacent owners - and the town attorney's structuring of the evidentiary hearing - gave community concerns institutional weight that general public comment would not have. The contrast with the nearby Kingsboro approval (same developer, same county, but no opposition) suggests site selection and community relationship matter more than project scale or developer identity.
Balico · Pittsylvania County, VA · Proposal withdrawn April 2025
Balico proposed one of the largest data center developments in Virginia history: a 3,500 MW hyperscale campus across 2,200 acres in rural Pittsylvania County, near the town of Chatham. Critically, the proposal included an on-site natural gas power plant as part of the facility infrastructure - not a grid interconnection, but dedicated fossil fuel generation built specifically to power the data center.
Pittsylvania County is rural, agriculturally rooted, and geographically distant from Virginia's established data center corridor in Northern Virginia (Loudoun County). Unlike Loudoun, which has decades of experience negotiating with hyperscale developers, Pittsylvania had no institutional precedent for evaluating proposals of this scale. The proposed 2,200-acre footprint represents a fundamental land use transformation for a rural county, not an incremental expansion of existing industrial zoning.
The county is also outside the PJM grid's most congested zones, meaning it lacked the established interconnection infrastructure that data center developers typically prefer - which may explain why an on-site gas plant was included in the proposal.
Promises: 390 jobs between the data center and the on-site power plant, and projected tax revenue of $50-184M annually - a wide range that itself became a source of community skepticism about the reliability of economic projections.
Concerns: Opposition coalesced around rural and cultural identity (residents framed the proposal as incompatible with Pittsylvania's agricultural character) as well as environmental concerns about the on-site gas plant's air quality and emissions impacts. The wide variance in the tax revenue estimate ($50M to $184M) undermined the economic case, with opponents questioning whether promised benefits would materialize or accrue locally.
Unlike Tarboro, where the community organized through an evidentiary hearing with formal legal standing, the Balico opposition appears to have operated primarily through public comment, county board pressure, and sustained civic engagement over several months. The developer's response (submitting a revised proposal) suggests it believed the opposition could be partially accommodated. The County Board of Supervisors nonetheless denied the revised proposal, and after months of postponed votes, Balico withdrew the project entirely in April 2025.
The withdrawal, rather than an appeal, distinguishes this outcome from Tarboro. Balico's decision to abandon rather than litigate suggests the company concluded that community opposition in Pittsylvania was structural, not procedural. No revised proposal would overcome the fundamental mismatch between the project's scale and the county's character.
Balico formally withdrew the proposal in April 2025 after the County Board denied the revised application. No appeal was filed. The project is cancelled. The site remains undeveloped agricultural land.[14]
The Balico case demonstrates that hyperscale projects face qualitatively different community dynamics than mid-scale ones. At 2,200 acres and 3,500 MW, the proposal would have transformed the rural landscape. The inclusion of an on-site gas plant was a significant strategic error: it converted an economic development question into an environmental and energy policy question, broadening the opposition coalition beyond adjacent property owners to environmental advocates and energy policy critics. The $50-184M tax revenue range suggests the economic case was not well-developed, undermining developer credibility at a critical early stage.
A complete analysis requires engaging the strongest case for data center expansion, not just the community opposition. Developers and federal advocates advance several arguments that deserve serious consideration.
| Dimension | Tarboro, NC (ESS) | Pittsylvania, VA (Balico) |
|---|---|---|
| Scale | 300 MW, 52 acres (mid-scale) | 3,500 MW, 2,200 acres (hyperscale) |
| Energy source | Natural gas + solar; own substation | On-site natural gas power plant |
| Site context | Heavy Industrial zoning, adjacent to residential | Rural agricultural land, no precedent for scale |
| Developer promises | 500 jobs, $11M/yr taxes (specific but contested) | 390 jobs, $50-184M taxes (wide range undermined credibility) |
| Primary concerns | Adjacency, property values, air/noise/light/water | Rural character, environment, economic skepticism |
| Opposition structure | Formal evidentiary hearing; 9 adjacent owners with legal standing | Sustained civic engagement; public comment and board pressure |
| Developer response | No documented revision; contested hearing | Revised proposal submitted; ultimately withdrew |
| Outcome | Permit denied 7-1; developer appealing in court | Revised proposal denied; developer withdrew entirely |
| Post-denial path | Legal appeal to Superior Court (outcome pending) | Project cancelled; no appeal filed |
| What drove the result | Residential adjacency + organized legal participation | Scale mismatch + gas plant + credibility gap on economics |
Despite their geographic and institutional differences, both cases share a common structure: a developer underestimated the importance of community fit and overestimated the persuasive power of economic projections. In Tarboro, the site selection placed an industrial facility directly adjacent to residential properties in a small town without the institutional capacity to negotiate detailed mitigation conditions. The result was a rejection rather than a negotiated approval. In Pittsylvania, the scale of the proposal and the inclusion of on-site fossil generation transformed what could have been an economic development conversation into a much broader values conflict about land, environment, and rural identity.
