May 2026
The Operations Brief
Ghibli Group
Three opportunities to as Ghibli scales through Hungary's most significant industrial buildout in a generation.
By Lightbloom · An AI Operating Partner
Potential savings: HUF 119M–214M per year from public sources
FY2022 to FY2024 · three years of exceptional growth
25.4B
HUF · FY2024 Revenue
Primary entity · CompanyWall confirmed
400+
Group FTE in 2025
Up from 85 in the primary entity in FY2022
Revenue doubled in two years. Headcount nearly quadrupled. The operational infrastructure running beneath that growth was built for a HUF 5B business, not a HUF 25B one.
Fig. 1 · Annual revenue · HUF billion · FY2022-FY2024
Ghibli Group is a 30-year-old Hungarian freight forwarder and warehouse operator that has repositioned itself as the logistics gateway between China and Central Europe. Revenue grew from HUF 11.9B in FY2022 to HUF 25.4B in FY2024 on the primary entity alone, a 114% increase in two years. Group-wide revenue including the absorbed Raktárlogisztika entity and the Cross Border Supply Chain Kft. tracks at HUF 27–30B. Szabó Zoltán has guided 30% further growth for 2025. The company processes over 55,000 customs declarations per day at Budapest Airport (37–38% of all cargo handled there) and is the default logistics operator for the Debrecen industrial cluster where BMW, CATL, and Semcorp are building.
Three key opportunities identified.
Assembled from public sources only: CompanyWall registry filings, Ghibli's website, press coverage including Világgazdaság and Transpack, the BSE BÉT50 profile, and the Atlatszo investigation into the military transport tender. Lightbloom has seen no internal data.
What is visible from the outside is typically 40–60% of what a full operational review surfaces once we have access to actual TMS data, functional headcount allocation, and the customs software operating costs. The opportunities below are conservative. The picture from inside is larger.
1
Opportunity One
·01 / 03
Finance that runs itself, across three entities
A multi-entity consolidation agent that eliminates the manual close process across three legal entities and seven sites, freeing 2–3 finance FTE and keeping both managing directors in the picture every morning.
Finding
Ghibli operates across three active legal entities: the primary Szállítmányozási Kft., the recently absorbed Raktárlogisztika operation, and the Cross Border Supply Chain Kft. that handles Chinese e-commerce flows through the CECZ shareholder structure. FY2024 payroll in the primary entity alone ran to approximately HUF 4.3B (294 FTE at HUF 1.22M per month). Group-level consolidated reporting, required both for the dual-shareholder governance structure and for any BSE Xtend preparation, is assembled manually with no evidence of an integrated group finance system in public sources. At HUF 25B in revenue and seven physical sites, the volume of transactions moving through that manual process is large enough that errors and delays have material consequences.
Three entities. Seven sites. Two managing directors. One manual consolidation process that runs as fast as the person doing it.
Our Solution
Lightbloom embeds a multi-entity finance consolidation agent: automated data ingestion from each entity's accounting system (NAV-compliant ERPs), intercompany elimination logic, and a unified management reporting layer. Szabó Zoltán and Wu Jiang open the morning to a dashboard showing group P&L, entity-level cash, receivables aging, and flagged anomalies. Built once. Running every day. The calculation is rule-based; the model explains the exceptions. If a BSE Xtend listing is still on the roadmap, this is the infrastructure it requires. The build timeline is 12 weeks; the first automated management pack replaces the manual version in week 9.
Estimated annual value
01 / 03
HUF 24M–43M
per year
2–3 finance/admin FTE freed from manual consolidation at HUF 1.0–1.2M/month average total cost x 12 months = HUF 24M–43M/year. Cash mechanism: attrition or redeployment as the agent absorbs workload. Plus avoided BSE Xtend preparation cost if listing is pursued.
2
Opportunity Two
·02 / 03
250 subcontractors. One rate card. No benchmark.
A procurement intelligence agent that reads every subcontractor invoice against live market benchmarks, surfacing the specific carrier relationships where Ghibli is paying above market and putting the data on the table before the next renegotiation.
Finding
Ghibli coordinates 250–300 subcontractor trucks for its domestic road freight operations. Each relationship involves rate agreements, performance tracking, and periodic renegotiation. At HUF 25.4B in FY2024 revenue, road freight is estimated at 40–45% of the total, with subcontractor cost running at 70–75% of that segment. That implies approximately HUF 7.1B–8.7B in annual subcontractor payments flowing through rate cards that were set at negotiation and reviewed opportunistically since. Without a system tracking actual rates paid per lane, per carrier, and per weight band against current market benchmarks in real time, the finance team has no visibility into where Ghibli is above market and no data to bring to a renegotiation. As volume grew 120% in two years, the subcontractor pool expanded faster than the ability to manage it analytically.
The rate was right when it was set. The market has moved since, lane by lane, without a system to track it.
Our Solution
Lightbloom builds a procurement intelligence agent: it ingests all subcontractor invoices from the TMS or ERP, classifies each payment by lane, load type, and carrier, and compares actual rates against benchmark data from Timocom spot market, Eurostat freight rate indices, and TransInfo for CEE corridors. The output is a ranked list of renegotiation opportunities, each formatted with the data Ghibli's team needs to open the conversation. The tool does not negotiate. Ghibli's team does. The shift is from relationship-based decisions to fact-based ones. At 1–1.5% improvement on a HUF 7.1B–8.7B subcontractor base, the recoverable range is HUF 71M–128M per year. The benchmark tools that produce this typically identify 2–4% in recoverable spend; 1–1.5% is applied here because Ghibli already manages rates actively.
