The Count Is Rising. The Money Is Already Here.

The Asheville-Buncombe Continuum of Care took its 2026 point-in-time count on February 10. It counted 824 people as homeless on a single night: 334 unsheltered and 490 in shelter or transitional housing. The total is up about 9 percent from 755 the year before, and up from 527 in 2021. Not every year rose, though: the 2023 count, 573, was lower than 2022's 637.

After Helene, the U.S. Department of Housing and Urban Development (HUD) told city staff that people in FEMA-paid hotels could be included in the count. The federally reported 2025 figure of 2,303 is not comparable to the rest of the series, because it includes 1,548 people in those hotels. So 2,303 should not be set beside 824 or 755 as if it measured the same thing. The cleaner reading is to look at the other years: the count is still rising, with unsheltered counts well above where they sat before the pandemic.

One question in the count tests a common assumption. The 334 people found without shelter were asked where they last had housing. About 60 percent said Buncombe County, and about 74 percent said somewhere in North Carolina, Buncombe included. That does not fit the idea that most of them came from elsewhere to use services. The answers do not cover the 490 people in shelters and transitional housing.

As for behavioral health, the 2026 count's own detail tables record 91 adults self-reporting a mental health condition and 62 a substance use disorder. The numbers come from a single source and are almost certainly an undercount. Those figures do not describe a population that is mostly addicted or mostly mentally ill, though a visible chronic minority clearly struggles with both.

Beneath the count is a supply problem. The 2025 regional housing needs assessment estimates the area needs about 34,358 units over five years, including a rental shortfall of roughly 6,441 units in Asheville alone. These are planning estimates, so treat them as rough.

HUD's January 2024 national count put homelessness at its highest level since record-keeping began in 2007. Research points to housing cost. In 2020 the Government Accountability Office found that every $100 increase in median rent is associated with about a 9 percent rise in the homelessness rate, after taking wages, poverty, and unemployment into account. It added that it could not conclude housing costs were the primary factor, only that the two were statistically associated.

Locally the arithmetic is stark. By the National Low Income Housing Coalition's 2026 Out of Reach report, a worker needs about $35.29 an hour to afford a two-bedroom at fair-market rent here. Pay in many of the service jobs the region runs on sits well below half that. By the same report, the area is one of the least affordable metros in North Carolina relative to local wages. The county's median home price was about $446,000 in the first quarter of 2026.

A 2022 survey of regional employers put housing cost and availability at the top of their hiring problems. Meanwhile, as of February 2024, nearly 1,700 new hotel rooms were under construction or in the development pipeline. That tally included rooms that were not yet being built. It starts in February 2021, when the city lifted its hotel moratorium. A July 2026 CoStar report on the western North Carolina hotel market counted two hotels under construction in the Asheville area, with 208 rooms between them. There is little housing nearby that the people staffing the new hotels can afford. Homelessness here is not separate from the labor shortage. It is the same housing market problem, measured at the bottom rung. Housing supply affects both.

The national fight over how to respond has hardened into two sides, and each is right about something different.

Housing First is permanent housing that does not require sobriety or treatment first and comes with support. The case for it is strongest on housing itself. One review of randomized trials found participants roughly 2.5 times more likely than people in usual care to be stably housed. Houston's coordinated, data-driven system has cut regional homelessness by around 60 percent since 2011. The weakness is just as real: an apartment is not a clinical treatment. The same review rated all four of its trials at high risk of bias. It found no clear difference from usual care in substance use or mental health over about two years, and no clear sign of harm.

The case for Treatment First, built on campus models like San Antonio's Haven for Hope, points to genuine recovery numbers. Its weakness is methodological: those numbers come without a control group and count the people who complete the program. That way of counting builds in survivorship bias and leaves out everyone who could not meet the conditions to stay.

The critique of Housing First lands hardest where the model is weakest: the chronically unsheltered minority with active psychosis or addiction. It lands softest as a general theory, because it cannot explain why high-rent regions, not high-addiction regions, have the most homelessness. A May 2026 post from the University of Pennsylvania put it this way: the most visible homeless people are also among the least typical.

The fairest way to put the two sides together is a division of labor. Housing First would be the structure for the majority, with a treatment track that is genuinely funded and actually available for the minority the housing model does not reach. The trap most places fall into is funding neither at scale.

The cleanest evidence on cost is a randomized controlled trial. In Denver's Social Impact Bond trial, the Urban Institute followed 724 people from 2016 to 2020. The control group cost about $25,544 per person per year in public services such as jail, emergency rooms, detox, and shelter. Part of that is what Medicaid was billed, not what it paid. People offered supportive housing used about $18,668 in other public services, roughly $6,876 less. That group included the one in five who did not move in. The savings covered about half of the program's own cost per person.

Among people who moved in early enough to be followed for three years, 12 percent had died within that time. Of those still alive, 77 percent were in permanent housing.

The National Academies reviewed six studies with a comparison group, two of them randomized. Of the six, three showed net savings, up to about $33,500 per person a year. The other three showed a net cost, from about $250 to about $3,100 per person a year. All of those amounts are in 2015 dollars. The review's own verdict: policymakers should not expect net savings. It added that some savings could show up in specific studies, such as those focused only on persistently heavy users of emergency medical services. Our verdict for high-need groups like these: supportive housing can pay back part of its cost, as it did in Denver, and in some studies all of it. For the broader population, the cost question matters less than the supply question.

