WEEKLY | June 30, 2026 | Macro Insights Queen
Hi everyone, last Friday’s PCE print was the test we’d been waiting for. It didn’t break the thesis. It sharpened it.
📌 The Regime: Mid-Cycle Reacceleration Is Tipping Into Overheat
Nothing about the framework changed today. What changed is how much evidence is now sitting on top of it.
May core PCE landed in line with consensus at 0.3% MoM, but the trend underneath it did not look benign. Core PCE has now run at 0.3% or higher for three straight months, including 0.4% prints in both January and February. The year-over-year rate has climbed from 3.0% in December to 3.4% in May, the highest since October 2023.
Supercore, the Fed’s preferred read on underlying services inflation, printed 0.5% MoM. Core capital goods orders rose 1.4% MoM, another confirmation that the AI capex cycle is accelerating, not slowing. Personal income and personal spending both came in at 0.7% MoM in May. Initial claims fell another 10k week over week. GDP was revised up to 2.1%, versus 1.6% expected and 0.5% in the previous quarter.
That combination matters. Growth is accelerating, the labor market is holding firm, capex is broadening, and inflation no longer looks like a one-off.
This is what a reaccelerating economy looks like.
Reacceleration, Not Recession
Nothing in this dataset — income, spending, claims, or capex — points to a slowdown story.
If growth and consumption were to weaken from here, the natural conclusion would be that price pressure fades with them. But given the strength in capex and still-loose financial conditions, it is difficult to argue for a growth collapse in this type of economy.
That asymmetry is why we are leaning into the overheat risk, rather than hedging for a recession the data simply is not showing.
From our view, reacceleration remains more likely than clean disinflation. But if that view is wrong, the risk is more likely disinflation than recession.
The key question is whether the Fed understands which cycle it is actually in.
🔑 The 5 Things That Matter
1. The trend is the story, not any single print
May’s 0.3% MoM core PCE print landed in line with consensus. But core PCE has now run at 0.3% or higher for three straight months, plus two 0.4% prints before that.
The year-over-year rate has climbed from 3.0% in December to 3.4% in May, the highest since October 2023.
One hot month is noise. A steady climb like this is not.
2. Supercore confirms this is not just energy
A 0.5% MoM supercore print is the clearest signal yet that the pressure is coming from services and wages, not one volatile category.
This matters because the market was previously treating inflation as mostly an oil story. That view looks too narrow now.
3. Business investment is broadening past the hyperscalers
Core capital goods orders, the Census Bureau’s broadest monthly read on private-sector equipment investment, rose 1.4% MoM in May.
This is not an AI-specific number. It covers machinery, electronics, metals, and other parts of the broader economy. But a print this strong is consistent with AI-driven demand spilling beyond hyperscaler balance sheets and into the wider capex cycle.
That same dynamic is now visible beyond the orders data.
Apple’s June 25 price increases hit MacBooks, iPads, and some other hardware categories, while iPhones were left unchanged for now. Microsoft has also announced Xbox price hikes, and PC makers have been under similar pressure from rising memory and storage costs.
The driver is the same AI-fueled memory shortage sitting underneath the capex numbers. TrendForce estimated that conventional DRAM contract prices surged roughly 90–95% QoQ in Q1 2026, while memory prices were expected to rise sharply again in Q2. Analysts also expect iPhone pricing pressure later this year as higher component costs move further into consumer hardware.
That is the transmission mechanism. AI demand is pulling real input costs into finished-goods prices right now. This is not a future or hypothetical channel.
It is exactly the kind of broadening we would expect if this capex cycle is inflationary before it becomes disinflationary.
4. Equities have retraced, not broken
ES is still holding above 7,495, the 50-day moving average, and remains well above the 7,232 low. Even a move toward 7,000 or 6,800 would still look like a healthy correction rather than a broken trend.
The bigger threat is underneath the surface. Tech and AI names are flooding the market with debt and equity issuance to fund capex. SpaceX is the live example.
That issuance is hitting right as a hawkish Fed is compressing valuations. Nevertheless, we still see issuance risk and rate-hike fear as an opportunity to buy the dip.
5. The market’s reaction does not match the data
Bonds rallied, and the short end started pricing peak hawkishness from Warsh. The market also took comfort from hopes for lasting peace in the Middle East.
But the data argued for the opposite.
That is exactly the kind of mispricing we have been flagging. We disagree with the peak-hawkishness take, but we still believe up to 75bps of additional hikes would not derail the AI boom rally.
Put together, these five points are why our call has not moved.
📍 Our Call
Nothing here has changed. We are still in a mid-cycle reacceleration regime.
The data has simply made the reacceleration case harder to dismiss, and the Fed’s room to stay patient is shrinking faster than the market wants to admit.
Equities Are Retracing, Not Reversing
The SpaceX story still matters because it highlights the risk in capex-heavy, debt-funded names that lack profitability.
Its new debt is mainly refinancing the xAI bridge loan, not funding new growth capex. That is debt paying off debt, not productive investment.
