davidremland.ai AI Systems Integrator · Commercial Real Estate
← Recent work
AI Systems Integration · Recent work · Build-to-resell underwriting

A spec-home model that says noto 100 of 101 listings.

Buy a lot or teardown, build new, resell. Every active West Palm Beach listing ran through the same cost stack, priced against closed new-construction sales in its own subdivision. One cleared the screen.

101
active lot and teardown listings screened (MLS, July 2026)
36
had enough nearby sales to support a resale price
1
cleared the margin and land-cost tests
The deal that passed

330 Marlborough Place, Prospect Park South

$1.45M teardown → 3,767 SF new home
33405 · 1952 house, 1,635 SF · 58 × 140 ft lot · 18 days on market
PASS
$5.78M
resale at $1,536/SF
$2.61M
pretax profit
45%
margin on sale
82%
return on cost

Resale is set by two closed new builds in the same subdivision: 265 Marlborough Road at $1,618/SF and 3218 Olive Avenue at $1,454/SF. Breakeven is $798/SF, so the price could fall by almost half before the deal loses money.

Why location decides it

New-build resale prices swing 2.5× inside one city

The 22 new homes (built 2022 or later, 3,500 SF and up) that closed in West Palm Beach in the six months before the screen. Median $1,146/SF. A zip-code average would have credited premium prices to streets that have never sold a new home.

Highlighted: the two Prospect Park South sales that price the subject. Source: Beaches MLS closed sales, exported July 20, 2026.

The screen

Top of the ranking, and why the rest fail

AddressSubdivisionAskResale $/SFProfitMarginResult

Same assumptions on every row: $238/SF all-in vertical cost, 24-month hold, 10% contingency, 5% commission plus 1% closing, all equity. Pass = 30%+ margin with land under 30% of the sale price.

How it was built

One person, one afternoon per refresh

Pull the listings

Every active lot and teardown in the city from the MLS export, duplicates merged (330 Marlborough was listed twice).

Price the resale

Closed new-build comps by subdivision first; asking prices with a haircut only where no sales exist.

Run the cost stack

Land, demolition, site, construction, soft costs, contingency and selling costs for all 101 rows in a live Excel model.

Rank and stress

Margin, return on cost, land share and breakeven $/SF, with sensitivity grids on resale price and build cost.