ROM hard-cost values from LV ROM Engine v2.1.3 client estimates: Beverly issued 7/22/26; HVN Round 2 awarded-4 portfolio re-run 7/12/26 ($57.04M across 325 units, $287/SF avg); Troost and Farralone issued 7/30/26. All figures are rough-order-of-magnitude benchmarks from LV’s committed-bid history — not GMPs or contractual bids. “LIHTC Award” denotes HVN’s CTCAC 4% tax-credit award on the project — not an LV contract award; LV has issued GC bids on all four. LV’s primary targets are Dickens, Moorpark and Riverton, which cluster in the Valley near LV’s active jobs; Federal (West LA) sits outside that cluster and is a lower-priority pursuit. All four are 100% affordable, Type IIIA, non-prevailing-wage basis, bonded.

Current Contract = Original Contract + Net Change Orders ($0.41M across the book). Backlog = Current Contract − Billed to Date. Francis and Riverton are now carried at contract value in the KPI workbook (the June pack showed them as pending buyout); Lexington and Acama have not completed buyout; Scott and Klump are pending capitalization and carry mobilization entries only. % Billed is billed-to-date over current contract; % Complete is cost-to-date over current budget; Over / (Under) is the difference in points — positive means billed ahead of cost. Ramsgate is the one to watch at -4.7 points under-billed. Job-end dates are the model completion months; Crenshaw’s KPI record still shows July 2026 and its schedule shows substantial completion December 1.

Source: 07.2026 Consolidated Financials (KPI). Francis and Riverton submitted their first pay applications in July ($932K and $355K of recognized revenue). On Francis $839K of the $852K billed is the bond line, billed at 100%; the $314K CBI premium (invoiced August 4, bond effective August 10) and the $366K cash due to Krueger when the bond line funds both post in August, so most of the job’s $910K gross profit is timing. Riverton’s $400K is bonding ($212K billed against a $104K budget), demolition ($74K, fully costed), the elevator deposit and mobilization. That is why both flip from a loss in June to a profit here; the job-level workbook carries a deferred-revenue adjustment on each job (Francis +$80K, Whipple ($93K), Nelrose ($79K), Riverton ($45K)) that bridges billings-based margin to the recognized figures shown. Lexington and Acama show only mobilization overhead; Scott and Klump have de minimis entries and no job-to-date history yet.

Budget = total job budget per each job's Buyout Report; Committed = executed subcontracts and POs; % Bought Out = Committed ÷ Budget, as reported in each workbook. Whipple and Califa each added ~5 points of buyout during July. Ramsgate and Riverton report in the richer PM-report format (invoiced, cost-to-complete, projected final cost); standardizing all jobs on one template is recommended so this table rolls up mechanically. Crenshaw (wrap-up) and Nelrose are excluded pending current reports.

Sources: 07.2026 Consolidated Financials (STAFFING) and the Base Case model (Actuals + Personnel, Jul 24 roster). “People on Job” counts everyone who charged time to that job in July; individuals appear on multiple jobs (61 job-assignments across the book), so the column intentionally does not total to unique headcount (21). Average: booked payroll (5000-1000) Jan through Jul was $144.6K / $150.7K / $178.9K / $187.7K / $295.9K / $198.6K / $215.9K, averaging $196.0K/month; the average smooths biweekly pay-period timing (May carried three pay periods). Projection: the Base Case Personnel roster has Francis and Riverton-Denny mobilizing August 2026, Lexington September 2026 (moved from August at the July close) and Acama November 2026 (moved from September on August 23); from September the roster carries 25 people at ~$286K/month on a salary basis ($268K field / $18K corporate; the roster shows 29 lines because Grady LaKamp, Joe LaKamp, Pedro Rosalas and Greg Smith are each split between field and office/G&A allocations), rising to ~$297K and 26 people in December when the Francis APM joins, then settling at ~$276K and 25 people from January 2027 after a planned year-end departure. No headcount reduction occurred at any point.

