Garrett Jin, identified by Odaily as the representative of the "1011 insider whale," said in a weekly market report that he has not changed his Bitcoin view and is still holding the position built around $60,000.
On gold, Jin said the metal had previously remained in an observation phase, but a recent breakout from its consolidation range has turned it into a long-term allocation opportunity. He said the investment case still rests on a longer-term weakening trend in the U.S. dollar.
Bitcoin stance unchanged after July low
Jin said that since Bitcoin approached a low near $57,700 in July, the market has continued to meet the conditions he looks for in a bottoming process. Based on that reading, he said there is no reason to alter his view, and he continues to maintain the position established around $60,000.
AI memory demand remains strong, but he is not chasing the rally
On AI memory, Jin said he had previously advised investors to gradually build positions in memory-chip names and buy on dips. The market, however, moved higher without going through the pullback he had expected. He said he has therefore sold half of his rebound position.
He stressed that the sale does not reflect any change in the underlying investment thesis. Instead, he believes the latest short-term advance was driven more by capital structure factors.
ETF-related risk has eased, but not disappeared
Jin said risks linked to Korean leveraged ETFs have not been fully released. He noted that JPMorgan said the related exposure had fallen 66% from its peak, while his own calculation puts the decline at about 38%.
According to Jin, the main reason for the gap is that net asset values have dropped, rather than investors exiting those positions. In his view, the leverage effect has weakened, but it has not disappeared.
Demand outlook extends into 2027
Even so, Jin said this should not be read as a bearish call on memory demand. He pointed to SK Hynix, whose 2026 capacity is already sold out, and Micron, whose orders extend through 2028. On that basis, he said AI memory demand is expected to remain strong through the second half of 2027.
At the same time, he cautioned that memory remains a cyclical industry. Related stocks have already risen by several hundred percentage points, and he said cyclical shares are unlikely to rely on valuation expansion for a sustained rally over the long run.
AI spending cycle is entering a new phase
Jin said the current AI investment cycle is moving from an early spending phase into a mid-cycle phase where returns need to be verified. Earlier, the market rewarded companies for increasing AI capital expenditure. Now, he said, attention is shifting to whether that spending can be converted into revenue.