📚 Context This is a follow-up to Korea Quality Re-Rating Watch 2026-06-04, AI Data Center CapEx at USD 5.3T, Broadcom confirms the 2027 AI semiconductor USD 100B frame and After NVIDIA: the AI semiconductor bottleneck. Related hubs: AI PCB and Substrate Hub, Korea Semiconductor Value Chain Hub and Foreign Investors in Korea Hub.
TL;DR
GigaVis has moved onto the watchlist as a Korean AI-substrate equipment name with both relative strength and a clear ownership handoff.
But the category matters. GigaVis is not Samsung Electro-Mechanics or Korea Circuit. It does not manufacture substrates. It sells inspection, review, repair, factory-automation and software equipment used in high-end FC-BGA / ABF substrate production. The investment question is not simply “more AI substrates.” It is whether finer AI GPU / ASIC substrates make yield-control equipment more valuable.
| Item | Read |
|---|---|
| Business position | Inspection / repair equipment for AI FC-BGA and ABF substrates |
| Latest price action | +6.7% over five sessions, top 5.2% of Korean equities |
| Latest flow | Retail sold; foreigners, institutions and programs absorbed supply |
| 2026 disclosed orders | Three contracts totaling KRW 29.41B, 56.1% of FY2025 revenue |
| FY2025 product mix | AOI/VRS/AOR 59.4%, FA 22.9%, software 4.0% |
| Margin structure | 2023 OPM 38.3% → 2024 -6.8% → 2025 23.1% |
| Main risk | Revenue-recognition timing, customer CapEx delays and valuation already pricing part of the thesis |
| Action stance | Watchlist / wait for confirmation |
1. What kind of AI-infrastructure stock is GigaVis?
GigaVis’s key products are AOI, VRS, AOR, FA and software. In plain terms, the company helps customers inspect high-end semiconductor substrates, review defects, repair some of them and automate / manage the process.
| Product | Role | Investment relevance |
|---|---|---|
| AOI | Automated optical inspection | Finds defects in fine-pattern, high-layer substrates |
| VRS | Review station | Validates and classifies defects |
| AOR | Automated optical repair | Helps improve substrate yield |
| FA | Factory automation | Automates the inspection / review / repair flow |
| Software | Data management | Extends the value proposition beyond hardware |
As AI GPUs, AI ASICs, HBM packages and networking chips grow larger and more complex, FC-BGA substrates need more layers, finer wiring and tighter process control. If yield slips, substrate cost and delivery schedules move immediately.
That makes the GigaVis thesis:
AI GPU / ASIC package complexity
→ finer and more complex FC-BGA / ABF substrates
→ harder inspection and repair
→ higher need for AOI / AOR / automation equipment
→ more volatile but potentially higher-value orders for GigaVis
GigaVis is therefore best viewed as a back-end yield-control equipment name inside the AI substrate chain.
2. The 2026 orders matter, but should not be overstated
Three major disclosed orders have been identified so far in 2026.
| Date | Region | Content | Contract value | Period |
|---|---|---|---|---|
| 2026-03-03 | Taiwan | Semiconductor substrate inspection / repair equipment | KRW 10.91B | 2026-02-27 to 2026-12-03 |
| 2026-04-24 | Korea | Semiconductor substrate inspection / repair equipment | KRW 9.04B | 2026-04-24 to 2026-10-30 |
| 2026-06-04 | Japan | Semiconductor substrate inspection / repair equipment | KRW 9.47B | 2026-06-02 to 2027-05-31 |
| Total | - | - | KRW 29.41B | - |
All three contracts are described as semiconductor-substrate inspection / repair equipment, and customer names are withheld for confidentiality. Payment terms include order, delivery, installation and inspection / acceptance milestones. That matters because confirmed orders do not equal same-quarter revenue for equipment companies. (Digital Today, Digital Today, Digital Today)
FY2025 revenue was KRW 52.44B. The disclosed 2026 order total equals roughly 56.1% of that revenue base.