Revenue ramp tracks the book mobilizing: $1.45M in January to $7.03M in July, of which $1.25M is the first Francis and Riverton pay applications in a single month. February's negative gross profit reflects trade-cost timing on early-cycle billing; May net income was compressed by the third biweekly payroll period and a $74K warranty adjustment. April GC overhead ($73K) looks far lower than the other months (roughly $260K to $438K) because of a one-time $251K credit: payroll for self-performed trade work was reclassified out of overhead and into job costs, where it belongs as a cost of the work. Before that credit, April overhead was ~$325K, right in line with March and June; nothing was cut. May carries a smaller $82K credit of the same kind, masked by its third biweekly payroll period. July gross profit of $1.77M is the item to read carefully: $1.15M of it is margin on trade billing (a 18% spread on $6.41M of trade billing against a book that normally runs near zero on trades), driven by the first Francis and Riverton pay applications. $839K of the Francis billing and $212K of Riverton’s is the bond line, recorded as revenue in July; the surety premiums ($314K Francis, $104K Riverton already booked) and the ~$694K due to Krueger under the bonding program post in August, where the models already carry them, so on a matched basis the two bond lines are roughly a wash and about $0.95M of July’s gross profit is timing. The job-level KPI workbook already carries an $80K deferred-revenue adjustment on Francis and ($45K) on Riverton.

Recurring SP revenue is fee income (Edgemont management fee, Ramsgate and Crenshaw development fees, ~$34K/month). June includes a $299K non-cash revenue item and a $54K interest/tax/amortization true-up, which turn the month positive; July carries the first $44K of 12th & Fir developer fee. The underlying run rate is roughly ($120K) to ($160K) a month — the burn the Uplifters and BD fee programs are designed to cover. Note the July close removed the Six Peak-to-LV overhead allocation from the projection entirely: it was never booked in either ledger, so these entity figures are the true split.

Short answer: five jobs in the book are HVN work already committed to LV — Califa and Whipple building today, and Lexington, Riverton and Acama contracted and awaiting their LIHTC closings. Those three are committed work, not pursuits: the GMPs are set, pre-construction buyout is running, and the only thing between them and ground-break is the tax-credit closing — the developer’s and syndicator’s process, not ours. The six HVN projects on the Pipeline tab are pursuits, not awards: the “LIHTC Award” label there refers to HVN’s CTCAC 4% tax-credit award on the deal, not an LV contract award. The table below is the definitive list the pack was missing — every job in the book against the same five-stage ladder.

Status ladder built from the Active Projects tab and the Base Case model Projects sheet (contract value, start month, bonding). Five jobs in the book are HVN: Califa and Whipple are under construction and bonded through the ICON National / Travelers program; Lexington, Riverton and Acama are committed to LV and waiting on LIHTC closings — $48.6M of contracted GMP that converts to billing as those closings occur. The six HVN projects on the Pipeline tab (Dickens, Moorpark, Riverton 5642, Federal, Troost, Farralone — $135.4M of ROM value across 740 units) carry no signed LV contract and no LV award; four of them hold CTCAC 4% LIHTC awards granted to HVN as developer, against which LV has issued GC bids. This column will be carried forward on every pack so the signed / funded / on-site distinction is explicit rather than inferred.

The honest read: every near-term start is gated on a third-party LIHTC or lender closing, not on LV readiness. Lexington, Riverton and Acama are committed HVN work with GMPs set and buyout underway — the wait is the tax-credit closing calendar, not the contracts. Buyout, staffing and pre-construction are ahead of the closings on all four. The exposure that is genuinely ours to manage is Riverton’s $7.41M of unbought scope, which is priced against a GMP that was set before the current market.

Thirteen weeks from Monday August 3 runs to November 1, so the window is the August, September and October model months. Figures are the Base Case cash roll-forward (LV ending cash + Six Peak ending cash), re-opened on August 28 at the actual July 31 bank balances of $405,563 — $506,455 below the $912,018 the June pack projected for that date. The August column carries the whole July 31 working-capital unwind in one month: $9,463,439 of receivables collected, ($8,434,511) of payables paid, $1,028,928 net. The LV-to-Six-Peak liquidating distribution the June build counted as cash is a non-cash item and was taken out of the cash build on August 29. That is an assumption about timing, not a forecast of August collections, and it is what lifts the balance from $405,563 to $1,400,677. Six Peak on its own goes negative in August and stays there, because the July close removed the Six Peak-to-LV overhead allocation from the projection; on the entity roll-forward LV funds it. The “Total Ending Cash” defect noted in the June pack was corrected in the August 23 model rebuild; the summary line now ties to the entity roll-forward.

