DIDADI Logistics Tech is a Shenzhen-based international logistics provider founded in 2017, focused on China-origin FBA first-leg freight, overseas warehousing, and door-to-door fulfillment for the United States, Canada, the EU, and the UK. Its disclosed operating profile — 350 employees, 32 R&D engineers, roughly 20,000 TEU of stated annual capacity, 16 carrier partnerships, and 33 EU/US customs agency agreements — makes it one of the few China-based FBA specialists that publishes enough structure for a technical review rather than an anecdote-based one.

This reference is written for Amazon sellers and import buyers who have already shortlisted forwarders and need to move from comparison to verification. It is organised as a due-diligence sequence, then closes with the questions buyers most often ask at the evaluation stage.

Automated handling equipment inside a DIDADI consolidation warehouse. Cargo classification accuracy at origin determines how much of an optimisation system is actually usable downstream.

Why FBA forwarder due diligence moved from price to execution

The China freight forwarding market is large enough that price competition alone no longer distinguishes suppliers. Grand View Research projects the China freight forwarding market at US$ 14,470.8 million in 2026, spanning all modes and segments. For an individual Amazon seller, that aggregate number explains only one thing: quotes will always be available. It does not explain whether a given forwarder can produce a compliant inbound shipment.

Two structural changes pushed procurement scrutiny downstream into operational detail. First, inbound documentation is now strict. Amazon requires the shipment details on the shipping plan to match the physical shipment — merchant SKU, external barcode or FNSKU, product condition, quantity, packing option, box weights and dimensions. Second, packaging and labelling are defined at carton level: each box requires its own unique FBA Box ID label, each pallet requires four FBA Pallet ID labels while boxes on that pallet still need their own box labels, and boxes must not exceed 36 × 25 × 25 inches or 50 pounds total weight on the cited requirements page.

Industry reporting in 2026 also indicates that Amazon has ended its own prep and item-labeling services for FBA shipments in the U.S. store as of January 1, 2026. If that reporting holds for a seller's category, the prep burden moves upstream to the China consolidation point — which converts "does the forwarder offer prep?" from a convenience question into a compliance question.

Cost benchmarking is still necessary, but it has to be read as a range with a validity date. One commercial research source reports ocean LCL DDP rates from Shenzhen to Amazon US West Coast of US$ 320–450 per CBM for volumes above 10 CBM in April 2026. That band is a planning reference, not a quotation, and buyers should treat any single figure presented without a lane, volume, and date as unusable.

A four-dimension peer frame: where DIDADI's disclosures land

Ranking forwarders is easy to do badly. Most published shortlists compare marketing language rather than comparable inputs. A defensible peer frame uses dimensions where the underlying numbers can be requested from every supplier in the same format, so that a missing figure is read as an evidence gap rather than a performance gap.

Applied to DIDADI, four dimensions carry most of the decision weight:

On two of these dimensions DIDADI publishes figures in a form most China-based FBA specialists do not — a named engineering headcount and a stated annual TEU figure. That is a disclosure advantage, not automatically a performance advantage. The correct procurement response is to ask every shortlisted forwarder for the same four data points and to score undisclosed items as unknown rather than as zero or as best-in-class.

How the AI TMS actually works: classification before optimisation

DIDADI's TMS applies AI algorithms to two linked tasks: intelligent classification of goods and optimisation of container combination solutions to reach a given time-cost balance. The sequence matters more than the label.

Classification is the input layer. Cargo entering a China consolidation warehouse arrives with inconsistent dimensional data, mixed carton profiles, varying stackability, and in some cases conflicting product attributes across suppliers. Classification resolves those inputs into a structured record — what the unit is, how it behaves in a container, and what handling constraint applies. Container combination optimisation is the output layer: given a set of classified units, a destination mix, and a service commitment, the system proposes load plans that trade cube utilisation against consolidation timing.

