Construction-to-Permanent Financing for the Sunset Park Rebuild: Mechanism, Jumbo Threshold, and Florida Lender Options
The question
Verbatim: "How do construction-to-permanent loans work for a custom new build at $1.5-1.8M all-in, for a household with $315k gross income, ~$150k liquid capital, and no existing mortgage — and what are the current Florida market rates and lender options?"
Context: this is the general-mechanics-and-rates half of home-rebuild-2027 Critical Open Question #4. The flood-zone-specific half was already answered on 2026-08-04 and is deliberately not restated here.
⚠️ TWO PREMISES IN THE QUESTION ARE WRONG. Corrected before any math was run.
1. Liquid capital is not ~$150k — it is ~$777.5k deployable. The live Monarch pull of 2026-08-01 shows cash $290,150 + taxable brokerage $487,381 = $777,531 deployable without penalty, plus $179,155 locked retirement and $36,998 spendable Mercury business cash. [[2026-08-01-capital-deployment-brainstorm]] Likely provenance of the $150k: the project README's parents'-home down-payment gift line item ("Founder gifts ~$150k"), which is Project B, not the founder's liquid position. [[readme]] All analysis below uses $777,531, and I flag where the answer would change under the $150k reading (it changes completely: at $150k liquid this project is not financeable at all).
2. There IS an existing mortgage. $415,622 outstanding on 1212 S Suffolk Dr as of the same 2026-08-01 pull, against a $1,047,100 Zillow estimate — roughly $631k equity. [[2026-08-01-capital-deployment-brainstorm]] This is load-bearing: the construction lender needs first lien on the same parcel, so that balance gets paid off out of loan proceeds or out of cash before a dollar of construction is funded.
The $315k combined gross ($190k phData base + $125k wife W-2) is correct per [[readme]].
What we already know (from the vault)
- The build is construction cost on already-owned land, not a purchase. ~3,000-3,300 sf, $1.5-1.8M all-in; the $1.8M gate pencils to ~$545/sf on owned land against a $770/sf turnkey comp three streets over, implying ~$2.5M finished value. [[2026-06-11-comp-3-streets-over]] So loan-to-value has enormous headroom — collateral was never going to be the constraint.
- Income, not capital, is the binding constraint. $26,250/mo gross → 28% front-end $7,350, 36% back-end $9,450, current PITI $4,094. The README's own conclusion: "Practical mortgage ceiling on income alone: ~$1.0-1.2M." [[readme]] The independent affordability pass reached the same place from the other direction — comfortable at a ~$1.45-1.5M home (PITI ~$7,500), stretch to ~$2.0M only "with heavy cash down." [[2026-06-04-home-affordability-build-vs-buy]]
- Underwriting-history risk is already flagged. The phData W-2 started 2026-05-26; the vault warns most lenders want a 2-year history for variable comp (bonus, cert escalators), and README Open Question #3 is specifically about bonus inclusion. [[2026-06-04-home-affordability-build-vs-buy]] [[readme]]
- The flood-zone financing layer is already resolved. Builder's-risk-with-flood at first draw → NFIP/private at walled-and-roofed; design-stage EC supporting the stem-wall draw, as-built EC at closeout; 4-6 phase draws with independent inspection. Teardown is classified as new construction, sidestepping the 50% substantial-improvement rule. [[2026-08-04-flood-zone-ae-construction-loan-requirements]]
- The 2027-01 start target in the backlog entry conflicts with the live decision posture. As of 2026-06-30 the standing call was build-ready but hold groundbreaking until the family-gravity signal (BIL's post-residency location) resolves in ~12-18 months; 2026-07-13 confirmed STAY PUT and killed the sell-and-rent branch. [[2026-07-13-rent-buy-build-interim-decision]] [[2026-07-13-rent-affordability-sell-scenario-financials]]
- Temp housing during the build is priced at ~$8k/mo, not the README's $4k/mo. The 7/13 dialogue repriced the Tampa rental band upward. [[2026-07-13-rent-buy-build-interim-decision]] The README's $40-50k temp-housing line is therefore understated by roughly 2x.
What the web says
- 2026 conforming limit, Hillsborough County: $832,750 (one-unit, national baseline — Tampa Bay carries no high-cost designation; Monroe County at $990,150 is Florida's only above-baseline county). This figure is corroborated across multiple secondary lender sources. Sourcing caveat: FHFA's own conforming-loan-limit page, which I did fetch, confirms Florida is not a statutorily designated high-cost area (only AK, HI, GU, USVI are) but does not state the 2026 dollar figure inline — it links out to a county XLSX I did not open. So treat $832,750 as high-confidence secondary, not primary-verified.
- Everything about this project is therefore jumbo. Any plausible loan here lands ~$1.1-1.6M, comfortably above $832,750. That resolves the "Fannie/Freddie product vs portfolio lender" fork in the backlog note: the GSE single-close construction products are simply out of range. The choice is jumbo-portfolio or jumbo-portfolio.
