When I glue all these notes together my conclusion is that the burden will be on the bulls until we speak next week (the trader’s perspective is fade the rally). There is always an if; my if is NVDA — “sell on the news” is what the street says, but in this AI frenzy people are paying crazy forward multiples of sales for new IPO’s such as Cerebras, and NVDA has the electricity to send this game into overtime. Your best bet is to sit this one out until Friday and enjoy the show — let the market show us what’s going on. We don’t always need to get front row seats.

There’s a moment in every cycle when the bond market stops being background noise and becomes the story itself. We are in one of those moments. The G7 government bond yield complex just printed its highest reading in more than twenty years, and Dr. Torsten Slok at Apollo captured the diagnosis cleanly in The Daily Spark this week: this is not a one-factor move. Four forces are pulling in the same direction at the same time — renewed inflationary pressure from elevated energy prices as the Middle East conflict disrupts global oil supply; persistently large government deficits requiring ever-increasing bond issuance; the end of central bank quantitative easing, with the Fed balance sheet potentially shrinking; and investors, finally, demanding higher term premiums and higher inflation premiums in a world that is visibly deglobalizing. None of those four are close to resolution. As Slok puts it, the era of artificially suppressed yields appears firmly behind us — rates will stay higher for longer, and investors should plan accordingly.

You can see the framework playing out tick by tick. Treasuries breached three levels at once into Friday’s options expiration: the 2-year cleared 4.00%, the 10-year cleared 4.50%, the 30-year cleared 5.00%. The breakouts were not just a U.S. story. Japan’s 10-year JGB printed its highest yield since 1997. The UK 10-year reached levels not seen since 2008. France’s 10-year took out its 2009 high. When the entire G7 long end moves together like this, it stops being a national story and starts being a regime story. The proximate catalysts were easy to inventory — WTI crude up roughly 10% on the week to about $105 a barrel, no real progress on Iran, hotter-than-expected inflation readings, and the formal handoff to a new Fed Chair. But the underlying engine is what Slok described. The bid that suppressed yields for a decade is gone, and the new buyers want to be paid.

Layered on top of all of this is a piece of geopolitics that did not deliver. The Trump–Xi summit took place over the last forty-eight hours, and the market had been hoping for at least a gesture from Beijing on Iran in exchange for movement elsewhere. It did not come. The only deliverable, by most readings, was soybeans. There was also visible tension on Taiwan that the desk should not ignore — a chip-funding decision is due from the White House inside of a few weeks, and Xi was more assertive on Taiwan posture than he has been in prior meetings. Wharton’s Jeremy Siegel, on Closing Bell, summed up the market reaction with a number he is not known for using lightly. Asked to grade the summit on a 1-to-10 scale, he gave it a minus two. He pinned the oil pop and the yield backup directly on the disappointment. He also did the back-of-envelope math everyone on the energy desk has been quietly doing: if there is no diplomatic action, WTI rises roughly $2–3 a week and gasoline rises 2–3–4 cents a week, which is not the kind of arithmetic any administration wants compounding into the midterms.

Siegel’s broader argument is worth pulling through, because it is the most useful framing of the current tape that I have heard this month. He is not bearish. He is, in his own words, still bullish on the overall market. But he wants the froth off, and he is unusually direct about where the froth is. The rest of the market, he points out, is trading at roughly 17 times forward earnings, which is reasonable. The chip slice, by contrast, is being priced as if its current rate of earnings growth extends to infinity. That is a category error. About 90% of a stock’s value, in a clean DCF sense, is earnings more than two years out. If you double earnings for four years and they then revert, you only earn a roughly 20% increase in fair value on the original purchase. Semiconductors have always been cyclical. AI is a real and powerful secular force layered on top of that cyclicality, but the layering does not erase the substrate. He stops well short of calling it a bubble. He calls it irrational, which is a different and more useful word.

He had the sharpest line of the segment on the Fed. Bond yields, in his read, are the market sending a message directly to incoming Chair Kevin Warsh: do not cut quickly into this. There were three dissents at the last meeting before the handoff. CPI and PPI are running hot. The 10-year, in Siegel’s reading, is at 4.60% — and as he put it, you can never ignore the ten-year in any stock market calculation. He expects Warsh’s first meeting to be a fight to defend an easing bias, with the most likely outcome a move at least to a neutral bias. He does not expect a hike in June. But the market that was pricing seven cuts at the start of the year is being asked, in real time, to reprice closer to one or two.