Responding to camps costs cities money too. A study of four cities, done for HUD and HHS, found that their responses to encampments cost $1,672 to $6,208 per unsheltered person in fiscal year 2019. Across the four, outreach was the largest cost and clearing camps the second. The totals did not include fire department or emergency medical responses, or police responding to the areas around the camps.

The clearest local win is Compass Point Village, a $17.5 million conversion of a former Tunnel Road motel into 85 permanent supportive housing units. It opened in September 2023. In June 2026, very few of its rooms were empty. The building also became the sharpest local version of the national disorder debate, with police calls and neighborhood pushback that we cover in The 92 Percent Who Stayed. The case management inside the building worked for the people living there.

The clearest disappointment is a 2022 plan. The city hired the National Alliance to End Homelessness for a roadmap to halve unsheltered homelessness. Dogwood Health Trust paid the $72,974 cost. The roadmap's 116 recommendations went largely unimplemented, and the count rose. Chief among them were a system overhaul and a single low-barrier shelter. That low-barrier shelter is the piece the city has repeatedly named as missing and repeatedly failed to build.

The opioid-funded Community Paramedics Response Team gives a local example of the gap between a referral and a home. In 2023 it logged 986 overdose responses and referred 88 people to housing. It landed nine. The team's figures do not say why so few gained housing, but the gap is the kind a housing shortage would produce.

One recent change cuts the other way, toward coordination. In 2024 the area's loose advisory committee was replaced by a formally chartered Continuum of Care, now several hundred members strong. In June 2025 it adopted its first three-year strategic plan, with goals to reduce unsheltered homelessness, increase exits to permanent housing, and expand prevention and rapid rehousing. The coordinating structure that was missing for years is now in place.

Two policy developments frame what Buncombe can do. Both are volatile and worth watching closely.

At the federal level, Executive Order 14321 was issued in July 2025. It directs HUD and HHS to end support, in their homelessness programs, for Housing First policies that "deprioritize accountability and fail to promote treatment, recovery, and self-sufficiency," in the order's words. The order says this directive applies to the extent the law allows. HUD wrote the order's principles into a November 2025 funding notice that would have capped permanent housing at a minority of grant dollars. The notice sent Continuums of Care nationwide scrambling. The order also tells the Justice Department, HHS, HUD and the Transportation Department to look at whether their discretionary grants can give priority to grantees in states and cities that enforce bans on open illicit drug use, urban camping and loitering, or urban squatting, among other tests.

In December 2025 a federal trial court in Rhode Island issued orders blocking that November notice. The trial court refused to lift those orders, and HUD appealed. On April 1, 2026 the federal appeals court refused to pause them while that appeal went on. HUD said it remains committed to its approach. The fiscal year 2026 notice that followed was a $4.04 billion competition whose first tier covered projects worth only 60 percent of renewal needs. That notice also gave points for providing treatment and for requiring people to take part in services. On June 29, 2026 a court vacated the 2025 notices outright. On August 7, 2026 the same court set the fiscal year 2026 notice aside in its entirety. It held that HUD could not create its $1.3 billion set-aside without first taking public comment.

HUD appealed. The challengers later said they were prepared to give up their public-comment claim. At their request, the appeals court vacated the August ruling on September 24, 2026. That left the fiscal year 2026 notice in force. Applications under it closed on September 30. The cases are back with the trial court. Nothing about the merits is settled, so the threat to existing supportive-housing funding is still real.

Locally, the Continuum of Care's own application would end two of Homeward Bound's supportive-housing grants, worth about $397,000 a year. That money would fund new transitional housing instead. The application's stated basis for the cuts is project scores, the priorities the Continuum's board approved for the first tier, and HUD's priorities. The full federal turn against Housing First, and the roughly $1.4 million of Homeward Bound's supportive-housing money at stake, is the subject of Strings Attached.

At the state level, the camping-ban approach has resurfaced. House Bill 781, the "Unauthorized Public Camping and Sleeping" bill, passed the North Carolina House in 2025 but stalled in the Senate. On June 10, 2026, its core language was revived in a Senate committee's rewrite of House Bill 437, originally a measure to establish drug-free homeless-service zones. The rewritten bill passed the Senate in June and cleared its final House vote 73 to 40 on June 30, 2026. Governor Stein vetoed it on July 8, 2026. Republican leaders have pledged to try an override, so the outcome is not settled.

Starting June 30, 2027, the bill would bar local governments from allowing regular camping or sleeping on public property. It makes exceptions, among them a state-certified sanctioned site: land the local government owns within its boundaries and designates by majority vote for a continuous period of up to a year.

The site the bill would allow is a place to camp or sleep outdoors, not a shelter, so it is a different thing from the low-barrier shelter the city has tried and failed to build. The bill provides no state money for such a site. Designating one would not lift the bar on camping on any other public property. The provider community can push on what local government builds alongside the site. We take the camping ban apart, and price what obeying it would cost Buncombe, in A Ban Is Not a Bed.

What is unusual right now is that several large pots of money are already committed. The task is matching each kind of dollar to the job it is suited for: one-time money to build, recurring money to sustain treatment, and a small prevention pot to keep people housed. The task also includes a hard stop on spending more to criminalize than it costs to house.

In ranked order, the actions worth considering: