Cultiva / Association 3 — Counter-Proposal

For: Association 3 cultivation facility On the table: Green Jungle offer dated 12.03.2026, total €82 583.92 From: Gozo Max Date: 11.04.2026

In one paragraph

The Green Jungle offer is €82 584 on paper but needs another ~€10–12 k in parts that aren't quoted to become an operating facility, and the flowering HVAC is sized for ~50 % of the actual heat load which means the crop cooks during the first July heatwave. I'm proposing a fully-specced counter-build at ~€88 k that fixes the HVAC sizing, uses Samsung LM301H chips with a real warranty instead of anonymous Chinese boards, replaces the cloud-dependent Growlink + Autogrow dual control islands with one locally-running Home Assistant stack, and includes everything the first build is missing: CO₂ hardware, carbon scrubbers, backup power, security cameras, door interlocks, drip lines, tank temperature management, and sign-off by a licensed electrician. Same order of money, an operable facility that survives the first summer, and no cloud accounts that can be turned off without warning.

Why I'm writing this

You have an offer from Green Jungle for a commercial cultivation facility at your Cultiva lease site. I've read the equipment spec (6 tables dated 24.03.2026) and the pricing offer (€82 583.92 dated 12.03.2026) in detail. This document walks through:

1. What's technically wrong with the Green Jungle offer — not hand-wave, specific numbers
2. What's missing from it that you'll discover on install day
3. My counter-proposal, same order of budget, fixing the issues
4. What you give up by choosing the DIY approach — the honest tradeoffs

You can verify every number here. Every claim in this pitch traces back to a specific teardown document in this project. The goal is not to sell — it's to make sure you're making the decision with full information.

The three technical findings that matter most

Finding 1 — Flowering HVAC is sized for half the real load

The Green Jungle offer puts one 15 kW ducted unit (UCHA60-TDC) in each flowering room at €3 200 each. Let's do the math:

• Lighting heat (12 × 800 W + 18 × 120 W): 11.76 kW sensible heat, continuous during light-on
• Dehumidifier reject heat (276 L/day × 0.7 kWh/L continuous): ~8 kW sensible, unless the dehu is ducted outdoors (not mentioned in the offer)
• Ambient / solar / motors: ~2 kW
• Total real sensible cooling demand per flowering room: ~20 kW

The 15 kW UCHA60 unit then gets derated for Malta peak ambient (38 °C outdoor → ~25 % capacity loss on cheap Chinese commercial equipment) and SHR (consumer equipment wastes capacity on latent load the dehu already removed). Effective delivered sensible capacity ≈ 9.6 kW at peak conditions.

9.6 kW of cooling against 20 kW of heat. The room hits 30–35 °C canopy temperature during the first heatwave. Terpenes volatilise off, buds cook, yield and quality crash.

Worst case: you lose a full crop (€15–30 k) every time Malta has a real summer day. Best case (if the dehu is ducted outside, which the offer doesn't confirm) the demand drops to ~13.7 kW and the unit is only 30 % under — still too small but survivable at reduced yield.

My counter-proposal: 6 × Midea Xtreme Save 18 (5.2 kW nameplate) wall splits per flowering room. Total 31 kW nameplate → ~20 kW effective at Malta peak → sized correctly, and any 2 of the 6 units can fail simultaneously and the crop still survives (N+2 redundancy). Cost: €2 800 per flowering room vs vendor's €3 200 (cheaper AND properly sized).

When a commercial HVAC tech in Malta takes 4 weeks to source a replacement UCHA60 part, you've lost a crop. When one of six Midea units fails, I swap in a €500 replacement in 3 hours. This is the single most important argument in the whole counter-proposal.

Finding 2 — €399 per "800 W LED" is below Samsung chip cost

The Green Jungle offer line for flowering top lighting is €399 per 800 W fixture, 24 total. This price point is below the cost of a real Samsung LM301H board — a genuine 800 W equivalent board with Samsung chips from Kingbrite (the gray-zone AliExpress supplier that the serious DIY community trusts) sells for €450–600.