Estimated annual value
02 / 03
HUF 71M–128M
per year
FY2024 revenue HUF 25.4B. Road freight estimated at 40–45% = HUF 10.2B–11.5B. Subcontractor cost at 70–75% of road freight = HUF 7.1B–8.6B. 1% improvement = HUF 71M–86M; 1.5% improvement = HUF 107M–129M. Range: HUF 71M–128M.
3
Opportunity Three
·03 / 03
Seven sites. One view. Currently assembled by hand.
A multi-site warehouse reporting agent that connects to each of Ghibli's seven locations, generates automated client KPI packs for BMW and CATL, and gives operations management a live cross-site utilisation dashboard instead of a weekly summary they wait for.
Finding
Ghibli operates 120,000 m² of warehouse space across seven locations: Budapest Csepel Free Port, Budapest Airport cargo area, Vecsés CTPark, Debrecen Airport, the new 22,000 m² Debrecen South Industrial Park (opened April 2025 adjacent to CATL and BMW), Pécs, and Szeged. Each site runs its own warehouse management operations and produces its own client KPI reports. The company website promises online inventory access and transparent reporting as client-facing features, implying reporting is a manually produced deliverable rather than a live automated feed. Multi-site warehouse businesses at this scale typically employ 2–4 people across the network whose primary job is aggregating site-level data into client reports. At Ghibli's headcount growth rate (85 to 294 FTE in the primary entity in two years), the reporting function has grown as headcount, not as automation.
Seven locations. Clients like BMW and CATL. The inventory picture is assembled by hand, once a week.
Our Solution
Lightbloom builds a multi-site warehouse reporting agent: connectors to the WMS at each of the seven sites, normalisation into a unified schema, and automated client report generation on a schedule or on demand. Internal management sees a live cross-site utilisation dashboard showing which sites are at capacity, where space is available, and where throughput is lagging. BMW and CATL procurement teams, who expect real-time inventory visibility as a baseline for automotive supply chain relationships, get a portal view of their own inventory position without an email cycle. The agent means site N+1 (Debrecen is adding 80,000 m² more by 2030) does not need an additional reporting person. Calculation: 2–3 reporting FTE freed or avoided at HUF 1.0–1.2M/month = HUF 24M–43M/year in direct cost. The client retention value from digital reporting at the Debrecen automotive clients is not counted in that figure.
Estimated annual value
03 / 03
HUF 24M–43M
per year
2–3 reporting/admin FTE freed or avoided per site expansion cycle at HUF 1.0–1.2M/month total employment cost: 2 FTE x HUF 1.0M x 12 = HUF 24M/year (low); 3 FTE x HUF 1.2M x 12 = HUF 43M/year (high). These are avoided incremental hires as much as direct reductions: each new Debrecen site would otherwise require a reporting person. The agent absorbs the marginal workload.
Before anything else
We validate the figures first. Then we build.
Nothing here becomes a commitment until the figures are validated against Ghibli's actual TMS data, accounting records, and WMS exports. If the numbers hold, Lightbloom builds the fixes specific to Ghibli's stack. They are AI agent workflows that keep running. We earn 20% of Confirmed Annual Value per opportunity, once, after the savings land in Ghibli's accounts. Nothing before that.
-
Week 1
01 / 04
Systems audit and data access
Read-only access to the accounting system in the primary entity, the TMS or ERP handling road freight, and one WMS site. Data quality assessed against all three opportunity workstreams. No process changes proposed. The goal is to confirm what is real before anything is scoped.
-
Week 2
02 / 04
Subcontractor rate benchmark
Pull the last 24 months of subcontractor invoice data. Cross-reference the top 50 carrier-lane combinations against Timocom and Eurostat benchmarks. Rank by estimated annual saving if market rate is achieved. Deliver a first renegotiation opportunity list before week 3.
-
Week 3
03 / 04
Finance consolidation proof of concept
Run the first automated group P&L consolidation alongside the manual version. Identify the intercompany flows between the primary entity and the Cross Border Supply Chain Kft. Resolve discrepancies. Confirm the reporting layer is producing numbers the finance team trusts.
-
Week 4
04 / 04
Joint readout. Commitment or not.
Validated baseline against Ghibli's actual books. Scope and sequencing of the build agreed. If the numbers hold, the engagement begins. If not, both sides walk away with a cleaner picture of the cost base. Either outcome is honest.
Together, three opportunities
Opportunity One
HUF 24M–43M
Opportunity Two
HUF 71M–128M
Opportunity Three
HUF 24M–43M
· sum ·
HUF 119M–214M
per year, recurring
Identifiable from public sources only. Internal access to TMS invoicing, functional headcount, and WMS data is where the full picture lives. Lightbloom's estimate of full-engagement potential from a complete discovery: HUF 375M–593M per year.
Where this goes next
A thirty-minute call. Nothing more until it is mutual.
We walk through the brief together: your reactions, what we got wrong, what we missed. If the four-week diagnostic still makes sense at the end, we scope it. If not, we shake hands. Lightbloom only earns when Ghibli does.