On Wednesday (2026-08-26) US stock markets barely moved: SPY rose 0.02%, QQQ rose 0.09%. Over the past 5 trading days, SPY fell 0.39%, QQQ fell 0.66%. Outwardly calm, the main themes are actually hot: on one side NVIDIA again delivered results above expectations, on the other side trade and geopolitical frictions are heating up; the market is being pulled between “growth” and “defense.”

What you should watch today is not an isolated headline but two threads advancing at the same time. The first thread is that tech-driven growth remains very strong. The second thread is that trade and sanction risks have not eased; they are instead being ratcheted up. Putting these two threads together explains why the indices haven’t moved much, but sector divergence will become increasingly pronounced.

Start with the tech story. NVIDIA released its latest earnings: sales reached 962.2 hundred-million USD, up 106% year-on-year, and it has beaten Wall Street expectations for 16 consecutive quarters. Put simply, this isn’t a “barely pass”; it continues to push the market’s imagination about artificial intelligence higher. What the market cares about most isn’t just how much the company earned this quarter, but that management also provided large upside for next year’s sales. In plain terms, the company is saying orders aren’t fully digested yet and future demand could be larger than the market currently thinks.

Why does this matter? Because over the past period the market has been worried that big US tech has run too far and valuations might be too expensive. Valuation, simply put, is how high a price tag the market is willing to give a company. But if profits and revenues keep sprinting ahead, that high price tag becomes less scary. So capital will re-evaluate: if AI demand is still accelerating, the leaders along the chip, cloud, and data-center chain could realize profits faster than previously expected.

But it’s not that simple. At the same time tech tailwinds are appearing, trade friction is also intensifying. The trade dispute between the US and Canada continues to ferment; related reports bluntly point out this is not a small episode of “shouting first then backing down.” Meanwhile, the US is relaxing some beef import restrictions to push down consumer prices. This detail is crucial because it indirectly admits one thing: tariffs don’t just “hit others,” they also raise domestic prices at home.

The transmission chain is actually clear. Once tariffs and retaliatory measures are implemented, imported goods become more expensive, firms’ procurement costs rise, and consumers may pay more. With higher costs, inflationary pressure is harder to bring down. Inflation, quite simply, means things getting more expensive overall. If inflation proves sticky, the Federal Reserve will be reluctant to cut rates easily. Cutting rates reduces borrowing costs. If rates aren’t cut, borrowing costs remain high. Thus, rate-sensitive sectors, such as real estate, small banks, and smaller companies, will feel the pain more heavily.

At the same time, geopolitical risk is rising. The UK officially supports pushing Iran further toward “economic isolation,” and the US intelligence chief visited Moscow to convey a warning not to touch NATO members. Put plainly, this shows the West is not only verbally tense; policy and security signals are tightening in parallel. Escalating sanctions and conflict warnings will push part of the market into defensive assets like gold, and they will also raise uncertainty around energy supply.

So today’s main thread is not simply “technology is great” nor is it just “risks are large”; rather, it is the coexistence of a growth narrative and a risk narrative. NVIDIA pulls up risk appetite — investors’ willingness to take risk — while trade and geopolitical friction push up demand for safety. The result is that headline indices appear unchanged, but seats are being rearranged inside: some money keeps chasing tech, while other money moves toward more risk-resistant assets like gold and energy.

One additional point is worth noting. The market’s “currency-devaluation trade” has heated up again. The so-called “devaluation trade,” in plain terms, is driven by worries that government debt and fiscal pressure are too great and that currency purchasing power will be slowly diluted in the future, so investors seek assets like gold and bitcoin that don’t rely on a single government’s credit. This topic isn’t the sole mainline today, but combined with the geopolitical risks above, it strengthens gold’s appeal.

But there are countercurrents. On Wednesday gold-related ETFs pulled back significantly in a single day: GLD and IAU both fell 1.58%. That shows not all defensive assets move straight up in the short term. To use an analogy, the safety trade is climbing a mountain but will pause for breath; it’s not a daily sprint to the peak. Such retracements look more like consolidation after a rise and do not mean the main thread has immediately ended.

Next, look at several key directions given by ETF Radar, but first distinguish two things: below I’ll describe what happened over the past 5 trading days, and then the model’s forward-looking 1-to-3-month outlook — these are two different matters.