At €399/fixture the Green Jungle product is almost certainly a Chinese LM281B+ relabel or an Epistar clone. There's no brand, no chip name, no efficacy rating, no DLC listing, no warranty term in the offer. For €9 576 you're getting 24 unbranded boards with no paper trail.

My counter-proposal: for roughly the same per-fixture price (€520 in bulk), Mars Hydro FC-E8000 or Spider Farmer G8600 from their EU warehouse in Germany. These are:

• Verified to use Samsung LM301B or LM301H chips (independent testing on YouTube, HLG Photon reports)
• 2-year manufacturer warranty, honoured in Europe
• 5–10 day DDP shipping to Poland with VAT included (no customs surprise)
• DLC-listed (useful for future commercial rebate programs and insurance)
• IP65 rated (won't fail in 60 % RH flowering rooms)
• Published spectrum graph and PPF (µmol/s) measurements

Same money. Verified supply chain. The extra ~€2 500 across 24 fixtures is the cheapest insurance policy in the whole build.

Finding 3 — Automation is two cloud-dependent islands with no fallback

Green Jungle bundles €11 346 of Growlink equipment (one kit per flowering, veg, mothers, dry room) plus €4 989 of Autogrow IntelliDose dosing hardware. That's €16 335 of control gear. The problem is that Growlink and Autogrow are two separate vendors with two separate cloud services and no integration between them.

• If Growlink's cloud goes down, you lose alerting, logging, and remote monitoring across all 5 rooms. Local schedule execution continues — the lights still cycle — but a sensor fault at 3 AM goes unnoticed.
• If Autogrow's cloud goes down, same story for the fertigation side.
• If either company folds in the next 5 years (Growlink is a small US company, Autogrow is a mid-size NZ company), that half of the facility control turns into paperweight. Precedent: Revolv, Wink, Insteon, Automatic — all killed their clouds and bricked customer devices.
• You cannot see everything in one place — 5 separate Growlink dashboards + the IntelliDose web portal = 6 places to look for facility status.

My counter-proposal: Home Assistant on a Raspberry Pi 5 with UPS + NVMe + daily snapshots, driving Sonoff POW R3 / 4CH Pro / TH Elite / SNZB-02P devices flashed with Tasmota (no cloud), orchestrating all climate + lighting + fertigation + security from one local host. We keep the Autogrow IntelliDose because it's the one piece of the vendor build that's genuinely well-specced — we just bridge it into HA via Modbus instead of paying €208 for the Autogrow cloud bridge.

Cost: €5 416 for the full automation stack vs vendor's €11 346. Functionally equivalent, €5 930 cheaper, and it runs forever with no internet connection.

What's missing from the Green Jungle offer

These are line items that aren't in the €82 584 bid but that you'll need on day 1 of operation. I've put them in my BOM; Green Jungle left them out so the bid looks cheaper:

Real cost of operating the Green Jungle facility: ~€94 900, not €82 584. You're comparing the wrong number.

The honest cost comparison

Apples to apples: my proposal is ~€6 600 cheaper and covers everything.

But the real story is what you get for the money. Here's the same €88 k spent two different ways:

5-year total cost of ownership

Capex is a fifth of what you'll actually spend operating this facility over 5 years. Operating costs matter more than bid price.

~€23 500 saving over 5 years, concentrated in the two areas where vendor cut corners (cheap LED chips that fail, single-point-of-failure HVAC that can't be repaired fast).

What you give up by choosing my approach

I want to be explicit about the downsides. If any of these are dealbreakers, the vendor approach may be the right call for you — and I'd rather you know now than find out six weeks into an install.