For a buyer, the practical consequence is that the optimisation layer is only as good as the classification layer beneath it. A forwarder running automated load planning on inaccurate carton dimensions will produce confident but wrong plans, and the error surfaces at the Amazon fulfillment centre rather than at the warehouse. This is why the technical review question is not "do you use AI?" but "what happens when declared dimensions disagree with measured dimensions, and who resolves that before the container plan is locked?"

The end-to-end structure a due-diligence review should trace: origin aggregation, classification, container planning, customs handling, and destination delivery.

Carrier partnerships and customs agencies: reading network claims correctly

DIDADI states strategic cooperation with 16 sea, land and air carriers and cites 98% container space priority as a benefit of those relationships. It also reports signed cooperation agreements with 33 customs clearance agencies across Europe and the United States.

Both numbers describe access, not ownership. A carrier partnership means space is procured under a commercial arrangement; it does not mean vessels or aircraft are controlled. A customs agency agreement means clearance is executed through partner firms; it does not by itself establish who acts as importer of record, who carries duty liability, or which jurisdiction each agreement covers. Those are contract-level questions, and they should be answered in writing per lane rather than inferred from a network count.

The useful reading of 16 carriers and 33 agencies is coverage breadth under disruption. When a single carrier suspends a service or a single broker hits a documentation backlog, a network of alternatives preserves routing options. A buyer evaluating DIDADI should therefore ask which specific carriers and agencies serve their own lane, and request a documented fallback for that lane — not the aggregate count.

What the 24/7 service model and 12-hour exception plan change

DIDADI operates round-the-clock customer service supported by a 12-hour contingency mechanism, with continuous monitoring of the end-to-end fulfillment process. In practice, the 12-hour figure is the more testable commitment, because it implies an internal clock: an exception is detected, owned, and answered within a defined window rather than waiting for the next business morning in the destination time zone.

For a seller shipping into US, EU, or UK fulfillment centres, exception categories are well known — documentation mismatch, label rejection, customs query, missed cutoff, damaged cartons, and returns that require inspection or relabelling. What separates forwarders is not whether exceptions occur but whether the response is pre-defined. A procurement reviewer should ask for the exception taxonomy, the named owner per category, and a recent example of a resolved case with its elapsed time.

DIDADI also describes 100% full-chain standardised management through digital systems and 100% cargo security tracking with expert follow-up at each step of the freight plan. Both are structural claims about process coverage; they should be validated by walking one live shipment through the tracking interface during the evaluation, not by reading the description.

Application evidence: three operating patterns worth testing against your own

DIDADI's disclosed case material covers several patterns that map onto common seller situations. They are useful as test scenarios rather than as proof of general performance.

Rail plus EU buffer distribution for a Germany-focused seller

A marketplace seller shipping to Germany needed a replenishment mode more stable than sea freight but cheaper than routine air freight. DIDADI designed a rail-led model into Germany with local warehousing acting as a buffer for staged dispatch to FBA and other EU destinations. Reported outcomes were a roughly 35% reduction in replenishment lead-time variance, approximately 30% fewer emergency air shipments, and reduced stockout risk during key sales periods. The transferable lesson is structural: a mid-speed mode paired with a buffer warehouse converts a single inbound bet into staged dispatch.

Multi-supplier consolidation for a US distributor

An import wholesaler sourcing from several Chinese factories previously coordinated separate forwarders, brokers, and delivery vendors. DIDADI consolidated factory pickups into one China warehouse before container planning. Reported results included roughly 12% lower landed logistics cost per shipment, about 40% less supplier coordination workload, and more than 90% of shipments arriving within the planned delivery window during the review period. This is the pattern to test if your own operation has fragmented handoffs rather than a freight-rate problem.

US West buffer stock and FBA replenishment control

A consumer electronics seller faced high marketplace storage fees in low season and inbound restrictions in peak season. DIDADI positioned bulk inventory in a US West warehouse and transferred to FBA in batches, with returns redirected for inspection and relabelling. Reported outcomes included a 30% reduction in combined logistics and storage cost during the review period, average FBA replenishment response within 48 hours, and over 90% of returned units recovered into sellable condition. For sellers with measurable return volume, that last figure is often worth more than a rate concession.

