Record revenue in July, a 16.17% share-price fall in October

LG Innotek closed October 12 at KRW 479,500, down KRW 92,500 or 16.17% from the preceding session. Back in July it had reported a record KRW 11.0621 trillion in first-half 2026 revenue and KRW 541.1 billion of operating profit.

These figures are not directly contradictory. Revenue and profit describe results already achieved; the share price incorporates expectations for the next reporting periods and market risk. A strong first half does not lock in margins for the third quarter.

Against that backdrop, the company has also been meeting robotics, autonomous-driving and advanced-packaging specialists in the US and collecting sensor data from real Korean roads. The strategic question is how to distinguish current revenue, new profit engines and the still-emerging capabilities that must serve future customers.

Q3 consensus was revised downward before the selloff

Local reporting on October 12 cited an October 8 consensus Q3 operating-profit estimate of KRW 202.9 billion, down 32.8% from KRW 302.0 billion a month earlier. Yuanta Securities offered a more conservative KRW 78.7 billion estimate. These are different forecasts, not finalized Q3 results.

Analysts pointed to a weaker won/dollar translation environment for earnings and a mismatch between raw-material purchasing and the timing of shipments tied to a major customer's new products. The broader KOSPI index also fell 3.65% on October 12.

Public information cannot quantify how much of LG Innotek's decline each factor caused. It would therefore be unsound to interpret the share price alone as evidence that its robotics partnership or substrate investments had failed.

Optical concentration and customer concentration are not the same

First-half segment disclosures put optical solutions revenue at KRW 9,128.5 billion, around 82.5% of consolidated revenue. Package solutions generated KRW 935.6 billion (8.5%) and mobility solutions KRW 998.0 billion (9.0%). Camera-centric optical manufacturing still provides the scale of the business.

A separate first-half report disclosed KRW 8,841.5 billion of combined optical/package sales to a single, unnamed customer exceeding 10% of total revenue. That works out to roughly 79.9% of consolidated sales. Korean press associates the account with Apple, but LG Innotek did not itself name the customer.

One percentage measures segment exposure and the other customer exposure; they cannot be added, nor can all of the unnamed customer's sales be assigned to optical alone. Taken separately, they show why protecting today's execution and broadening the future earnings mix can be simultaneous priorities.

Packaging delivers 18.4% of operating profit on 8.5% of revenue

Package solutions earned KRW 99.6 billion of first-half operating profit from KRW 935.6 billion in revenue: about a 10.6% margin. It represented around 18.4% of total KRW 541.1 billion operating profit. Optical solutions earned KRW 409.9 billion on KRW 9,128.5 billion in sales, a roughly 4.5% margin; mobility earned KRW 31.6 billion on KRW 998.0 billion, around 3.2%.

The optical business has not stood still: second-quarter revenue grew 48% year on year, versus 20% for packages and 10% for mobility. Even a successful emerging division can struggle to gain a much larger revenue share when the incumbent segment remains huge and is also growing.

One half-year margin is not a guaranteed long-term ranking. Nonetheless, the contrast shows why judging diversification solely by sales share misses a potential shift in where operating profit originates.

Package-substrate expansion is a yield and execution challenge

LG Innotek announced expansion of its semiconductor-substrate manufacturing footprint in Haiphong, Vietnam, with a planned completion in May 2027, complementing its site in Gumi, Korea. Management targets package-solutions revenue above KRW 3 trillion by 2030 and operating profit around KRW 1 trillion by 2031. These are forward-looking targets.

Product lines include RF-SiP, FC-CSP and FC-BGA substrates. Supplying larger or finer-pitch substrates for demanding processors is not just a question of additional floor space. Circuit density, materials, process control, manufacturing equipment and production yield all affect whether customers can be served repeatedly at scale.

The related capability boundary runs between designing the product and proving it can be manufactured at the customer's quality and yield levels. Those experiences overlap but are not identical—and these role illustrations are not announcements of specific new vacancies.

From making cameras to validating a robot's eyes and touch

On July 28, LG Innotek announced a Physical AI partnership with Japan's TDK to develop next-generation vision and tactile sensing modules. It described integrating optical and precision-engineering capabilities with inertial, positional and other sensor technologies, targeting initial modules in 2027.

On September 29 the company disclosed Korean-road operation of autonomous-sensing data-collection vehicles with US software company Applied Intuition. The fleet is planned to grow to about 20 data and development vehicles by 2028, combining camera, radar and lidar data in sensor-fusion development.

The move adds real-road evidence to component engineering. Rain, glare, construction-zone markings and sensor-calibration errors can expose problems not visible in a single sensor's specifications. Yet a data-collection fleet or collaboration announcement is not proof of finalized new customer contracts or performance gains.

Why meet robotics and packaging specialists in Los Angeles?

In August, LG Innotek hosted its Inno Connect talent event in Los Angeles. About 30 graduate-degree holders and experienced specialists from 16 US universities—including Stanford, Caltech, Yale and UCLA—participated. CTO Min Jon and CHO Lee Dong-hoon attended alongside R&D executives.

Highlighted fields included robotics/autonomous driving, advanced semiconductor packaging and next-generation optics. Attendance should not be mistaken for confirmed hires; the company did not provide an actual conversion count.

The profile of expertise reveals three distinct operational demands: maintaining high-volume optical quality, ramping advanced substrates with reliable yield, and verifying multi-sensor performance in physical environments. A generic search for 'robotics talent' would miss the specific work to be done.

BANSEOG VIEW | Revenue scale, profit mix and future capability use different clocks

LG Innotek's optical segment generated 82.5% of first-half revenue, while packaging produced around 18.4% of operating profit from 8.5% of sales. The company is also financing substrate capacity, building sensing partnerships and gathering on-road test data. These commitments operate at different scales and stages.

October's share-price shock followed revised Q3 profit expectations, a short-term earnings judgment. Whether the Haiphong footprint ramps as planned, sensor fusion wins customers, and packaging becomes a larger share of sustainable profit must be demonstrated later rather than inferred from a one-day stock move.

A company's workforce architecture should mirror those distinctions: proven optical operating expertise, advanced-package manufacturing yield, and sensor-fusion field validation represent different evidence. The question is not whether the company has announced a new business, but whether it can build, verify and deliver it repeatedly.

BANSEOG VIEW

Banseog View — Present revenue, profit contribution and future capabilities

Optical's 82.5% revenue share and an unnamed major customer's ~79.9% revenue share are two different measures of concentration; they cannot be added.

Package solutions' 8.5% share of revenue produced 18.4% of operating profit in H1 2026, while long-range package and sensing targets remain plans.

Robotics recruiting events and live-road sensing validation indicate desired expertise; neither proves the cause of the October 12 selloff nor confirms new hires or contracts.

SOURCES

Primary sources and references

As of 13 October 2026. LG Innotek's 12 October close was KRW 479,500 (−16.17%); the KOSPI fell 3.65% the same day. Q3 operating-profit consensus KRW 202.9bn and Yuanta's KRW 78.7bn are October 8 forecasts, not reported results. Segment revenue/profit for H1 2026 are reported from interim disclosures, with shares and margins calculated from the published values. The KRW 8,841.5bn unnamed customer's sales span optical and package segments; Apple is media inference rather than company identification. Targets for packages by 2030/2031, 2027 sensing modules and a ~20-vehicle fleet by 2028 are forward-looking. About 30 LA talent-event participants are not confirmed hires. Capability inferences are Banseog interpretation, not LG Innotek staffing or stock-price causation statements.