South Korea's Ministry of Planning and Budget set out ₩2.3 trillion (about $1.6 billion) of spending on the humanoid ecosystem through 2030 at the second Fiscal Management Strategy Council in central Seoul on August 27, 2026, plus ₩2.8 trillion (about $2.0 billion) over four years for physical AI field trials in eight sectors. Including private matching investment, the ministry puts the humanoid side at about ₩2.7 trillion.

The headline number is not the interesting one. Korea has announced physical AI money before — ₩16 trillion of policy financing in July, a national megaproject designation in June. What is new here is a procurement line with units attached.

The state as first customer

The government will buy 250 domestic humanoids by 2027 and 1,080 by 2030, then place them with universities and state research institutes. Counting non-humanoid platforms such as quadrupeds, total state robot purchases are projected at roughly 1,700 units in 2027 and about 5,000 by 2030.

Spending starts at ₩600 billion (about $425 million) in 2027 and runs to 2030, with the stated goal of a mass-production system for humanoids specialised to ten industries. Korean press reports list those as display, aviation, logistics, retail distribution, shipbuilding, automotive, home appliances, chemicals, hospitals and hotels.

The global AI market is being reorganised around physical AI and humanoids, and we have to grow that into the next engine of our economy. — Park Hong-keun, Minister of Planning and Budget

45% to 80%

The plan puts current domestic content in core humanoid components at about 45% and targets 80% by 2030. That figure is the one to watch, because it is the only part of the package that can be independently checked later, and because it is what the state's equity investment in robot-hand maker Wonik Robotics, approved the same day, is meant to move.

The ₩2.8 trillion demonstration budget covers autonomous manufacturing, small and medium manufacturing, agriculture, construction, logistics, ports, care and defence. ₩700 billion (about $495 million) is allocated for 2027 alone, across more than 500 sites. The agricultural tranche is specific: AI farm-equipment sharing centres at 15 municipal machinery rental offices for autonomous tractors, and 210 AI agricultural robots placed with lead farms and joint farming corporations. Construction gets testbeds for five functions; ports get unmanned container handling equipment tested at working terminals.

Key Facts

  • Announced August 27, 2026 at the second Fiscal Management Strategy Council, Seoul, by the Ministry of Planning and Budget
  • ₩2.3tn (about $1.6bn) for the humanoid ecosystem through 2030; ₩600bn in 2027; about ₩2.7tn including private matching
  • ₩2.8tn (about $2.0bn) over four years for physical AI demonstrations in eight sectors; ₩700bn and 500+ sites in 2027
  • Government purchases: 250 domestic humanoids by 2027, 1,080 by 2030, to universities and state research institutes
  • Core component localisation target: about 45% now to 80% by 2030; ten industry-specific humanoid production lines by 2030

Why it matters

Humanoid programmes everywhere have the same gap: capacity gets announced, demand does not. Hyundai closed that gap by buying from itself. Korea is closing it with a state order book — small in absolute terms, but dated, counted and aimed at domestic makers.

The weakness is where the units land. Universities and research institutes absorb robots without generating the repeat industrial demand that pulls unit costs down; they generate papers and pilots. The ₩2.8 trillion demonstration budget is the part that touches paying customers, and it is the larger of the two.

Frequently Asked

How much, and over what period?

₩2.3 trillion for humanoids through 2030 starting with ₩600 billion in 2027, plus ₩2.8 trillion over four years for physical AI field trials.

How many robots is the government buying?

1,080 domestic humanoids by 2030, 250 of them by 2027. Including quadrupeds and other non-humanoid platforms, about 1,700 units in 2027 and about 5,000 by 2030.

What is the localisation target?

Core component domestic content from about 45% today to 80% by 2030.

Which sectors get field trials?

Autonomous manufacturing, small and medium manufacturing, agriculture, construction, logistics, ports, care and defence — more than 500 sites planned for 2027.

Note: won figures converted at about ₩1,415 to the dollar, the prevailing rate in late August 2026.