Target Allocation
Fixed Income 25%
Gold/Silver 10%
Crypto 5%
Cash 5%
Equity 55%
Fixed Income
The yields of my current holding rang from 6% to 8%. With the coming rate cut, expect lower yield in 2026. However, this yield is still significantly higher than those elsewhere in the world, like China and Japan.
Should increase holding on oil pipeline AMLP.
Gold/Silver
- Hedge to all uncertainties (Geopolitical, US dollar devaluation, AI burst)
- Looks at the 5 year chart of Gold and Silver. What a world we are facing???
Cash
BIL currently yield 4.19%, Apple cash only offers 3.75%
Crypto
NOT A HEDGE TO ANYTHING, not even dollar. Hugely volatile. Very sensitive to market liquidation. Still not quite sure what its meaning is for the portfolio but guess I need to tiptoe on it to get a taste.
Hold Bitcon and Ethereum.
Equity
No expectation for 2026. Goal is to avoid loss.
Big Tech
Currently in 25%, hold the position.
I am a firm believer of platform economy. A “platform” company taxes on everyone and every company. GOOGL, AMZN, MSFT, APPL and META are the platform companies. TSLA not yet is, NVDA definitely not.
But who will be the platform company in the AI era? Don’t know. APPL has the biggest potential, but it is making little progress. GOOGL may be a good contender, but far from sure.
A Platform
- Must be to “C”. Has hold of individual consumer.
- Has an Ecosystem to suck everyone in.
Don’t believe in rotation to mid, small cap. These platform suckers will keep sucking every blood of the economic.
AI Play
AI is still the biggest play.
- Scaling law has hit the wall (Ilya recently said, “We are moving from the age of scaling to the age of research.” Lecun Yann, as always, said LLM is dead-end.) AGI is NOT in sight.
- 2025 is supposed to be the year of AI agent, it obviously runs short of the expectation. We still don’t have even one “Agent” that can really work. The news just broke that Meta to acquire Manus. I would think that is more or less Zuckerberg’s “病急乱投医“ (“A drowning man will catch at a straw”) .
- But even with its current capability, LLM is still hugely useful. Inference Token demand is still exploding.
- AI infrastructure play will test its limit in 2026 (APPL and TECHY has set back their AI capex. Are they 春江鸭？Ducks that first know the river warms in spring.) 20% – 30% correction is possible. If it comes, it will drag every asset class down with it (Gold may be the only exception) ，along with the whole US economy. Trump’s mid-term election is also on stack. So, Trump would do everything he can to prevent that from happening, but just can he?
- NVDA is not a platform company and not for long term hold. It is still the best bet on AI. It can easily earn 200B in 2026, so even if market cap grows to 6T (current at 4.5T), the PE is just 30. But it would well be the peak at that point. It has grown to be my biggest single holding, takes 10% of the portfolio. Looking to harvest some gain in 1H of 2026.
- Beneficiaries are hard to come by, with the first round of the AI utilization. Don’t count on software. So just stick with the big suckers (the hyperscale).
- To the end, both AI token and Crypto are tied to electricity power. kWh is the future UM (unit of measure) of the economic. World currency will be based on kWh as well. GEV and LNG are my bet on the kWh.
