Themes series: we take a market trend that global analysts are talking about, retell it with the source credited, then extend it with our own Thai-market + RS lens. This round's source: the Business with Brian channel — [watch the original](https://www.youtube.com/watch?v=M8_DnpYV2hY)
When most people picture AI, they picture giant data centers. The point Brian raises here is the interesting one: the next chapter of AI isn't in the data center — it's in the devices around you — your phone, your car, security cameras, robots, factory machines. This is Edge AI: the "brain" runs on the device itself, with no trip to the cloud — faster, more private, always on.
What he covers in the clip
- AI's two waves — the first is the data-center build-out (growing ~25%/yr, spending heading toward ~$5T by 2030). The second is Edge AI chips running AI on real devices — which Brian argues grows faster than the data center (30%+/yr). - "Whoever wins the edge wins the AI race" — he anchors on quotes from chip CEOs (Jensen/NVIDIA, Cristiano/Qualcomm). - Companies that "own" the shift (US names): Ambarella (vision chips for cameras/robots, ~80% of revenue is Edge AI), Mobileye (EyeQ chip in 230M+ vehicles), Cohu (chip-test equipment, orders +57% YoY), Calix, and NVIDIA (profit up 40× in 6 years: <$3B → ~$120B, >70% margin). - A point I like: he flags that many are pure-plays still investing hard for growth — profits come later.
Note: the names above are US stocks the source discusses. We cite them to explain the trend, not as buy calls.
Our view: how Edge AI connects to Thai stocks
(This section is MOEasymmetry's own extension — the clip is about the US; we're the ones connecting it to Thailand.)
Thailand doesn't design AI chips — but Thailand is the manufacturing base for the supply chain Edge AI needs: printed circuit boards (PCB), chip assembly/test, and power. As AI moves into more cars, cameras, robots and devices, the number of chips, boards, test cycles and power components rises across the whole chain — and that's exactly what Thai electronics does.
| Edge-AI layer (global) | Where Thailand sits in the chain | Related Thai stock (current RS) |
|---|---|---|
| Vision / automotive (Ambarella, Mobileye) | Automotive PCB | KCE (RS 95) |
| Chip test / assembly (Cohu) | Assembly–test–sensors | HANA (RS 92) · CCET (RS 91) |
| Power for AI/EV/edge devices | Power management / EV | DELTA (RS 93) |
| End electronics manufacturing | EMS / device build | SMT (RS 98) · CCET |
The notable part: this group is already leading on RS — KCE 95, DELTA 93, HANA 92, SMT 98 — not names that were just talked into a move, but ones whose price/relative strength was strong first. That fits our principle: **the theme says "why," RS says "which / when."**
How we use it as RS investors
We don't chase a theme on emotion — we use it as a map, and let RS + chart structure decide the timing:
1. Screen with RS — within Thai electronics, the names with strong RS building clean bases (KCE/DELTA/HANA) are where institutions are actually accumulating. 2. Wait, don't chase — enter on a breakout from a base + holding above the 50-EMA, not at peak hype (that's when it's expensive). 3. Remember we're the supplier, not the IP owner — Thailand gains on volume, not on owning the technology — and strong RS means the market has priced in some good news already, so the stop must be clear: when you're wrong, you're out.
In short: Edge AI = why Thai electronics is interesting · RS = which ones are real · the stop = the discipline that keeps you alive.
Source: a summary of points from The Last EASY Trade Before It Explodes — the Business with Brian channel — [watch the original](https://www.youtube.com/watch?v=M8_DnpYV2hY) · we summarize and extend it with our own Thai-market + RS view, not a word-for-word translation; the figures/US tickers reflect the time the clip was recorded.
This article is for education and research, not investment advice · RS is current and changes · manage your own risk.