1. Build timeline

• Green Jungle: quote says 6–8 weeks, probably realistic
• Mine: ~3 months, because AliExpress shipping is 14–21 days, automation commissioning + tuning is 80–120 hours, and there's more physical install due to 18 split ACs vs 5 ducted units

If you need flowers on the shelf before August, vendor wins on schedule.

2. Operational complexity

• Green Jungle: Growlink and Autogrow dashboards are designed for operators. Point and click.
• Mine: Home Assistant has a learning curve. I'll build read-only dashboards for day-to-day operation that hide the complexity, but if I become unavailable and you need to debug an ESPHome deployment, it's harder than calling Growlink support.

Mitigation: I document everything. I train your staff. I write runbooks for common failure modes. But there's real single-integrator dependency on whoever builds the HA stack.

3. Insurance unknown

• Green Jungle: commercial vendor-installed equipment fits standard insurance profile
• Mine: DIY-built control panels + AliExpress-sourced lighting is novel territory for Maltese cannabis insurers

Action required before committing: get insurance quotes on both builds. If DIY is uninsurable or materially more expensive to insure, the savings evaporate. I can help with the insurer conversation but the customer has to decide.

4. Insurance / licensing DLC requirement

If Malta licensing requires DLC-listed fixtures, the Mars Hydro / Spider Farmer EU warehouse route (both of which are DLC-listed in their production runs) works. But if a local interpretation rejects specific models, we have a problem.

5. Insurance-backed warranty on controls

Growlink and Autogrow have commercial insurance and support contracts. My DIY HA stack has none — if something breaks after I walk away, the remediation is "find another HA consultant" or "figure it out".

6. Airflow uniformity

Ducted HVAC gives ±1 °C canopy uniformity; 6 split AC units give ±2–3 °C. For a yield-maximising grow chasing absolute peak g/m², ducted is technically better. My approach trades ~5 % of yield uniformity for redundancy + cost + local control. Whether that's the right trade depends on your priority.

The three things I'm asking for

1. Read the comparison/cost_comparison.md and comparison/risk_comparison.md — those are the two documents where I show my work, line by line. If anything doesn't add up, tell me and I'll correct it.

2. Ask Green Jungle these 5 questions before signing anything, regardless of which approach you choose:

   • Is the UCHA60-TDC dehumidifier hot side ducted outside the room, or does it reject into the room?
   • What brand + chip + µmol/J rating + warranty term is the €399 "800 W top light"?
   • Where are the CO₂ tanks + regulators + solenoids in the €82 584 total — implicit or genuinely missing?
   • Is carbon filtration for odour control included anywhere, or do I buy it separately?
   • What is the spare-parts lead time for UCHA60-TDC in Malta when one fails mid-flower?

3. Get an insurance quote on both builds from a Maltese cannabis insurer before committing. This resolves the one genuine unknown in my proposal.

Next steps if we move forward

1. Insurance check (blocking)
2. Incoming electrical service capacity confirmation with landlord / EPD (blocking, applies to both builds)
3. Bulk quotes from Mars Hydro + Spider Farmer EU warehouse for 24-fixture order
4. Midea Xtreme Save availability confirmation from Rotenso PL for 16-unit order
5. F-gas installer quote from Galea Refrigeration / Cool Services / AC Gozo for 18-unit install
6. Logistics plan: consolidated shipment to Poland, van transport to Malta
7. Timeline: ~14–21 day shipping, ~3 week physical install, ~4 week automation commissioning + tuning = 12-week realistic build window
8. Staged payment aligned with delivery milestones

Further reading (inside this project)

• Full vendor teardown by subsystem: ../vendor_teardown/
• Full DIY proposal with BOM: ../diy_proposal/
• Cost comparison line-by-line: ../comparison/cost_comparison.md
• Risk heatmap: ../comparison/risk_comparison.md
• Honest tradeoffs: ../comparison/tradeoffs.md
• Research that informed the proposal (LED chips, AliExpress sourcing, Sonoff, Midea): ../../research/
• Source documents (Green Jungle offer): ../source/