Mill Street Research

@millstreetresearch.com

Mill Street Research strategist Sam Burns, CFA, provides proprietary institutional research & tools on asset allocation, stock selection and the economy.

Slightly odd day today in markets. Equity indices are currently flat to down. Quite unusually, oil is also near flat today. Bond yields are roughly flat, up a bit from their latest low. But gold and silver have surged 4% today. Suddenly everyone's on vacation now except precious metals traders.

The capex boom is mostly a US story, as projected capex relative to sales for Europe and Asia Pacific do not look anything like the North America (mostly US) capex/sales data. Which helps explain why projected free cash flow has grown slower in the S&P 500 than for ACWI Ex-US since early 2025.

BildBild

Big buying spree going on in the cap-weighted indices, led by Tech, but not particularly broad. Palantir earnings giving a boost (PLTR +25%), and Semis +5% Only 4 S&P 500 sectors up, 7 down right now. The Russell 2000 is gaining as well, likely helped more by falling bond yields and oil prices.

Still a tough market to trade due to the ridiculous volatility in geopolitical and regulatory headlines. Treasury giving war updates while also intervening in currency markets. War has been both over and getting bigger dozens of times (pity the oil traders) Tariffs all over the place.

The Mag 7 is back, with the Bloomberg Mag7 index up nearly 4% today and up almost 10% in the last 3 trading days. This is lifting a lot of the Tech/AI complex and all the major indices today, while the low-beta sectors and Energy are lagging.

This economist article summarizes much of the skepticism of the longer-term ROI on the massive AI capex. That skepticism is (IMHO) why stock prices have not (so far) responded as positively to strong Q2 earnings as they did to strong Q1 earnings, and big AI-tied companies have been so volatile.

Duncan Weldon@duncanweldon.bsky.social · 2d ago

‘A rough calculation finds that covering AI capex through identifiable AI income requires revenue on the order of $2.5trn per year, more than tech’s entire combined revenue today (and far higher than what would have been needed a year or so ago, when capex plans were more modest)’.

For the eleventy-seventh time, headlines about talks with Iran are pushing oil and bond yields down, and non-Semiconductor stocks up. S&P 500 pushing back up to the high end of its trading range since May. Russell 2000 also up, but struggling to keep up ever since the late June index rebalance.

Looks like some month-end buying coming in as indices are rebounding from the early pullback. Amazon doing a lot of the work in the SPX and NDX, along with Alphabet, NVIDIA, and Microsoft. Apple still the big drag on the indices. Materials sector down -2.5%. Broad market roughly flat.

Some mild drama in stocks at the start as the big up open in Semis/Tech has faded pretty quickly. Small-caps are lagging notably, Russell 2000 -0.75% while NDX is +0.5. Semis still up 1%, but down -4% from the opening high.

Yikes, 10-year bond yield jumping sharply this morning, now at 4.73%, very close to the Jan 2025 peak. Partly driven by oil prices rising again, and partly the continued reaction to the Fed. Two dissenters from Wednesday's meeting who wanted higher rates have put out statements today.

Positive earnings news from AMZN and a dramatic rebound from leverage-driven selling in chip stocks in Korea is the good news this morning, while higher oil prices and bond yields are the bad news. So NASDAQ futures are up while S&P futures are about flat. Tech leadership back today.

Good reminder by @gregdaco.bsky.social that spending is being financed more by non-income sources, with a significant chunk coming from financial market returns (which are not included in personal income figures). The economy is more dependent on the stock market than in most past cycles.

Gregory Daco@gregdaco.bsky.social · 6d ago

⚠️Real disposable income rebounded in June, but the income/spending gap has widened over past 2 years: Spending is increasingly financed by savings, credit and wealth...

Big day for Microsoft after its strong earnings report. It is up 16% and thus adding about 77bps to the S&P 500's daily return all by itself. That's about half of the S&P's 1.5% gain today. Micron, Amazon, NVIDIA, and AMD contribute most of the rest of the index's gains.

So the macro data today was decent, not likely to cause big changes to the overall view. But now, no one is sure what the Fed will think of it, since Warsh has decided to be coy. June core PCE +0.1%, a tick better than expected, putting Y/Y core PCE at 3.3%. Improving but still too high.

Market's "reaction function" seems to be that Warsh did a very poor job in communicating today. Tech, Industrials, and Financials all down sharply, major indices closing at their lows down -1.5 to 2.0%. Only Energy up meaningfully due to oil. Long-term bond yields up notably, 10-year at 4.67%.

The markets are settling on a more negative view of the Warsh Fed meeting, with 10-year bond yields going out near their highs and stocks now trading at lows of the day. S&P 500 down more than 1%, same with Russell 2000 and NASDAQ.

Fed said no rate hike today, but had 3 of the 12 voters dissent in favor of a rate hike (Hammack, Kashkari, and Logan). This was what most people expected, but had unusual uncertainty around it. Press conference at 2:30 could be interesting.

Looks like correlations within stocks are ticking up today, with the CBOE's Implied Correlation index (COR1M Index on Bloomberg) trading above 10 for the first time this month. Dispersion index thus starting to pull back from recent historic extremes. VIX up to 20 now, VIXEQ at 48.

Bild

Bond yields up, following oil going back up again (isn't this fun?), while stocks are down broadly today. Semis getting hit hard again as the leverage unwind continues and worries about future capex have grown. Rest of the market doesn't like the oil/rates move and slightly worried about the Fed.

Fed day today, though earnings and AI are still much bigger overall market impacts. Most expect no rate change but potentially hawkish comments, and less info/guidance from Warsh. Market pricing 35% chance of a hike today, and 100% chance by September.

At the macro level, this is a lot of why earnings are so strong late in a cycle, and why core inflation is still above target: Fiscal policy is still unusually loose based on the deficit/GDP ratio, and it is quite rare for it to be this loose while the unemployment rate is this low.

Bild

Short covering has been a significant factor this year, and particularly in recent weeks. The Bloomberg US Short Interest long/short factor (based on days to cover) has had an unusually large jump, indicating heavily shorted stocks outperforming. Pattern looking like 2020-21.

Bild

Starting off today with big selling in Semis, and more factor rotation. Value, low-volatility, and high short-interest stocks are all outperforming. Momentum, liquidity, and earnings revisions factors are all negative today.

NVIDIA the biggest drag on the S&P 500 by far today, pulling it down 35bps by itself with its 4.5% decline, keeping it flat. News of more circular financing with OpenAI is worrying investors, so they are rotating into mostly non-AI areas, including AAPL, and some recent losers like GOOGL and MSFT.

Which act of the story are we in? The AI drama remains the biggest driver of equities and to some degree the economy (huge fiscal deficits and a lax Fed are the others). Lots of debate about AI itself and which companies will profit near-term vs long-term as the story evolves . . .

Starting off with rally in small-caps/broad market, and selling in Semis. Oil is down on another Lucy-with-the-football TACO from Trump over the weekend, helping bond yields ease. Huge number of earnings reports coming out in the next two weeks.

Low-volatility stocks doing well today, outperforming high-beta stocks by a good margin. SPLV +0.9%, SPHB -0.4% Value (+1%) outperforming Growth (+0.2%) today as well. And high short-interest stocks outperforming according to Bloomberg's factors.