- Dated Florida rate quote (the only hard one I obtained). MIDFLORIDA Credit Union's construction-to-permanent program page, rates as of 2026-08-03: 1-year ARM 5.875% / 6.881% APR; 30-year fixed 7.000% / 7.129% APR. Interest-only during construction. Down payment as low as 10% (fixed) / 20% (ARM); min FICO 700 ARM, 720 fixed; ARM adjusts annually at 1-yr T-bill + 2.75% margin, 6% lifetime cap. Critical caveat: those posted rates illustrate a $300,000 example loan. A $1.1-1.6M jumbo will price differently, almost certainly higher.
- Soft market band (low confidence, not fetched): a search summary reported construction-to-perm rates of 6-8% during construction, roughly 1.0-2.0% above conventional, as-of 2026-03-02. I did not open that source. Treat as directional only; five-month-old rate color is close to worthless for a Q4 2026 application anyway.
- Named Florida-active lenders (aggregator page, last updated 2026-04-27): Goldwater Bank — one-time-close C2P, portfolio/jumbo to $1.5M, 720 min FICO, 12-month construction term, interest-only, two draws monthly. TD Bank — OTC C2P to $1.5M at 20% down, to $3M at 30% down, 720-740 FICO desired, rate locked before construction begins, 10% of construction costs held in reserve, no prepayment penalty. MIDFLORIDA CU — OTC C2P, jumbo available, 10% down, interest-only, builder review + inspection before each draw. Seacoast Bank — OTC C2P, fixed and adjustable, local underwriting and closing decision-makers. Space Coast CU — caps at $650k, too small here. Separately, BuildBuyRefi (broker/aggregator, all 50 states incl. FL) markets a One-Time-Close Jumbo to $4.5M: 720 min FICO, fixed interest-only period up to 18 months as a portfolio ARM, auto-converts at completion with no separate refinance, in-house draw management with inspection before each disbursement, contingency reserve required (percentage not stated), primary residence only, no self-build or non-approved contractors.
- Single-close vs two-close, the real tradeoff. Single-close: one closing, one underwrite, one appraisal (subject-to-completion), one title policy; saves a reported $3,000-5,000 in duplicate closing costs; interest-only on the drawn balance only; converts automatically at completion. The cost is that the permanent rate is set at the initial closing — BuildBuyRefi's jumbo OTC explicitly notes the rate is fixed at initial closing and is not flexible, unlike their two-time-close and hybrid structures. Two-close: separate construction loan then a take-out refinance, two sets of costs, but you shop the perm rate at completion — at the price of being fully re-underwritten (income, credit, appraisal) 12-18 months later.
Convergences and contradictions
- Convergence, and it is the whole finding: every independent line of vault analysis (README DTI ceiling ~$1.0-1.2M; the 6/04 affordability pass at ~$1.45-1.5M comfortable) and the web's LTV/LTC mechanics agree that the down payment is not the problem and the monthly payment is. He has roughly six times the down payment the question implies and cannot service the loan the build would need.
- Contradiction with the backlog framing. The backlog entry says the vault has "zero coverage of the actual financing mechanism." That is not true — [[2026-08-04-flood-zone-ae-construction-loan-requirements]] already covers draw schedules, inspections, insurance sequencing, and EC delivery points for exactly this project. What was genuinely missing is the sizing and product-class layer, which is what this brief adds.
- Contradiction with the stated timeline. The backlog's "2027-01 start" predates the 6/30 hold and the 7/13 STAY PUT. A construction-loan application supporting a January 2027 groundbreaking would need to be in motion in Q4 2026, which runs directly against the live posture of holding until the family-gravity signal lands. Either the target moved or the decision did; the vault does not record a reversal.
Synthesis for RDCO
The mechanism, concretely, for this specific parcel. A single-close construction-to-permanent loan on 1212 S Suffolk Dr is one note, closed once, that funds in stages and then modifies into a 30-year mortgage without a second closing. At the initial closing the lender takes first lien on the land, which means the existing $415,622 mortgage is retired out of loan proceeds — that is the first "draw" in economic terms even though no construction has happened. The already-owned land counts as the equity contribution: with an as-is land value plausibly in the high six figures and total cost basis (land + ~$1.6M construction) around $2.4-2.6M against a ~$2.5M appraised-on-completion value, the loan-to-cost and loan-to-value tests are trivially satisfied. During construction he pays interest-only on the drawn balance, which ramps from near-zero to the full note rate as draws release — Goldwater runs two draws monthly, most programs run 4-6 phase draws, each gated on an independent inspection, with the stem-wall draw being the one where the design-stage elevation certificate has to land per the 8/04 brief. Some programs capitalize an interest reserve into the loan so the borrower writes no check during construction; the jumbo OTC program I looked at does the opposite and makes the borrower pay. Conversion triggers on completion — Certificate of Occupancy, final inspection and final draw, as-built EC, clear lien waivers under Florida's construction-lien regime — and on a single-close that conversion is a modification, not a refinance: no new note, no new title policy, and on some programs a re-verification of employment. The rate, however, is set at the initial closing, 12-18 months before he moves in. Float-down availability is program-specific and unverified here; it is the single most valuable term to ask each lender about, because on a $1.2M note a 75bp float-down is roughly $600/mo for thirty years.
