The competition in AI is getting more aggressive.

OpenAI has told Cursor that access to its models will be removed from the platform over security concerns, and I expect moves like this to become more common as the major labs fight for control over distribution, developers, data, and ultimately the AI ecosystem itself. The next phase of AI competition will be about much more than who has the best model. It will increasingly be about who controls the platforms, infrastructure, and customer relationships around those models.

At the same time, the market just worked through 2 major catalysts: Nvidia earnings and Kevin Warsh’s first Jackson Hole appearance as Fed Chair. So far, both have been relative nothing burgers for the broader market. We have had Nvidia earnings, Jackson Hole, major AI developments, and yet none of it has been enough to push the indexes decisively out of their current range.

The Nasdaq has effectively gone sideways since early May, giving us nearly 4 months without a sustained directional move. There have absolutely been opportunities underneath the surface, but this has not been an easy market where you can simply buy beta and let the trend do the work. In a choppy environment, trying to trade every small move is one of the fastest ways to burn (mental) capital and overtrade.

For now, this remains a stock-picker’s market rather than a broad trending market.

The last month of price action has been anemic. There is almost no sustained momentum in either direction, and from an index perspective, there simply has not been much to talk about.

The S&P 5ßß broke out of the multi-month consolidation that ran from early May through late July, but the breakout has produced very little follow-through. Normally, after spending several months building a base, you want to see price expand quickly after the breakout. Instead, the index has mostly drifted sideways near the highs. This is still positive but not enough conviction to turn it into a meaningful momentum move, at least not yet.

September is here.

Historically, September has been the weakest month of the year for the S&P 500, with an average decline of 0.8% since 2006. Of course, is not a prediction but serves a general guide and it’s important to treat it with enough context.

The worst Septembers historically have usually started from a much weaker market. 9 of the 10 worst Septembers on record came when the S&P 500 was already negative for the year heading into the month. That is very different from what we have today.

The S&P enters September up roughly 12% YTD, sitting close to record highs and comfortably above its 200-day moving average. Historically, September outcomes have been considerably better when the market enters the month with a positive performance rather than already being under pressure. Strong markets tend to stay strong until the data gives you a reason to think otherwise.

We’re set to close out August with the S&P already up double digits YTD, and historically that has been a very positive sign for the market going-forward.

Since 1950, there have only been 20 years where the S&P entered September with a double-digit YTD gain. In just 1 of those 20 years did the market suffer a double-digit drawdown during the rest of the year.

So yeah, September can still be volatile. We could easily get a 3%, 5%, or even larger pullback along the way. But for now September does not look as bad as many might think.

The S&P 500 has now gone 22 straight trading days without falling 1% or more. At the same time, the VIX has stayed at or below 16 for 18 straight days. On Friday, it closed around 14.4, its second-lowest level since December 2025 and about 20% below its 1-year average.

The market is surprisingly calm. That can become a risk by itself. The 12 to 14 range has historically been near the bottom for the VIX. When volatility stays this low for too long, fewer people hedge, dips feel harmless, and everyone starts expecting the market to keep moving higher. The problem comes when something suddenly changes. Investors rush to hedge at the same time causing volatility to rise suddenly and turning seemingly small moves into much bigger ones.

And we are now entering a time of year when volatility has historically started to rise.

earnings right now are not just strong. They are historically strong.

S&P 500 EPS guidance momentum has climbed to its highest level since Bloomberg Intelligence began tracking the data in 2011. It is the highest reading on record.

Companies are raising expectations faster than they are cutting them. We’re in an environment where earnings estimates are moving higher and fundamentals are giving investors very little reason to question the broader bull case. That is exactly what you want to see in a healthy bull market.

At some point, expectations can become so strong that simply beating estimates is no longer enough. Companies then need to beat by a wide margin, raise guidance, and show that growth is still accelerating. The higher expectations move, the smaller the margin for error becomes.

We are not there yet. Right now, earnings revisions are still moving in the right direction, and that remains one of the strongest arguments for staying constructive on the market.

e’rebeOf the 465 S&P 500 companies that have reported Q2 results, 69% beat revenue expectations. Among companies issuing Q3 guidance, 64% came in above consensus, the highest share since the post-COVID recovery period. W now looking for roughly 27% adjusted earnings growth in 2026, which would among the strongest fundamental backdrops in more than 50 years. And I do not see anything in the near term that looks strong enough to derail it.

Earnings are justifying this massive move higher in concentration. The top companies in the world are earning more than they ever have before.

But it’s not just the largest companies in the world. Earnings are rising across large caps, mid caps, and small caps at the same time.

Small and mid caps are still dealing with tighter financial conditions and higher sensitivity to rates, so seeing their earnings move higher alongside large caps is especially important. So, it is likely that fundamentals is spreading beyond the biggest AI and technology names.

The bigger picture still looks healthy, but this remains a difficult trading environment.

The major indexes have not broken down. S&P 500 is still near its highs, and the Nasdaq continues to hold its key moving averages. That’s is still positive. The problem is underneath the surface momentum has faded, there is little volume, and large parts of the market is still moving sideways.

Right now we are seeing rotation rather than broad deterioration.

Semiconductors and many of the former AI infrastructure leaders have weakened, but that capital has not disappeared. It has rotated into software, cybersecurity, fintech, biotech, crypto-related names, and a few other pockets of strength. That rotation has kept the major indexes supported and prevented weakness in the crowded AI trade from developing into a much broader correction.

