Tesla's Q2 Results: Growth Stalls and Profitability Slips, and Musk Has No Easy Answer

The decline came at a time when interest in Tesla had relatively waned after the SpaceX listing, and analysts see internal changes at the company as being reflected in the numbers.

The segment had grown rapidly and accounted for a large share of total gross profit, but its gross margin has fallen sharply recently, a decisive factor in the overall drop in profitability.

In particular, low-price offensives from Chinese companies such as BYD are analyzed as having directly hit the energy business's profitability. Tesla has not responded effectively to these external pressures and is experiencing a weakening of competitiveness in its core business.

The decline in gross margin is not merely a problem in one business segment; it raises fundamental questions about Tesla's cost management and strategy for maintaining competitiveness across the company.

The plan aims to produce in-house the chips required for robots, self-driving cars, and data centers, reducing external dependence and strengthening technological independence.

Terafab is targeting enormous chip production capacity and is expected to become a core pillar of Tesla's long-term growth strategy. Through it, Tesla plans to further strengthen its ability to integrate hardware and software.

Such seemingly unrealistic production targets have raised concerns about the scale of the investment and how it would be funded. Skeptics question whether Tesla can build and operate capacity on this scale within a short timeframe.

Terafab is intended to serve as the core infrastructure for mass-producing these proprietary AI chips, particularly the chipsets equivalent to GPUs.

This is seen as a strategy to secure the high-performance computing chips needed for autonomous driving, robotics, and data centers without external constraints, and to gain a technological edge over competitors. In-house chip production is expected to further strengthen Tesla's integrated hardware-software ecosystem.

Through this, Musk aimed to build an independent AI ecosystem and escape external dependence.

From this perspective, the Twitter acquisition can be seen not simply as a business decision, but as a strategic investment to secure AI competitiveness across Musk's overall business portfolio, including Tesla.

Ultimately, the acquisition suggests that it was a crucial piece of the puzzle for making AI the core driver of the future business structure Musk envisions.

First, the completeness of the physical devices needed to collect data, such as robots and self-driving vehicles, is still insufficient.

Second, building a data accumulation system capable of effectively storing and processing the vast amounts of data collected through these devices remains an outstanding challenge. These four hurdles are challenges Tesla must overcome to realize its long-term vision.

Tesla's plan to build Terafab and then construct AI data centers faces serious spatial constraints. Building large data centers on the ground is not only limited by available physical space but also raises issues such as local residents' opposition and heavy water consumption. Tesla therefore faces the task of finding new ways to overcome these spatial limits and expand its data center capacity.

Optimus was originally expected to begin mass production much sooner, but it has not yet reached volume production in the thousands or tens of thousands of units.

This suggests that humanoid robot production goes beyond technology development and is running into unexpected obstacles in building the overall supply chain, including complex parts procurement, production process optimization, and quality control.

One of the biggest problems in humanoid robot development is the precise control of robotic hands. Critics point out that current robotic hands still lack the fine force control and complex dexterity needed to fully replace human workers.

The challenge is especially acute in tasks requiring delicate force adjustment, and this serves as a significant technical limit to robots fully replacing human labor. Even leading companies in the field admit that fully replacing the human hand is not easy.

Tesla has said it will sell the Optimus humanoid robot for $20,000 to $30,000, but critics argue this is far from a realistic mass-production price. For example, Hyundai's Atlas robot is reportedly priced at around 200 million won when produced at a volume of about 30,000 units.

To bring production costs down to the level Tesla targets, large-scale production on the order of automobiles, at least 500,000 to 1 million units or more, is essential. Given the current size and technical level of the robot market, price reductions through mass production are expected to take considerable time.

The ability to install the FSD Lite version in vehicles that are seven, eight, or even ten years old and use upgraded autonomous functions is widely viewed as a very positive signal.

This shows that Tesla has the capacity to improve autonomous driving performance through software updates regardless of hardware aging. Such technical progress is expected to raise the value of Tesla vehicles and let more users experience autonomous driving.

Tesla began developing its own compute chips across hardware generations 3 and 4, and currently outsources production to Samsung and TSMC. Over the long term, however, it aims to design and manufacture these chips independently.

At the center of this plan is the construction of Terafab, a large-scale in-house semiconductor plant. Through Terafab, Tesla hopes to reliably secure the high-performance chips needed for AI and autonomous driving and reduce reliance on external manufacturers to strengthen technological independence.

This is seen as part of a strategy to control not only software but also core hardware in-house, vertically integrating its entire technology ecosystem.

Tesla's large-scale data center plan has run into a major obstacle in the form of spatial constraints on the ground. As a solution, Elon Musk has proposed placing data centers in space.

This is an innovative idea to overcome physical limits on Earth and environmental issues such as water consumption and noise, and to expand computing infrastructure without limit as needed. Using the Starlink satellite network and SpaceX's rocket technology to build data centers in orbit reflects Musk's long-term vision.

Ultimately, Musk's remarks stress that physical AI is a crucial challenge with national destiny at stake, not just a matter of technology development, reflecting the view that such progress will shape the future of society as a whole.

AI companies have the resources for intelligence learning, namely vast data and computing power, but lack the medium through which to apply that intelligence in the physical world and accumulate more data.

In other words, to implement the intelligence AI has learned in the real world and collect additional data, a physical foundation such as a robot platform, automobile platform, or factory platform is essential. Korea's manufacturing sector has the capacity to provide such advanced physical platforms, making it an important partner for AI companies.