@ishmohit1 Too early to predict AI application and service winners. Isn't it? They may not even exist. The industry is at hardware and software platform stage.
@itsTarH Yes. Silicon photonics is altogether a unique area and is little different form opitcal chiplets. The other related areas are PICs and optical engines. They are well protected by IP and has very high entry barrier. They are already relavent. Interesting times. 1/2
@gurjota Inequlity is not the concern. Minimum income is the issue to be tackled. How much more to earn is individual responsibility. The minimum remuneration is society's responsibility
@VijayKedia1 Data centre is a high capex high churn business. You need a combination of operational excellence, vendor/tech management and financial prudence. After everything, your profits are not great.
@suru27 It needs to be automated and system should run by itself. Define risk and control it a the higher level. It may need about an year to show the real results. Even more time
@ishmohit1 Here are the potential winners who can have high RoE
High revenue, high RoE:
Chip companies, basic hardware providers, platform software and security software companies
Mid Revenue, High RoE:
Power equipment, connectivity
Low revenue, High RoE:
Software maintenance and operation
@ishmohit1 The suppliers to building data centers are niche tech companies. If they control the ecosystem, then yes, they make around 30-40% RoE. Otherwise, very less.