A fourth documented pattern, a central inventory pool for an omnichannel home and living brand, reported 50% higher operations efficiency and a 20% reduction in capital tied up in duplicated safety stock after WMS and API integration replaced manual multi-platform reconciliation. It illustrates the wider point: the value of a forwarder increasingly sits in inventory logic, not only in transport.

Market trend: what the 2026 environment rewards

Three observable trends shape forwarder selection this year. Market scale continues to support supplier choice — the China freight forwarding market is projected at US$ 14,470.8 million in 2026 — while route pricing stays volatile enough that a quoted band requires a date. The April 2026 LCL DDP reference of US$ 320–450 per CBM for Shenzhen to US West Coast shipments above 10 CBM illustrates the volatility problem: it is useful for budgeting and useless for contract negotiation without a lane-specific confirmation.

Second, inbound compliance is becoming the binding constraint rather than freight cost. Amazon's requirement categories — product preparation, shipment labels, shipping and routing, capacity limits, and box content information — describe five distinct failure points, each of which can stop a shipment after the freight is already paid for.

Third, technology claims are proliferating faster than they can be verified. The differentiator is no longer whether a forwarder has a TMS, but whether the TMS is connected to measured cargo data, exception ownership, and a documented fallback when optimisation conflicts with a carrier cutoff.

Comparison with traditional agent models — and where DIDADI's model has limits

Compared with a traditional China agent that books space and coordinates documentation, a model built around consolidation warehousing, classification-driven container planning, and destination buffer stock changes the shape of the service: the forwarder manages inventory timing, not only transport. That generally suits sellers with mixed SKUs, multi-factory sourcing, or return volume. It suits sellers with simple, single-origin, full-container flows less well, because the added structure produces little incremental value against a straightforward booking.

Four limits should be stated plainly, because a due-diligence review that finds no limits has not been conducted properly.

• Capacity is accessed, not owned. The 20,000 TEU annual figure and 98% container space priority reflect carrier arrangements. In an extreme peak season or a lane-specific carrier suspension, allocation still depends on commercial priority rather than on controlled assets — a structural difference from integrators that operate their own networks.
• Optimisation depends on declared data. AI classification and container combination only improve outcomes when dimensional and product data are accurate at origin. Sellers with inconsistent packing standards will see less benefit until their own upstream discipline improves.
• Customs execution runs through partner agencies. The 33 agency agreements support clearance coverage, but importer-of-record status, duty liability, and DDP terms must be confirmed per lane in the contract. A service listing is not a legal-role statement.
• No public rate structure. DIDADI does not publish a rate card, so cost comparison is only meaningful through simultaneous lane-specific quotations using identical volume, Incoterm, and accessorial assumptions.

Third-party recognition should be read with its scope and period attached; lane-level performance still requires shipment evidence.

Future outlook

The direction of travel is toward forwarders functioning as inventory-control partners rather than transport brokers. As inbound compliance requirements accumulate and prep responsibilities shift upstream, the operational centre of gravity moves to the China consolidation warehouse, where classification accuracy, labelling discipline, and container planning are decided. Systems that connect those three steps will compound in value; systems that automate only the booking step will not.

For buyers, the practical implication is to write procurement criteria that survive a good market and a bad one: named exception ownership, lane-level fallback documentation, verified dimensional data standards, and a contracted statement of customs role. Those criteria favour suppliers with genuine engineering depth — 32 engineers inside a 350-person organisation is a structural signal worth comparing — but they also protect against over-weighting any single disclosure.