Both cases also suggest that the timing of community engagement matters. In Tarboro, nine adjacent landowners organized sufficiently to achieve legal standing in a formal evidentiary proceeding, a significant institutional achievement for a small community. In Pittsylvania, opposition sustained itself over months of postponed votes until the developer concluded withdrawal was preferable to further attrition. Neither community had a formal Community Benefits Agreement framework available to them; both relied on the blunt instrument of outright denial rather than conditioned approval.
The contrast with the nearby Kingsboro approval (same developer, same county as Tarboro, no opposition) and the Microsoft Catawba approvals (organized developer engagement, Duke Energy infrastructure deal, positive community framing) suggests that the determining variables are less about data centers as a category and more about site selection, developer credibility, and the presence or absence of binding commitments made before community opposition organizes.
Using the five-factor framework from the Decision Framework tab, how each case scores in hindsight, based on documented conditions at the time of the hearing. Higher scores mean more pressure against approval.
Result: Permit denied 7–1 · Outcome: Conditional/Pause zone
| Factor | Rating | Evidence |
|---|---|---|
| Resource constraints | Moderate | Dedicated substation planned; gas-plus-solar raises grid concerns; water demand contested in hearing |
| Economic impact | Mixed | 500 jobs and $11M/yr taxes cited; specific figures, but neighboring property owners disputed net benefit given residential impacts |
| Community sentiment | Actively opposed | Nine adjacent owners achieved legal standing; formal evidentiary hearing; 7–1 council vote against |
| Land use fit | Poor fit | Heavy Industrial zoning on paper; direct adjacency to residential properties undermined compatibility argument |
| Developer commitments | Partial / vague | No documented revisions or binding mitigation conditions; “net zero” and EMP claims were contested, not codified |
Result: Withdrawn after denial · Outcome: Pause & organize zone
| Factor | Rating | Evidence |
|---|---|---|
| Resource constraints | Severe | On-site natural gas power plant proposed for 3,500 MW; 2,200 acres of new land clearing; no existing grid infrastructure at that scale |
| Economic impact | Weak payoff | 390 permanent jobs for a facility of this scale; tax estimate ranged $50–184M — a 4× spread that communities read as unreliable |
| Community sentiment | Actively opposed | Months of sustained civic mobilization; repeated vote postponements; county board voted against revised proposal |
| Land use fit | Poor fit | Rural agricultural land with no precedent for hyperscale industrial; triggered a values conflict about rural identity, not just zoning |
| Developer commitments | Minimal / nonbinding | Revised proposal submitted but contained no documented binding mitigations; developer ultimately withdrew rather than negotiate further |
A 20-point gap separates Tarboro (63) from Balico (83). Tarboro was recoverable — a different site, better buffering, or earlier binding commitments could have shifted the outcome. Balico was not: the scale mismatch and on-site fossil plant created pressure that no set of conditions could fully absorb. The framework predicts both denials, but for different reasons and with different implications for what a developer could have done differently. Try the to test your own scenario.
Local legitimacy has increasingly become the bottleneck for data center construction. A sustainable path forward requires the federal government to condition grid access, standardize operational practices, and coordinate across levels of government.
The Policy: Each state establishes a permanent Data Center Authority (DCA) to manage siting, environmental review, and community standards for large facilities. The authority integrates staff from energy, water, land use, and community development. It develops binding standards for siting distances from residential areas, water usage, operational hours, and local benefit agreements. Federal permits are contingent on state-level approval from the DCA.
The Problem It Solves: Currently, data center siting decisions are fragmented across county and local planning boards with inconsistent standards and expertise. Tarboro was denied on residential adjacency; Kingsboro (same developer, same county) was approved with no opposition. This inconsistency creates unpredictability for developers and fear for communities. Rural communities like Pittsylvania lack institutional capacity to evaluate hyperscale projects, leading to blanket opposition rather than negotiated approval.
Why It Works: A state authority centralizes expertise, standardizes evaluation, and creates predictable rules. It prevents race-to-the-bottom competition where counties undercut each other for data center tax revenue by waiving environmental protections. It also prevents not-in-my-backyard (NIMBY) blocking where one locality's denial simply pushes the project to a less-organized neighbor. With clear siting standards (minimum distances from homes, water availability, grid proximity, environmental constraints), communities know what is acceptable and what is not. The authority becomes a convening body that negotiates Community Benefits Agreements (CBAs), ensures equitable impact distribution, and enforces operational standards post-approval.
For communities: A state authority means your community is not alone in evaluating hyperscale development. You benefit from state-level expertise, standardized environmental review, and guaranteed enforcement of conditions. You have more leverage, since developers cannot shop around for the most permissive county.
The Policy: The Department of Energy establishes federal standards for (1) water usage effectiveness (WUE), power usage effectiveness (PUE), and cooling methods; (2) demand-response capability requiring data centers to curtail non-critical loads during peak demand or grid emergencies. Projects meeting these standards receive priority grid interconnection and federal approval. FERC and state Public Utility Commissions enforce curtailment requirements and penalize non-compliance.