2026 disclosed order ratio
= KRW 29.41B / KRW 52.44B
= ~56.1%
That is positive. But it is not yet proof of explosive growth. Against a similar point last year, the order level looks more like a strong absolute base than a major YoY acceleration. Equipment revenue can also shift across quarters because delivery, inspection and acceptance timing matter.
The key question is where the orders expand: only FC-BGA, or also WLP / PLP RDL, glass substrates and more advanced inspection layers.
3. Product mix: not a single AOI tool
FY2025 revenue was KRW 52.44B. The mix shows that the recovery was not only AOI; AOR, automation and software also came back.
| Product | FY2025 revenue | Mix | YoY | Investment read |
|---|---|---|---|---|
| AOI | KRW 18.31B | 34.9% | +109.1% | Core optical inspection |
| VRS | KRW 3.10B | 5.9% | +69.9% | Defect review / validation |
| AOR | KRW 9.75B | 18.6% | +270.5% | Laser repair and yield recovery |
| FA | KRW 12.01B | 22.9% | +157.1% | Inline automation, loaders / unloaders |
| Software | KRW 2.09B | 4.0% | +90.0% | Inspection data and defect classification |
| Service | KRW 6.90B | 13.2% | +7.5% | Installed-base service |
| Rent | KRW 0.28B | 0.5% | -61.4% | Small |
| Total | KRW 52.44B | 100.0% | +100.6% | Recovery after the FY2024 downturn |
AOI, VRS and AOR together represented 59.4% of FY2025 revenue. Including FA and software, GigaVis looks less like a one-product AOI vendor and more like a process stack for inspection, review, repair, automation and data management.
That is the business-quality point. As FC-BGA substrates become larger, finer and more expensive, customers do not only want to find defects; they want to recover yield.
4. Margin: high-quality equipment, but lumpy quarters
GigaVis can produce strong margins when utilization and mix are normal. It can also swing hard when customer CapEx slows or revenue recognition shifts.
| Period | Revenue | Operating profit | OPM |
|---|---|---|---|
| 2023 | KRW 91.42B | KRW 35.01B | 38.3% |
| 2024 | KRW 26.13B | -KRW 1.79B | -6.8% |
| 2025 | KRW 52.44B | KRW 12.10B | 23.1% |
| 1Q26 | KRW 6.0B | -KRW 1.0B | about -16% |
The 2025 gross margin was 54.1%, and the 2025 operating margin recovered to 23.1%. But Q1 2026 revenue was only KRW 6.0B with an operating loss, after Q4 2025 had KRW 31.7B revenue and 35.0% OPM. That is not automatically demand destruction; it is the lumpy delivery and acceptance pattern of equipment revenue. (AwakePlus)
The next key checks are backlog, contract liabilities / advances, gross margin, AOR/FA mix and acceptance timing.
5. Moat: the recipe inside the customer’s process matters
The moat is not simply a camera, a laser or a stage. The more important asset is the customer-validated inspection / repair recipe and defect-data loop.
| Moat layer | What it means | Investment relevance |
|---|---|---|
| Performance moat | Optics, lighting, stage control and image processing | Harder as line / space gets finer |
| Yield moat | AOI detects; AOR repairs | Defect recovery has real economics on expensive substrates |
| Qualification moat | Equipment is tuned to customer process recipes | Switching equipment creates requalification risk |
| Data moat | DTS / GiDC-style defect tracking and classification | Installed base can improve software and lock-in |
Some pieces are replicable. Global players can source cameras, optics, stages, lasers and AI-classification software. Competitors such as KLA, Camtek and Inspec are relevant to the broader inspection-equipment field, and KB Securities has discussed this competitive frame. (KB Securities)
The harder-to-copy part is the combination of customer-specific fine-pattern recipes, AOI-AOR linkage, post-laser repair quality assurance, global substrate qualification and installed-base data.
