Problems to focus on in the AI era

24 Year Annual Topic: Service->Infrastructure->Device Cycle Deployment, Infrastructure-to-Device Expansion is Critical
Q2 Trends This Year Will Be Most Important Momentum This Cycle

The large flow of the cycle is as follows.
Services:Cloud
Infrastructure: AI
Device : (not yet)

23 years of service (AI, ChatGPT),
23-YEAR 24-YEAR INFRASTRUCTURE (NVIDIA) – AI will run to 30% across all servers in three years (currently 5%)
25-year device order (EX: Apple Vision Pro, etc. I don’t know what it will be, but it will be a leading item in the future)

In the past, it took two and a half years for an infrastructure called a data center to be built in the Internet era
-> If Nvidia cycle is conservatively scheduled for two years (there is still a year left considering that it started in May last year (NVIDIA, Hynix chain still has room to try)

The difference is whether to naturally expand to devices before the first half of next year or not because of the absence of special devices
In the end, what kind of KEY device comes out is the key to the new trend, and if it accidentally fits the consumer after various attempts, it will be a hit

Past Case >
In Internet Service Cycle, Cisco Stock Crashed, IT Bubble Smashed
The Wi-Fi cycle was also meaningless in the beginning
The era of Wi-Fi has arrived since the semiconductor revolution in 2003 increased the number of laptops from 2 hours to 8 hours
Since then, it has rapidly evolved into a laptop cycle (PC market)_The stock price has been off for a year and a half before that

Therefore, the point is whether the device will arrive by the first half of next year
Services and infrastructure, which were previously the leaders, will be maintained

Future Important Events Related to Devices >

1. WWDC Apple Developer Conference in June – Expect IOS18 to accept AI on a large scale (Apple should talk about AI now, timing shouldn’t be late, I think it’s Siri who has become very smart)
   PLITO (subcontracted by Apple, Korean company):
   Apple Direction Forecast
2. Reinforcement Learning – Recognizing and classifying situations by looking at pictures
3. Sensor-driven reinforcement learning – conversational patterns with temperature, room, and time
   Vision Pro is in high demand in China from Auto – virtual office and factory implementation
   There is also too much demand from people who live from home. If you use Vision Pro, it will be dramatically implemented as an office
4. Windows Large Update (4Q)
   Microsoft foretells large-scale updates, which is actually Windows 12, but it has copilot buttons
   Windows 12 is estimated to be changed due to AI with high probability
   From Microsoft’s perspective, it will move on to interactive AI -> ChatGPT paid version subscription, office copilot paid subscription
5. Azure (Cloud) 2. Windows 3. Office Subscription
   AI market share is increasing as a surprise for Azure, but the current situation -> device spread is important

Create demand for device replacement >
If AI-led devices become devices, existing PC specifications will be lowered too much, and the key is whether demand for PC replacement will appear
There is a case of a triple camera on a mobile phone. 1. Need to quality food photos on Instagram. 2. Need to quality up the YouTube shoot

Need to find beneficiary when NPU is added >
AMD, Qualcomm Will Evolve From Mobile To NPU On AP
There are no cars that only run NPU
If Co-pilot and ChatGPT enter the car, NPU-only companies will benefit unconditionally

LLW Memory -> Hynix is exclusively delivering to Apple Vision Pro
-> Wide IO says it increased the amount of data you touch at once
-> Amount of data (dram speed remains the same), faster response time
-> Since the response speed is very important from AI’s point of view
-> 20-30% higher unit price compared to mobile DRAM
-> Post-process_ Inspection equipment (tech wing), substrate (Daedeok, Simtech), probe card (TSE) for LLW

If you look at the sports team

1. Increase in utilization rate: Electronics, Materials (Hansol, Solve, Dongjin), Reno Industrial, ISC, ISU PETA
2. Post process_ Inspection equipment (tech wing), probe card (TSE) for LLW
3. On-Device AI Related: Intel, AMD, Qualcomm, MediaTek