Weekly review: 12–16 October 2026. Snapshot: 17 October.
Executive summary
Our portfolio is up 15.06% year to date and 103.48% since January 2024, after a 10.02% loss last week. That is a serious setback, and it deserves the same scrutiny as our gains. Strong semiconductor demand continues to support our investment thesis, but our concentration leaves several holdings exposed to the same reassessment of AI spending, financing and valuation.
The week brought both sides of that argument into view. ASML and TSMC reported strong revenue, while concerns about financing AI infrastructure weighed on the sector. Our task is to test whether scarce capacity can translate into profitable growth, without assuming that good businesses are protected from large share-price falls.
The scorecard puts the setback in context
The figures below show our portfolio, the S&P 500 and the Nasdaq-100, in that order:
- Week: -10.02% / -2.00% / -5.63%
- October month to date: -8.66% / -0.10% / -4.27%
- 2026 year to date: +15.06% / +11.84% / +15.77%
- Since January 2024: +103.48% / +57.57% / +69.57%
We now trail the Nasdaq-100 by 0.71 percentage points this year, while leading the S&P 500 by 3.22 points. Our longer-term lead over both remains substantial. The weekly loss belongs in that record alongside the gains.
Concentration creates shared risk
Nine chip-related positions account for 64.77% of our displayed holdings by value. NVIDIA and ASML alone represent 25.08%. Although the portfolio contains 16 holdings, several depend on the same AI spending cycle.
This concentration gives us a direct stake in scarce manufacturing and compute capacity. It also means that a change in expectations can affect several businesses at once. A broad reassessment of customer budgets can reach across the chain, from equipment and fabrication to the chips themselves.
The number of stocks is therefore only part of the risk picture. We also need to consider how much of their future revenue depends on the same customers continuing to fund expansion.
Strong demand meets funding fears
US consumer inflation reached 3.7% year on year in September, while producer prices rose 1.3% month on month. Concerns about AI financing also resurfaced alongside selling in chip shares and buying of US Treasuries. Our reading is that the market is demanding clearer proof that large infrastructure bills will produce cash returns.
The earnings evidence was encouraging. TSMC reported third-quarter revenue of US$46.72 billion, up 41.2% year on year in US-dollar terms. ASML reported €11.7 billion in sales and said planned supply for 2027–28 would fall short of customer demand forecasts.
This supports our “Economics of Chokepoints” thesis: own the toll booths where scarce capacity gives suppliers bargaining power. That scarcity has physical and financial foundations. Building chipmaking equipment takes time, and ramping new production carries costs. TSMC’s gross margin fell 1.3 percentage points from the second quarter as its 2nm ramp weighed on profitability.
Strong demand supports the business case. The return we earn still depends on the price we pay, execution and customers’ ability to fund expansion. Customer demand forecasts can also change. We should assess them as forecasts, rather than treating them as assured future revenue.
Recent additions need different tests
Moderna shows an open-position gain of 16.34% and a 6.43% weight in displayed holdings value. Palantir shows an open-position loss of 3.12% and a 5.45% weight on the same basis. These figures measure gains or losses on open positions, not their returns during the review week.
The two positions extend our exposure into medicines and enterprise data, but they require different evidence. For Moderna, the focus is clinical progress and cash use. For Palantir, it is growth, cash generation and valuation. An early gain does not settle an investment case, and a short-term loss does not remove the need to test it carefully.
The week ahead: follow the cash and the shared exposure
We remain focused on orders turning into revenue, revenue turning into cash, and how much shared risk sits across the portfolio. Delayed customer spending or lower valuation multiples could create further losses even if demand keeps growing.
Our record includes this drawdown. The responsibility now is to keep testing the thesis against new evidence and make position size reflect both conviction and risk.
How to follow or copy this portfolio
If this approach fits your horizon and risk tolerance, review the holdings, performance and risks on Razvan Ceapa’s eToro Popular Investor profile, razvancd, then use Copy if you decide the portfolio suits you.
Razvan’s recommended starting amount is $500, not a platform minimum. Your entry timing, position coverage, execution and returns may differ.
Copy Trading is not investment advice | Capital at risk | Past performance does not guarantee future results