6. Step-change triggers
The stock deserves a higher-quality re-rating only if several checks move from option to evidence.
| Trigger | What to watch | Meaning |
|---|---|---|
| FC-BGA expansion restarts | Additional orders from Japan, Taiwan and Korea | AI GPU / ASIC substrate CapEx becomes equipment revenue |
| AOR / FA mix rises | Higher AOR and FA revenue share | More yield-improvement value, not just inspection |
| WLP / PLP RDL enters production | OSAT or semiconductor-maker orders | TAM expands beyond FC-BGA |
| Glass substrate inspection advances | Pilot delivery or customer validation | Long-dated next-generation substrate option |
| Software / service expands | Software and service revenue growth | More recurring, less cyclical revenue layer |
So far, only the first trigger has early evidence. Taiwan, Korea and Japan orders are positive. PLP / WLP / glass-substrate and software expansion remain option value until specific orders or repeat revenue are disclosed.
7. Price action: it beat the market
The data window is the five trading days from May 28 to June 4, 2026. The latest local database close is June 4.
| Period | GigaVis |
|---|---|
| Five-day return | +6.7% |
| June 4 return | +11.2% |
| Five-day rank | 143 / 2,717 equities |
| Five-day percentile | Top 5.2% |
| One-day rank | 70 / 2,717 equities |
The median Korean equity was down 6.3% over the same five-day period. GigaVis was not simply bouncing with the tape; it was outperforming a weak market.
| Peer / related stock | Five-day return |
|---|---|
| Eugene Technology | +20.0% |
| Korea Circuit | +9.2% |
| GigaVis | +6.7% |
| Isu Petasys | +4.9% |
| EO Technics | +2.4% |
| Simmtech | -2.4% |
| Daeduck Electronics | -3.5% |
| Samsung Electro-Mechanics | -7.2% |
| TLB | -14.9% |
This places GigaVis in the second-line substrate / inspection-equipment strength bucket.
8. Flow: retail sold, foreigners and institutions absorbed
| Date | Close | Retail | Foreigners | Institutions | Read |
|---|---|---|---|---|---|
| 5/28 | KRW 130,700 | -KRW 2.30B | -KRW 1.64B | +KRW 3.89B | Institutions moved first |
| 5/29 | KRW 141,400 | -KRW 5.11B | -KRW 0.99B | +KRW 6.61B | Strong institutional buying |
| 6/1 | KRW 134,200 | -KRW 3.11B | +KRW 9.39B | -KRW 6.50B | Foreigners absorbed the pullback |
| 6/2 | KRW 125,400 | -KRW 1.07B | +KRW 3.30B | -KRW 2.61B | Foreign support |
| 6/4 | KRW 139,400 | -KRW 7.87B | +KRW 3.99B | +KRW 3.89B | Foreigners and institutions both bought |
| Investor | Five-day net flow |
|---|---|
| Retail | -KRW 19.46B |
| Foreigners | +KRW 14.04B |
| Institutions | +KRW 5.29B |
| Foreigners + institutions | +KRW 19.33B |
| Five-day turnover | ~KRW 129.77B |
Foreigners plus institutions bought about 14.9% of five-day turnover. That is meaningful absorption of retail supply.
Institutional flow was mixed internally.
| Institution type | Five-day net flow |
|---|---|
| Financial investment | +KRW 2.76B |
| Insurance | +KRW 2.19B |
| Investment trusts | -KRW 6.23B |
| Private funds | +KRW 1.60B |
| Pension funds | +KRW 4.80B |
Investment trusts sold, while pension, insurance, private funds and financial investment accounts bought. This is not a one-way institutional accumulation story, but some higher-quality pockets of capital are present.
Supporting data is also constructive.
| Item | Latest state |
|---|---|
| Program net buying | +KRW 14.10B over five days |
| Short-sale turnover | KRW 2.56B over five days |
| Short-sale share | ~2.0% of turnover |
| Foreign ownership | 2.63% on 5/28 → 3.32% on 6/4 |
The program bid supports the flow, but also means the foreign buying should not automatically be treated as pure long-only accumulation.
9. Why the stock is back on the screen
First, the AI-substrate discussion has focused heavily on substrate makers. But the more complex the substrate, the more valuable inspection and repair become. GigaVis fills that missing layer.
