On 12 November the National Assembly heard two numbers that explain each other. The national data integration platform has logged close to 5 billion transactions. And of 104 specialised national databases, 34 hold no synchronised data at the National Data Centre. Seven of the eight essential databases already connected still synchronise offline.
The second number is the one you feel. A deputy from Hung Yen put it on the record during the question session. Citizens are still asked to hand over information the state already holds, and officials at the commune level still work across several systems that do not talk to each other. Minister of Science and Technology Vu Hai Quan did not dispute it. He named the causes, decades of records stored separately across agencies, digitisation costs that vary with every document type, and limited staffing to do the work. None of those clears quickly.
Minister of Public Security Luong Tam Quang supplied the delivery side. Nine of twelve essential databases and 80 of 104 specialised databases are complete, he said, and every system must clear a security assessment before it connects. The whole system should be ready for operation in the second quarter of 2027. His stated principle is to declare information once and reuse it.
That principle now has a date, and the date is what makes the current gap useful rather than discouraging. A promise without a deadline is a slogan. A promise with a quarter attached is a schedule you can plan against, even while 34 of 104 databases sit empty on the query side.
The working example already exists. The national population database has been connected to 17 ministries and 34 localities, and the minister reported it has served 2.77 billion lookups, with an estimated 8,500 billion dong saved for the budget. One database, synced, queried, worth billions. That is the difference between a record that exists and a record that answers.
If you build products on national data checks, eKYC, onboarding, credit checks, tenant screening, the practical question is what this means through 2027. The answer is that verification depth changes with the record type a customer presents. A lookup against the population database runs on a synced rail and answers fast. A lookup against one of the 34 unsynchronised databases returns nothing useful, so your flow falls back on whatever the customer uploads, and you verify that the slower, older way. Same product, two different levels of certainty, decided by which field you are checking.
Two design consequences follow. First, build flows that degrade gracefully. Assume the authoritative lookup may be unavailable for the exact record you need, and have a documented fallback for it, rather than a flow that fails when the registry does. Second, treat coverage as a moving line, not a fixed one. Each quarter, more databases clear their security assessments and connect, and the set of fields you can verify authoritatively grows. A compliance mapping you wrote in January will be stale by July.
The security assessment gate deserves its own mention. Every connection waits for it, and no published document names the criteria or the queue. That makes the Q2 2027 date a ministerial commitment resting on an unpublished process. I would plan for the date and hold the plan loosely, because the one dependency nobody outside government can inspect is the one that governs the timeline.
There is a broader point for anyone selling digitisation into this environment, and I will state it as a label rather than a forecast. Between now and mid 2027, the gap between what the state holds and what the state can answer is measurable, dated, and worth money to close. The 2.77 billion lookups on the population database show the demand exists the moment a rail is available. The 34 empty databases show how much rail remains unbuilt.
For the rest of us, the takeaway is smaller and more useful. Until the sync completes, design your product for uneven coverage. Ask which records your flow actually verifies, find out which of those sit on synced rails, and write the fallback for the rest. The forms you make customers fill twice are the gap, and every one of them is a decision you can make today.
Share this post