Ed Yardeni, sitting alongside him, took the constructive side cleanly. He was not disappointed by the summit because he never expected anything from it. He pushed back on the idea that the rally is unsustainable: this is, in his memorable phrase, an “earnings-led meltdown.” Multiples have not stretched relative to earnings because earnings went to record highs. The forward P/E actually came down through Q1. That is why the bounce off the March 30 low was so violent — when the denominator surprises to the upside, the cheapness shows up in the math even if the price chart looks vertical. He frames the AI build-out as a generational story. Beyond land, capital, and labor, he argues, data is becoming the fourth factor of production. Yardeni concedes the tape may chop for a few weeks, possibly a few months, but the trend remains intact. He left his S&P 500 target unchanged at 8,250 for calendar 2026 and 10,000 by the end of the decade.

Tom Lee at Fundstrat held the same year-end SPX target as Yardeni and is unambiguously bullish on tech — but he is the most explicit of the three about near-term tests the market needs to clear: (1) how the Fed reacts to underlying inflation risk; (2) the market’s dislike of the combination of rising inflation concern and rising yields, with PPI in his read already signaling pipeline inflation that is shock-driven and energy-driven rather than structural; (3) IPO supply, which is heavy and growing. His most counterintuitive call was on DRAM: in a world worried about inflation and an energy shortage, the memory shortage is bigger, and DRAM ends up trading as defensive in tough macro because people buy what works regardless of regime. The data points he cited are the kind that get attention — the SMH up six straight weeks, NVDA up seven straight days, the DRAM ETF going from $77 million in AUM a month ago to $6.25 billion and then $9.5 billion in successive sessions, the QQQ now needing to fall roughly 12.5% just to retest its 50-day moving average.

Sitting beneath all of this is a market structure issue that explains why the moves both up and down have been as violent as they have been. Options volume has surged this month — seven of the biggest sessions of all time have occurred in May alone, including Thursday, which saw volumes top 49 million contracts. Saxo Bank’s Koen Hoorelbeke captured the mechanic in one sentence: when equity options positioning tilts this far toward calls, market makers typically carry elevated short-gamma exposure, meaning they are forced to buy on rallies and sell on declines, amplifying directional moves in both directions. That is not opinion. That is plumbing. And it is why a single FOMC headline, a single Iran cable, or a single NVDA earnings line can move the index the way it has been moving. The pipes are loaded.

Speaking of NVDA — that earnings print is the next event on the calendar, and it lands Wednesday May 20 after the bell. Expectations are high because the stock has run so hard. The company will almost certainly deliver strong results. But the asymmetry now sits on the other side of the trade. The most plausible reaction, regardless of the print, is a sell-on-the-news move in the chip space — though anything can happen in this environment. Siegel’s “irrational” framing and Lee’s “consolidation later this summer” framing converge on the same conclusion from different angles: the chip slice has earned a rest.

A few things to watch alongside it. The equal-weight S&P 500 versus the cap-weighted index is the cleanest single tell on whether the rest of the market is being left behind. Within the equal-weight, the heavily weighted cyclicals — financials and industrials — are the ones to watch for any signal that growth is broadening rather than narrowing. The second is the participation rate inside the most recent jobs report. April unemployment stayed at 4.3% even as jobs growth topped 100,000, and the rate did not fall because participation dropped a tenth to 61.8%. Falling participation paired with steady headline unemployment is usually how an economy quietly tells you it is slowing. The third, back to Slok, is the curve itself. If 30-year yields hold above 5% and the G7 long end holds these new highs, the regime change is real and the playbook needs to reflect it. Sticky inflation, higher yields, and narrow leadership argue for a far more selective market from here.

So where does that leave us. The macro is constructive and the bond market is not. Earnings are real and the chip slice is overheated. The Fed is in transition and the term premium is back. The summit disappointed and the options pipes are leveraged. None of these things are individually fatal. All of them together demand respect. Slok’s frame is the right one to carry forward: four forces, no resolution, rates higher for longer. Plan accordingly.

This week marked the first meaningful sign that markets may be transitioning from “macro fear acceleration” toward “selective stabilization and tactical risk re-engagement.” Geopolitical influence: elevated but less dominant. Macro influence: still the primary driver. Trend quality: improving, but not fully confirmed.

The environment increasingly favors tactical flexibility, relative strength analysis, shorter-duration positioning, and awareness of violent countertrend risk. Current leadership structure: agriculture remains strongest, metals attempting stabilization, energy volatile but improving, defensive macro positioning beginning to unwind.

The market is no longer purely liquidating risk — it is beginning to determine which sectors can sustain recovery momentum and which remain vulnerable to renewed macro pressure. Still dancing — close to the door.