Sizing, and why the answer is uncomfortable. Run it as sources and uses. Sources: $777,531 deployable liquid, plus the README's projected $150-200k of additional savings by Q1 2027, plus a permanent loan the vault's own DTI work caps near $1.0-1.2M. Uses: $415,622 mortgage payoff, $1.5-1.8M construction, and temp housing that the 7/13 repricing puts at **$8k/mo × 12-15 months ≈ $96-120k**, not the README's $40-50k. Take the favorable end — $1.5M construction, $1.1M perm loan, $175k of new savings — and sources are ~$2.05M against uses of ~$2.03M. It closes, with essentially zero margin and no reserve left standing, which directly violates the README's own "maintain 6-month expense reserve OUTSIDE the project budget" mitigation. Take the $1.8M end and it is short by roughly $400k. And the DTI ceiling itself is probably generous: $1.1M at 7% is $7,318/mo P&I before taxes and insurance, and this is a Zone AE parcel on a ~$2.5M finished basis in the post-Helene Florida insurance market — the escrow line is not a rounding error, and I did not price it here (see [[2026-07-22-florida-ae-zone-insurance-repricing-tampa]]). Layer on that jumbo construction lenders typically want months of post-closing PITI reserves, and TD's posted requirement to hold 10% of construction costs in reserve — $160k on a $1.6M build — and the tightness gets worse, not better. Under the question's erroneous $150k-liquid premise, none of this is financeable and the correct answer would have been "don't." Under the true $777.5k it is financeable at the bottom of the budget band and not at the top.
Which product class, and what to actually ask. GSE single-close is out on size alone, so the fork is between a balance-sheet portfolio lender (Seacoast's local underwriting is the archetype — a Tampa-decisioned bank can look at a nonconforming 68-ft lot and a 7-month W-2 with judgment rather than an overlay matrix) and a national jumbo OTC program (BuildBuyRefi-style, $4.5M ceiling, 18-month interest-only, mechanically clean but rule-bound). Goldwater's $1.5M portfolio cap is uncomfortably close to the likely loan; TD's $3M-at-30%-down tier is the one mainstream product on this list with real headroom, and 30% down is not a problem for him. A third category the vault has never considered and that fits his balance sheet unusually well: private-client / pledged-asset structures, where $487k of taxable brokerage is pledged rather than liquidated, avoiding the capital-gains hit of selling to raise cash down. I could not source a specific Florida institution's private-client construction program in this run, so that stays a category to investigate, not a recommendation. Two structural points that should shape the conversation regardless of lender: single-close is probably right for him specifically, not for the closing-cost savings but because two-close means being fully re-underwritten in late 2028 on variable comp the vault already flags as an underwriting risk — one underwrite at 7-9 months of phData W-2 history is a smaller exposure than two. And the loan is not the lever that makes an $1.8M build work; income is. The honest options for closing the gap are a smaller build, more cash, higher qualifying income, or waiting — not a cleverer loan product.
Why this is in the vault
Answers the sizing-and-product-class half of home-rebuild-2027 Critical Open Question #4 ("construction-to-perm vs separate construction loan"), the half [[2026-08-04-flood-zone-ae-construction-loan-requirements]] explicitly scoped out, and corrects two wrong premises ($150k liquid, no mortgage) that were about to propagate into the Q1 2027 go/no-go gate and the still-idle mortgage-broker consult.
Open follow-ups
- What the all-in monthly escrow actually is on a ~$2.5M finished-value Zone AE home in Hillsborough County — property tax at the new construction basis (does any Save Our Homes portability survive a teardown-rebuild on the same parcel?) plus wind plus flood. This is the number that sets the true DTI ceiling, and every affordability figure in the vault currently assumes an escrow line inherited from the 1963 house.
- Whether float-down provisions exist on jumbo single-close construction-to-perm products in the current Florida market, and what they typically cost in rate or fee. This determines how much the "rate locked 12-18 months early" downside of single-close actually stings.
- Whether pledged-asset / securities-backed structures are available for owner-occupied custom construction from institutions operating in Florida, and how a pledge is treated in DTI versus a cash down payment. Directly relevant given a $487k taxable brokerage and a founder who would rather not realize gains.