The Problem It Solves: Two interrelated crises: (1) Resource unpredictability: a single large data center consumes 60,000+ liters of water daily and 69 GWh of electricity annually, with no enforceable limits. Communities fear infrastructure overwhelm, driving moratoriums. (2) Grid inflexibility: interconnection timelines reach 4-7 years because utilities cannot accommodate always-on loads. An additional 44 GW of data center demand is needed by 2028, but only 25 GW is projected to come online, leaving a 19 GW shortfall. Inflexible loads worsen this bottleneck.
Why It Works: Clear standards address both problems simultaneously. For water/energy: measured efficiency expectations and predictable limits give communities concrete information to evaluate impacts. For grid: most computing is not time-critical. AI training and bulk processing can shift to off-peak hours with minimal operational cost. Research shows curtailing load just 0.25% of the time could allow the grid to accommodate an additional 76 GW of demand without new generation capacity. Data centers become grid resources rather than grid liabilities.
For communities: Standardized limits protect local infrastructure. Demand-response capability means more stable electricity prices (data centers help balance peak load rather than worsen it). Flexible data centers reduce pressure for emergency grid upgrades that would otherwise be socialized across all ratepayers. Communities approve projects knowing they have predictable, enforceable operational boundaries.
The Policy: The federal government provides state and local jurisdictions with technical assistance, impact modeling tools, and historical data on data center effects. Communities use this to model electricity price effects, water stress, and grid reliability risks for proposed projects.
The Problem It Solves: Federal expedited permitting (issued July 2025) has not accelerated projects because it ignores the real bottleneck: local legitimacy. Communities block projects even when federally approved because they lack the technical capacity to evaluate impacts.
Why It Works: Communities are not philosophically opposed to data centers; they are opposed to being blindsided by unpredictable costs and impacts. Federal technical support transforms local decision-making from reactive (opposing what they don't understand) to proactive (negotiating from a position of knowledge). When state and local officials can model impacts, they can set terms. When they set terms, projects move forward because communities feel heard and protected.
For communities: Federal-provided data and modeling tools enable informed decisions, shift power dynamics toward communities, and create accountability mechanisms that build trust.
Federal officials and industry partners present a compelling case: “Slower, more community-responsive permitting will delay AI infrastructure and compromise U.S. competitiveness. China added 429 GW of new generation capacity in 2024 while the U.S. added only 51 GW, and every month of delay costs the U.S. leadership in AI.”
This concern is legitimate. The U.S. is in a genuine competition for AI dominance, and energy constraints are real bottlenecks. These policies address why fast-track permitting is failing in the first place.
Federal expedited permitting has existed since July 2025, yet two-thirds of data center projects remain stalled. The problem is not approval speed but community legitimacy: federal agencies can issue permits, but communities continue to block projects or threaten appeals, moratoriums, and legislative action even when permits are granted, because they do not trust the cost-benefit calculation. Forcing faster permitting without addressing this trust deficit will not accelerate deployment; it will trigger more community opposition, appeals, and backlash.
These three reforms reverse that dynamic. By conditioning grid access on flexibility, standardizing efficiency, and supporting local planning, they give communities the tools to evaluate and negotiate projects while giving developers clearer rules to follow. Projects that meet these standards face less opposition because they arrive with guardrails communities helped define. The U.S. gets the infrastructure it needs while communities gain stability, and deployment actually accelerates because the bottleneck shifts from federal approval to community acceptance, which happens faster when communities feel heard.
The Decision Framework you explored earlier helps communities evaluate specific projects. These policy reforms create the environment in which good decisions can be made. With a State Data Center Authority, Operational Efficiency & Grid Flexibility Standards, and Multilevel Coordination in place, a community evaluating a data center proposal will have:
When communities use the Decision Framework with these guardrails in place, they can figure out what works best for them with state and regulatory backing.
The cases in this tool show what happens when these structures are missing. Tarboro was a technically sound project with real economic benefits, but it failed because the community experienced the process as imposition rather than partnership. The developer lost the deal, federal agencies lost a project they considered strategic, and the grid lost 300 MW of planned capacity along with it.
Conversely, Northern Virginia's established data center hubs approve projects at higher rates, not because they love data centers but because they have institutional capacity, historical experience, and negotiated agreements.
The United States cannot sacrifice AI competitiveness for local opposition. It also cannot sacrifice local stability and trust for federal speed. Data center deployment is not a federal-versus-local question. It is a federal-and-local question.
A State Data Center Authority, Operational Efficiency & Grid Flexibility Standards, and Multilevel Coordination bridge that gap. They enable federal policy to move at the speed of AI growth while enabling communities to move at the pace of informed decision-making. Together, they transform the current bottleneck into a pipeline, and in doing so, they protect both national interests and community stability.
The path forward is not faster federal permitting. It is smarter governance: state-level siting authority that coordinates with federal requirements, operational standards that are clear and enforceable, and multilevel coordination that shares information and builds capacity. The path forward starts with policymaking that includes all stakeholders from the start.