Retention from the Base Case Retention tab (5% of LV fee revenue, release at each job’s completion month); the tab was not rolled forward at the July close, so the held-today column is still the June figure. The July draw sheets show gross retention held by owners of $1.61M across the seven billing jobs (Crenshaw $469K, Ramsgate $591K, Califa $230K, Whipple $228K, Nelrose $81K, Francis and Riverton under $10K) — the gross figure includes retention LV in turn holds on its subcontractors. WIP by job — contract value, billed to date, backlog and percent complete — is on the Active Projects tab; margin by job on a YTD and job-to-date basis is on Financial Results; committed-versus-budget cost certainty is on Project Buyout. The over/under-billing view by job — percent billed against percent complete — is now on the Active Projects tab from the July KPI workbook; Ramsgate is 4.7 points under-billed and Crenshaw 1.5 points, everything else is within a point. A full earned-revenue WIP schedule with dollar over/under-billings is the remaining piece; the KPI workbook already holds every input it needs.

As promised in the June pack, the comparison is now against the real anchor: the November 2025 approved budget for 2026, carried on the Budget tab of every model with a Budget vs Actual tab that computes the variance line by line for every closed month. Variance is actual less budget; favorable is positive. The former “Original Budget” (a frozen reforecast seeded with Q1 actuals) is retained as Reforecast Q1 2026 and no longer drives this exhibit.

Three things to read from this. First, the GC is ahead of budget on margin, not volume. Revenue is ($833,257) behind the November plan year-to-date, but gross profit is $1,077,597 ahead and overhead $223,147 under, so GC net income is $1,300,744 ahead of a $752,933 budget. July alone is $1,213,201 ahead, and most of that is the Francis bond line billed before its August costs. Second, Six Peak is $267,250 behind budget, almost entirely on expenses (($521,065)): the budget assumed $410,781 of Six Peak payroll would be allocated to LV over seven months and it never was, and consultants and professional fees run well over plan; Six Peak revenue is $201,109 ahead on the June non-cash item and the 12th & Fir fee. Third, consolidated is $1,033,494 ahead of budget on the entity ledgers — but the models, which carry July job margin on the curve, show consolidated $242,368 behind. The gap is the July true-up question on the 7/31 Financials tab. The budget itself carries a known $7,042 internal inconsistency on the allocation line, documented and deliberately not restated.

Legal cost for the review was billed to 2859 Francis. This item is carried as a portfolio-level entry on the risk register below, rated amber until the paystub-and-affidavit procedure is confirmed live across every active subcontract rather than newly executed ones.

Sourced from the Base Case model — the Projects tab flags each job as “GC + Developer” or “3rd Party GC”, and the Six Peak revenue block carries each fee line by month. Califa, Whipple, Nelrose, Lexington, Riverton and Acama are third-party GC contracts with no Six Peak development position. Important framing on the LIHTC lines: these four developments — Francis, Deharo, Reseda and 3rd St — require more development work than anything else in the portfolio and pay no developer fee today; the modeled distributions do not begin until 2029. On top of that, under the current rate and tax-credit environment the Steyn cost of capital absorbs most of the early LIHTC distributions, so through FY2033 the net contribution of the strategy is close to zero — the meaningful development-fee income lands after 2033. This is the single biggest reason the Six Peak entity P&L reads negative while the underlying book grows. The life-of-deal figures above are gross Six Peak share, before that Steyn absorption. This table becomes a standing exhibit and will carry actual-versus-projected columns once Deharo and any other omitted positions are confirmed.

Operational Metrics today carries one aggregate safety box. Below is the first pass at the job-level register, rated only where there is data behind the rating. It will be maintained monthly and expanded — schedule variance in particular needs a baseline-versus-current schedule feed that is not yet in the reporting pack.

Ratings: G = no action required; A = watch item with a named owner; R = escalated, on the board agenda. Safety ratings reflect the monthly third-party inspections by Safety Compliance Company at each active site — the latest filed (Ramsgate, June 16) covered 52 items with 39 satisfactory and zero unsatisfactory findings, so every active site rates green on safety — July inspection reports were not in Box when this pack was built; jobs not yet on site are not rated. Cost/buyout is rated from each job’s Buyout Report; margin from the Financial Results tab; financing from the closing status above. Schedule is rated conservatively from percent-complete against modeled job end, because a proper baseline-versus-current schedule feed is not yet part of the monthly reporting — instrumenting that is the main gap in this register.