Second, the product mix is broader than a single inspection tool. AOI, VRS, AOR, automation and software together create a yield-management equipment stack.
Third, orders have been geographically diversified across Taiwan, Korea and Japan in 2026. Customer names are undisclosed, but the pattern looks more like a high-end substrate CapEx restart signal than a single-customer event.
Fourth, relative strength appeared during a weak tape. A stock that holds up when the market is weak tends to re-enter institutional screens quickly when the next round of risk appetite appears.
10. Risks
The first risk is category confusion. GigaVis does not sell AI chips. It sells equipment to substrate customers. If customer CapEx is delayed, orders and revenue recognition can slip.
The second risk is quarterly volatility. FY2025 revenue recovered to KRW 52.44B with operating profit of KRW 12.1B, but FY2024 was loss-making and Q1 2026 also posted an operating loss. Equipment revenue can be lumpy.
The third risk is price. After a +11.2% one-day move on June 4, chasing the stock is less efficient.
The fourth risk is flow quality. Foreigners bought aggressively, but program buying was also large. Part of the move may be basket / program-driven rather than long-only accumulation.
The fifth risk is competition and customer diversification. GigaVis may have a strong position in high-end FC-BGA inspection / repair, but global inspection-equipment suppliers can still pressure pricing and customers may dual-source over time.
11. Action view
Current stance: Watchlist / wait for confirmation.
| Condition | Meaning |
|---|---|
| Support around KRW 135,000 | Healthy pullback after the June 4 surge |
| Breakout above KRW 144,400 with renewed turnover | Short-term trend resumes |
| Two or more sessions of foreign + institutional buying | Ownership handoff continues |
| Program buying holds and short-sale share stays low | Mechanical selling pressure remains limited |
| Break below KRW 125,000 | Weakens the flow-turn thesis |
| Q2-Q3 revenue conversion | Confirms Q1 weakness was timing, not demand |
| OPM returns toward 30% | Key proof for a high-margin equipment re-rating |
Fund-manager summary:
GigaVis is not a direct AI-chip bet. It is a way to express the AI substrate yield-control bottleneck. The five-day ownership handoff is constructive, but price discipline matters after the June 4 surge. The better setup is confirmation around KRW 135,000 support or a clean breakout above KRW 144,400.
Evidence Classification
[Fact]
- GigaVis rose 6.7% over the latest five trading days, ranking in the top 5.2% of Korean equities. Data window: 2026-05-28 to 2026-06-04, local database.
- Five-day flow was retail -KRW 19.46B, foreigners +KRW 14.04B and institutions +KRW 5.29B.
- The three identified 2026 disclosed orders total KRW 29.41B, equal to 56.1% of FY2025 revenue. (Digital Today, Digital Today, Digital Today)
- FY2025 product mix was AOI 34.9%, VRS 5.9%, AOR 18.6%, FA 22.9%, software 4.0% and service 13.2%.
- FY2025 gross margin was 54.1% and operating margin was 23.1%. 2023 OPM was 38.3%; 2024 OPM was -6.8%.
- Q1 2026 revenue was KRW 6.0B with an operating loss of about KRW 1.0B. (AwakePlus)
[Inference]
- GigaVis should be classified as an AI FC-BGA / ABF substrate yield-equipment beneficiary, not as a direct AI chip or substrate maker.
- The latest flow is a retail-to-foreign/institutional ownership handoff.
- If AOI/AOR/FA/software mix keeps expanding, GigaVis can be valued less as a simple inspection-tool vendor and more as a yield-management equipment platform.
- Because program buying was also large, the foreign buying cannot yet be treated as pure long-only accumulation.
[Blocked]
- Customer names, final package applications, equipment-level ASP, total backlog and 2027 revenue-recognition pace are not fully confirmed in public materials.
- WLP / PLP RDL and glass-substrate inspection are confirmed as development / validation areas, but mass-production revenue and customer timing remain unconfirmed.
- June 5, 2026 intraday data was not included.