Semiconductors remain the biggest technical concern. The SOXX semiconductor index continues to make lower highs and its structure is clearly weaker than the broader market. There is no reason to be concentrated there while relative strength is deteriorating. Leadership changes, and there is no reason to keep forcing exposure into the groups that led the previous move simply because they worked before.

In fact, some of the most interesting opportunities are starting to appear elsewhere. New groups begin to strengthen, new relative-strength leaders emerge, and individual stocks start separating themselves even while the broader market remains stuck. That is roughly where we are now.

For now, patience matters more than activity.

There are periods when you should press the gas and periods when preserving capital matters more. For weeks, I have been arguing that this is not the environment to aggressively increase exposure, and my trading activity reflects that. I would rather avoid burning mental and financial capital trading a sideways market and have both available when momentum returns.

The main macro catalysts I am watching are a resolution or meaningful de-escalation in the Iran conflict, lower oil prices, and lower Treasury yields. Any combination of those could remove pressure from financial conditions and help restart risk appetite.

But the most important signal remains earnings.

This is the one thing investors can track quarter after quarter without needing to predict every geopolitical headline or short-term market move. Watch actual earnings growth. Watch forward estimates. Watch revisions. As long as corporate earnings continue rising and analysts are not meaningfully cutting future estimates, the fundamental foundation of the bull market remains intact.

The walls of worry will always be there. There will be another geopolitical scare, another political headline, another inflation concern, another reason to sell. That is normal. Bull markets rarely move higher in a straight line with everyone feeling comfortable.

So I am still bullish, just less aggressive.

Nvidia’s latest quarter was another reminder that the AI boom is still very real.

Revenue was $96.2 billion in Q2. That is more than $1 billion per day. And to put things into perspective, Nvidia now makes more than 3x as much revenue in a single quarter as it made in the entire fiscal year 2022.

But the profit growth is even more insane. Operating income reached $63.7 billion, up $10.2 billion from the previous quarter. That increase alone is bigger than the $10 billion Nvidia made in operating profit during all of fiscal 2022. And it’s happening at a much bigger scale than before.

Data center revenue came in at around $89 billion, up more than 100% from a year ago. It’s pretty obvious now that almost all of Nvidia’s business is now tied to AI infrastructure.

And to top it off, Nvidia also guided for $108 billion in revenue next quarter. What makes that even more interesting is that Nvidia is assuming basically no data center compute revenue from China in that guidance.

So China could still add upside. Even relatively small H200 shipments can move the numbers. If Nvidia sells around 20,000 H200 chips at roughly $30,000 each, that is already about $600 million in revenue.

Even with zero China business, Nvidia is still expected to generate $412 billion in revenue in 2026, $718 billion in 2027, and nearly $1 trillion in 2028. If those estimates come close to reality, we are watching one of the fastest expansions of earnings power ever recorded. No other company in history has combined this level of growth, margins, pricing power, and market dominance at this scale.

And even after this incredible growth over the last few years, Nvidia is trading at its lowest forward PE in over 10 years. The last time was in May 2016 and we all know what happened since then.

But what’s even more important is that Nvidia is still supply constraint. Jensen said customer forecasts suggest demand could support around 100% growth next year IF they had enough supply. That’s why they expect closer to 70% growth because it simply cannot keep up with demand.

Demand is still not a problem right now. Unsurprisingly, supply continued to be the key issue. Nvidia is still trying to get enough chips, memory, packaging, and other components to keep up with what customers want.

That pretty much kills every bear argument for an AI bubble, at least from the demand side.

The only slightly concerning point this quarter was gross margin. Nvidia expects margins around 74% next quarter, and management said they could fall to around 71% to 72% before improving again.

The reason for that is memory pricing. We’ve seen how steep the increases in memory prices have been especially for HBM (high-bandwidth memory). Although Nvidia has already announced price hikes across the board and can pass a lot of those higher costs on to customers, but probably not all of them right away. So, Nvidia is taking some cost pressure. That’s why margins may come down a few percentage points. Still, margins above 70% are extremely strong for a company selling hardware.

Another interesting part of the quarter was Nvidia’s supply commitments.

Those commitments jumped from around $119 billion last quarter to $279 billion. That’s a massive increase. Nvidia is basically reserving future chip capacity, memory, packaging, and other key parts of the supply chain well ahead of time. Since supply is the key issue, procuring as many resources as possible is the best antidote and gives it much better visibility into future production.

Plus, it also puts pressure on competitors. Google, Broadcom, MediaTek and other companies building custom AI chips all need many of the same things Nvidia needs. They need advanced manufacturing capacity, HBM, packaging, networking, and power.

There is only so much supply available. So if Nvidia is still growing around 70% while already locking up huge amounts of future capacity, some of the more aggressive estimates for competing AI chips may be too high.

Of course, there are still real risks.

AI spending could slow eventually. Companies could build too much capacity. Nvidia’s valuation could still be too high. But the current business numbers are not confirming any of those. They are showing the opposite.

Revenue is still growing at a massive rate. Data center demand is still exploding. Customers want more chips than Nvidia can currently supply. Nvidia is spending huge amounts to secure future production. The company also returned $26 billion to shareholders in the quarter while doing all of this.

We might now be early in the AI buildout, but this is certainly not anywhere close to the end.