The Rules of Survival to the End: Which Stocks Survive?

The Rules of Survival to the End: Which Stocks Survive?

Does history repeat itself? (The dot-com bubble vs. the AI boom)

Opinion on today's AI market is evenly split between concerns that it is 'a second dot-com bubble' and claims that it is 'a structural shift of the Fourth Industrial Revolution.' What matters is that whether the bubble bursts or continues, there will always be companies that survive, and they will take the whole market.
 

When the dot-com bubble burst in 2000, countless 'dot-com' companies disappeared, but today's Big Tech was born from those ashes. Let's first dig into the secrets of their survival.

 

 

2. The DNA of dot-com bubble survivors 

The companies that survived when the bubble burst all had 'tangible infrastructure' and ‘an irreplaceable moat.’

 

① Amazon: "Connecting clicks to logistics"

  • Situation: During the dot-com bubble, it went through a crisis in which its share price plunged 90%.

 

  • Why it survived: It was not just a website. It poured enormous sums into building fulfillment centers and delivery systems (physical infrastructure). While competitors spent money on marketing, Amazon invested in the essentials: 'customer experience' and 'logistics innovation.'

 

  • Lesson: Companies that combine technology with real services/logistics to generate 'cash flow' survive.

 

② Microsoft: "The monopoly faucet called the OS"

  • Situation: It stood firm even amid the dot-com frenzy.

 

  • Why it survived: It was already earning huge amounts of cash through its powerful operating system, Windows. Internet or not, you needed Microsoft's product to turn on a PC.

 

  • Lesson: Companies with overwhelming market share and the character of an essential good (Utility) are not shaken by external shocks.

 

③ Cisco: "The pickaxe-and-jeans strategy"

  • Situation: It reached No. 1 in market capitalization at the time but was hit hard when the bubble burst. However, the company itself did not fail and remained a dominant player of the internet era.

 

  • Why it survived: It sold the 'routers' and 'switches' essential for using the internet. No matter who succeeded with internet services, the infrastructure had to be laid.

 

  • Lesson: 'B2B equipment/component' companies that supply infrastructure survive regardless of whether services succeed or fail.

 

 

3. Today in 2026: Survival conditions and risks in the AI market

Let's apply the lessons of the past to today's AI chip and talent war.

 

① 'Talent' is the 'moat' (The Talent War)

  • Current state: Global Big Tech companies such as Elon Musk's Tesla, Google and Broadcom are 'snapping up in advance' Korean semiconductor engineers, offering annual salaries of 400 million won plus stock compensation.

 

  • Implications: If the 'factory (Fab)' was the source of competitiveness in the past, design capability is now the core competitive edge. The movement of talent caused by the gap in performance bonuses between Samsung Electronics and SK hynix is not just a morale problem; it means an accelerating technology gap.

 

  • Conclusion: Companies that lose talent will fall behind, while companies that attract or retain talent (SK hynix, Tesla, etc.) will survive like 'Amazon.'

 

② From 'general-purpose' to 'custom' (New Infrastructure)

  • Current state: Big Tech companies dream of escaping NVIDIA's GPU dominance. Demand for chips tailored to their needs, namely 'custom HBM' and 'application-specific integrated circuits (ASICs)', is exploding.

 

  • Implications: Just as Cisco laid down routers in the past, today companies that design and build 'custom AI chips' play the role of essential infrastructure.

 

 

4. Final investment strategy: Finding the next Amazon/Microsoft

Here we summarize the sectors and stocks that will survive even if the bubble bursts, or that will take the lead when the market is reshaped.

 

1) Irreplaceable infrastructure (the next Cisco/Microsoft)

The faucets and roads of the AI era are 'HBM' and 'foundries.'

 

  • SK hynix: It currently leads the HBM market and has been successful in retaining talent with high compensation. Like 'Microsoft' in the dot-com era, it occupies the position of an essential-goods supplier.

 

  • Hanmi Semiconductor: It supplies equipment essential for making HBM (TC bonders). It is a gateway that chipmakers must pass through.

 

2) The designers of custom chips (the next Amazon - an innovation model)

Big Tech companies want to make their own chips, but they cannot do all the design themselves. The 'design houses' that fill this gap are key.

 

  • Gaonchips / ADTechnology: They are the bridge connecting Samsung Electronics' foundry with fabless (design) companies. With the technology to optimize complex AI chip designs, they are the biggest beneficiaries of the 'custom chip' trend.

 

3) Monopoly power in future technology (IP companies)

  • OPENEDGES Technology / Chips&Media: These companies license semiconductor designs (IP). Since they do not make chips themselves, they have no inventory risk, and their structure means royalty income grows as AI chip development increases (operating leverage).

 

 

[Summary] The lesson of the dot-com bubble is clear. "Expectations without substance disappear, while companies with infrastructure and proprietary technology survive."