- How jumbo construction lenders treat bonus and cert-escalator income at under 12 months of W-2 tenure — portfolio lenders may have materially different overlays than the 2-year-history norm the vault currently assumes, and the difference is worth several hundred thousand of qualifying loan.
- Whether land equity on an owned free-and-clear-after-payoff lot counts toward the down-payment requirement at appraised as-is value or at original cost basis. Lender practice differs and it swings the required cash contribution substantially.
Related
- [[readme]] — home-rebuild-2027 project plan; Critical Open Question #4, the DTI ceiling, and the $150k parents'-gift line this brief traces the bad premise to
- [[2026-08-04-flood-zone-ae-construction-loan-requirements]] — companion brief; the flood-zone layer of the same question (draw gating, EC timing, insurance sequencing)
- [[2026-08-01-capital-deployment-brainstorm]] — the verified 2026-08-01 Monarch balances used to correct both wrong premises
- [[2026-06-04-home-affordability-build-vs-buy]] — independent affordability pass reaching the same income-bound conclusion
- [[2026-06-11-comp-3-streets-over]] — the $770/sf turnkey comp establishing ~$2.5M finished value
- [[2026-07-13-rent-buy-build-interim-decision]] — STAY PUT decision trail and the ~$8k/mo temp-rental repricing
- [[2026-07-13-rent-affordability-sell-scenario-financials]] — the final STAY PUT call that closed the sell-and-rent branch
- [[2026-07-22-florida-ae-zone-insurance-repricing-tampa]] — the insurance-cost layer this brief defers to for the escrow line
- [[zoning-flood-zone-findings]] — 68-ft nonconforming lot width, the open appraisal/LTV thread that interacts with jumbo underwriting
- [[milestones]] — Q1 2027 go/no-go decision gate this brief feeds
Sources
Research caps: 3 QMD queries, 3 WebSearch, 4 WebFetch — one over the 3-fetch cap, because the FHFA fetch returned no usable figure and a substitute was needed.
Vault:
rdco-vault/01-projects/home-rebuild-2027/README.mdrdco-vault/01-projects/home-rebuild-2027/milestones.mdrdco-vault/01-projects/home-rebuild-2027/zoning-flood-zone-findings.mdrdco-vault/01-projects/home-rebuild-2027/2026-06-11-comp-3-streets-over.mdrdco-vault/01-projects/home-rebuild-2027/2026-07-13-rent-buy-build-interim-decision.mdrdco-vault/01-projects/home-rebuild-2027/2026-07-13-rent-affordability-sell-scenario-financials.mdrdco-vault/04-finance/2026-08-01-capital-deployment-brainstorm.mdrdco-vault/04-finance/2026-06-04-home-affordability-build-vs-buy.mdrdco-vault/06-reference/research/2026-08-04-flood-zone-ae-construction-loan-requirements.mdrdco-vault/06-reference/research/2026-07-22-florida-ae-zone-insurance-repricing-tampa.md
Web (fetched and read in this run):
- FHFA, Conforming Loan Limit Values — https://www.fhfa.gov/data/conforming-loan-limit (confirms FL carries no statutory high-cost designation; does not state the 2026 dollar figure inline)
- MIDFLORIDA Credit Union, Construction-to-Permanent Loans — https://www.midflorida.com/personal/home-loans/mortgages/construction-to-permanent (rates as of 2026-08-03: 1-yr ARM 5.875% / 6.881% APR; 30-yr fixed 7.000% / 7.129% APR; example loan $300,000)
- BuildBuyRefi, One-Time Close Construction Loan Programs — https://www.buildbuyrefi.com/otc-one-time-close-construction-loan-programs-for-fha-usda-and-va (Jumbo OTC to $4.5M, 720 FICO, 18-mo interest-only portfolio ARM, rate set at initial closing)
- Biglaw Investor, "4 Best Construction Loans in Florida" — https://www.biglawinvestor.com/marketplace/construction-loans/florida/ (page last updated 2026-04-27; Goldwater, TD Bank, Space Coast CU, MIDFLORIDA, Seacoast terms)
Web (search-summary only, NOT fetched — lower confidence):
- 2026 Hillsborough County conforming limit $832,750 / Monroe $990,150 — corroborated across mintrates.com, truebluelending.net, and rocketmortgage.com search results; not primary-verified
- Construction-to-perm rate band 6-8%, ~1.0-2.0% above conventional, as-of 2026-03-02 — locallifehomes.com search summary; directional only
Unverified in this run and explicitly not treated as fact: as-is land value for 1212 S Suffolk Dr; the property-tax and insurance escrow line on a ~$2.5M finished-value Zone AE home; Florida retainage and lien-waiver draw practice (stated as general practice, not sourced); post-closing reserve requirements on jumbo construction loans beyond TD's posted 10%-of-construction-cost figure.