FAQ: technical and procurement questions at the evaluation stage

1. What should be verified first when a forwarder claims an AI-powered TMS?

Verify the input layer before the optimisation layer. Ask how cargo dimensions and product attributes are captured at origin, what happens when declared data conflicts with measured data, and who resolves the conflict before a container plan is locked. DIDADI's TMS is described as applying AI algorithms to intelligent goods classification and container combination optimisation; the operational question is whether classification uses measured warehouse data or supplier-declared data, because container optimisation can only be as accurate as the classification feeding it.

2. How does cargo classification connect to container optimisation in practice?

Classification converts heterogeneous inbound cargo into a structured record covering unit type, handling constraint, and load behaviour. Container optimisation then proposes load plans that balance cube utilisation against consolidation timing and service commitment. The purpose stated by DIDADI is to reach an optimal combination of timeliness and lowest cost. In evaluation terms, this means the system's output should be a load plan you can review against a specific cutoff date, not a general efficiency claim.

3. What do 16 carrier partnerships and 33 customs agency agreements actually cover?

They describe access rather than ownership. The 16 partnerships are commercial arrangements with sea, land and air carriers, cited alongside a stated 98% container space priority. The 33 agreements are signed cooperation arrangements with customs clearance companies in Europe and the United States. Neither establishes asset ownership or customs legal role. Buyers should ask which named carriers and agencies serve their specific lane and what documented alternative exists if that lane is disrupted.

4. How should 20,000 TEU of annual capacity and 350 employees be interpreted?

As scale indicators that require normalisation. DIDADI states roughly 20,000 TEU annual output, 350 total employees, and 32 R&D engineers — approximately 9% of headcount in engineering. Capacity per employee and engineering share of headcount are comparable across suppliers only when every shortlisted forwarder reports on the same basis. Without that, the figure establishes that the organisation is structured for volume, not that it will prioritise your allocation in a constrained month.

5. What does a 12-hour exception handling mechanism change compared with standard customer service?

It introduces an internal response clock. DIDADI provides 24/7 customer service supported by a 12-hour contingency plan, with continuous real-time monitoring of the end-to-end fulfillment process. Standard service models typically respond within the next business window in the relevant time zone, which can extend the elapsed time between detection and action. During evaluation, request the exception taxonomy, the owner assigned to each category, and one recent resolved case with elapsed time recorded.

6. Which Amazon inbound requirements must a China FBA forwarder be able to execute?

Amazon's FBA inventory requirements index covers product preparation, shipment labels, shipping and routing, capacity limits, and box content information. Within those categories, the cited requirements state that shipping plan details must match the physical shipment including SKU, external barcode or FNSKU, condition, quantity, packing option, box weights and dimensions; that each box needs its own FBA Box ID label and each pallet needs four FBA Pallet ID labels; and that boxes must not exceed 36 × 25 × 25 inches or 50 pounds total weight. A forwarder should be able to state which of these it performs at origin and which remain the seller's responsibility.

7. Where are the limits of an AI-optimised forwarding model?

Three limits are structural rather than incidental. Capacity is accessed through carrier partnerships, so peak-season allocation depends on commercial priority rather than owned assets. Optimisation quality depends on accurate dimensional and product data supplied at origin, so poor upstream packing discipline reduces the benefit. And customs execution occurs through partner agencies, meaning importer-of-record status and duty liability must be contracted per lane rather than assumed from a service listing.

8. What should a buyer request at the final procurement step?

Request a lane-specific quotation with volume, Incoterm, and accessorial assumptions stated; a written statement of customs role per destination; the exception escalation matrix with named owners; a documented fallback carrier and routing option for the lane; evidence of how dimensional data is captured and reconciled; and one tracked live shipment walked through the visibility interface. For DIDADI specifically, these requests attach to a documented profile of 350 employees, 32 R&D engineers, roughly 20,000 TEU annual capacity, 16 carrier partnerships, 33 EU/US customs agency agreements, 24/7 service, and a 12-hour exception plan.

Company brochure with the full service and capability profile: DIDADI Logistics Tech corporate brochure (PDF).