Start with QQQ. QQQ fell 0.66% over the past 5 trading days. The ETF Radar model is biased positive for QQQ over the coming period, but the signal is modest. Why not especially strong? Because short-term price action is a bit weak, but mid-term evidence hasn’t broken. First, Nasdaq futures in the CFTC data are overall net-short at 6.1771 ten-thousand contracts, but week-on-week the net short was reduced by 2.7354 ten-thousand contracts, an increase of 31%. CFTC positions, in plain terms, show how much large players are net-long or net-short in futures. Net-short reduction indicates some previous short bets are being pulled back. Second, since 2026-08-10, QQQ has seen net creations of 97.88 hundred-million USD, accounting for 2.02% of assets. Net creations, in plain terms, mean real money continues to flow into the ETF. Third, the prediction market’s yes_prob for another rate hike in 2026 is 0.46. yes_prob, in plain terms, is the market’s probability estimate of an event on a 0-to-1 scale; 0.46 is less than half. This reading is not very hawkish. Hawkish, in plain terms, means the central bank is more inclined to keep rates high. To be honest, QQQ has fallen over the past 5 trading days, but the model remains mildly positive, meaning it’s looking at the next 1 to 3 months and short-term volatility hasn’t fully negated the mid-term logic.

Next look at GLD. GLD rose 1.81% over the past 5 trading days. The ETF Radar model is positive for GLD going forward, with consistent signals. This ties directly to today’s main threads because geopolitical friction, fiscal worries, and safe-haven demand all transmit into gold. Quantitatively, first, CFTC gold net-long is 14.1648 ten-thousand contracts, up by 3986 ten-thousand contracts week-on-week and sitting at the 100th percentile of its own range. Net-long, in plain terms, means bullish positions significantly outnumber bearish ones. Second, GLD’s net creations since 2026-08-10 amount to 45.29 hundred-million USD, accounting for 2.91% of assets, and 30-day cumulative inflows equal 2.79% AUM. AUM, in plain terms, indicates how much money the fund manages; the larger the share, the more meaningful the inflows. Third, the US 10-year real yield fell to 2.32%, down 9 basis points over 20 days. Real yield, in plain terms, is the interest rate after subtracting inflation; when it falls, non-yielding assets like gold lose less relative disadvantage. Note a divergence here: GLD fell 1.58% in a single day on Wednesday, but it is up over 5 and 20 days and the model is positive, indicating the short-term pullback hasn’t overturned the mid-term evidence.

Third, look at KRE. KRE fell 0.56% over the past 5 trading days. The ETF Radar model is negative for KRE going forward, with consistent signals. This ties to the earlier transmission that tariffs may hold back disinflation and make Fed easing less likely. What do regional banks fear most? They fear high funding costs, weak loan demand, and continued pressure on the asset side. Quantitatively, first, since 2026-08-19, KRE has had net outflows of 3.90 hundred-million USD, about 9.4% of AUM; 30-day cumulative outflows are 7.36% AUM. Persistent outflows, in plain terms, mean capital is leaving the sector. Second, the put_call_oi_ratio is around 19.06, flagged as put_crowded. This metric, in plain terms, measures how crowded protective put positions are in options; a value this high indicates many are hedging against declines. Third, CFTC data show the 2-year Treasury net short is -1,243,004 and the 10-year net short is -2,229,013. Treasury net-short, in plain terms, indicates the market is betting rates won’t fall quickly. If high rates stick around, regional banks will find it harder to be comfortable. There is a small divergence here: KRE actually rose 0.34% on Wednesday, but it is weak over the past 5 and 20 days and the model is focused on the next 1 to 3 months.

Fourth, look at EWC. EWC rose 0.65% over the past 5 trading days. The ETF Radar model is positive for EWC going forward, with consistent signals. However, this doesn’t fully align with today’s main threads and in fact embodies a divergence that is itself interesting. News of rising US-Canada trade friction intuitively looks bearish for Canadian assets; yet the system’s mid-term readings remain positive. Quantitatively, first, since 2026-08-20, EWC has seen net inflows of 74 million USD, accounting for 1.07% of AUM; 30-day cumulative net inflows equal 5.00% AUM. Second, over a longer window EWC has gained 7.50%, outperforming EFA by 3.97%. Excess return, in plain terms, is how much it outperformed the reference index. Third, HY credit spread OAS is 2.7%. Credit spread, in plain terms, is how much more companies pay over Treasuries; this number hasn’t widened noticeably, suggesting the market isn’t pricing in large credit stress yet. Why might this be? Because energy has a meaningful weight in the Canadian market, and prior support on the energy side partly offsets trade pressure. So to be clear, today’s news is biased against Canada, but the model’s mid-term stance still leans positive; this divergence between headlines and system signals requires continued observation.

Putting these ETFs together completes the picture: QQQ represents that growth expectations aren’t extinguished; GLD represents ongoing defensive demand; KRE represents the most vulnerable corners under a high-rate environment; EWC reminds us the market isn’t one-dimensional — trade pressure and resource support can coexist.

The most important thing today is not to guess whether the index will rise or fall tomorrow, but to see the more real transmission chain inside the market: NVIDIA continues to strengthen tech growth, trade frictions may slow disinflation, and geopolitical and fiscal worries lift safe-haven demand, so the tug-of-war of “strong tech, strong gold, and pressure on rate-sensitive assets” continues. Reminder: the above are ETF Radar system quantitative model directional forecasts for general market conditions and do not constitute personalized investment advice; consult a licensed investment advisor before investing.