But this remains a concentrated problem, not a market-wide one. Russell 2000 strength tells us breadth is still healthier than the mega-cap “stress” narrative suggests. Russell 2000 has recorded ATH while ES and Nasdaq retrace.
The Bigger Issue Is Rates
The market is leaning into peak hawkishness.
That means investors believe the Fed has already reached the most hawkish point of the cycle, another hike is less likely, and the path from here eventually flattens or eases.
That can create a modest bull steepening in the near term, with the front end leading yields lower.
But we think the market is getting too comfortable too quickly.
Lower oil prices and the mid-June US-Iran memorandum have helped inflation swaps move lower. At the same time, the market is buying the idea that AI is structurally disinflationary.
We understand that view. Over time, AI can be disinflationary through productivity, automation, cheaper compute, and lower marginal costs.
But right now, this cycle is still capex-heavy and infrastructure-intensive. It needs power, memory, cooling, data centers, networking, construction, labor, and financing.
That is not a clean disinflationary impulse yet. It is still an investment boom.
That is why we are not ready to buy the bond rally.
Real Rates Still Do Not Confirm Peak Hawkishness
If the market uses falling inflation expectations to argue real yields are already restrictive enough, we think that conclusion is fragile.
Shorter real yields, especially around the 2-year area, still do not look restrictive enough while near-term inflation expectations remain elevated.
So yes, peak hawkishness can be priced. Yes, a limited bull steepening is possible.
That does not mean the regime has changed. And nowhere is that disagreement sharper than in rates.
⚖️ Two Sides of the Tape
We flagged this split last week, and today’s reaction is that divide playing out in real time.
The Market’s Side
One side looked at the same data we did and chose to buy bonds, price out additional hikes, and lean into the peak-hawkishness narrative.
Some believe inflation is moving back toward a Goldilocks path. Others believe the consumer is weakening, labor demand is softening, and AI capex is not broad enough to offset the slowdown.
There is also a deeper argument behind the long-end rally. The market increasingly believes AI is disinflationary through productivity gains and lower marginal costs.
Our Side
We understand the argument. We just think the timing is wrong.
Long-run structural disinflation from AI does not cancel out the near-term cyclical inflation the buildout itself is generating.
That is why we are still on the other side.
A Core PCE trend that has been climbing for several months, a hot Supercore reading, resilient consumption, accelerating capex, and a labor market that is not clearly loosening are not the ingredients for materially lower rates.
They argue for the opposite.
Friday’s print did not resolve that debate. What comes next might start to.
🔎 What’s Next
The PCE report is behind us, but the question it raised is not.
Watch whether upcoming Fed commentary reinforces or pushes back against the market’s peak-hawkishness pricing. Kashkari has already leaned hawkish. Warsh’s next major remarks will also matter, because the market is trying to decide whether the Fed is genuinely done or simply pausing while inflation risk stays alive.
NFP Is the Next Test
The next data point to watch is Thursday’s NFP. Consensus for June is 114k, versus 172k previously.
A strong print would meaningfully support our reacceleration thesis.
In the near term, a modest bull steepening is still very possible if June NFP comes in in line or slightly softer. The front end would likely lead yields lower as markets price out another hike.
But that would not automatically mean a recessionary or disinflationary regime has arrived. It may simply mean the market no longer believes the Fed needs to hike again immediately.
The Dollar Setup
For the dollar, the setup is more nuanced.
A modest bull steepening can pressure USD temporarily if front-end yields fall. But we do not think that becomes a sustained bear trend.
The Fed still looks relatively more hawkish than most major central banks. US growth continues to outperform. The Middle East ceasefire also remains fragile, with this past weekend’s strikes being the latest reminder.
That keeps a safe-haven bid and some upside inflation risk alive.
We would expect any USD downside to stay limited unless the data clearly confirm a softer growth and inflation path.
After Peak Hawkishness
The bigger question is what happens after the peak-hawkishness trade is priced.
If incoming data continue to show resilient growth, sticky inflation, strong capex orders, and AI spillovers into the broader economy, the market will eventually have to question whether it got too comfortable buying duration.
Our base case is simple. The market can still price peak hawkishness, and we can still see a bull steepening into next week. But NFP comes first.
Our disagreement with the market was never that peak hawkishness cannot be priced. It probably can.
Our disagreement is with what comes after.
Here is where that shows up on the charts.
📊 Key Levels
1. Equities
Same zones as last week.
Bulls want 7,510 to 7,489, the 33-38% Fibonacci retracement, or 7,392 to 7,374, the 64-66% retracement, to hold.
A break below shifts focus to 7,000 and 6,800 as the next buy-the-dip zones.
We would only revisit the broader bullish equity thesis if ES fails to hold above 6,800.
2. Russell 2000
The Russell 2000 continues to show healthier breadth than the mega-cap stress narrative suggests.
If the market were genuinely worried about broad economic weakness, we would expect small caps and cyclicals to deteriorate more meaningfully.
Instead, weakness still looks concentrated in highly leveraged, capex-heavy names where profitability and balance-sheet quality matter most.
3. U.S. 2-Year Treasury Yield
This is where the market is expressing the peak-hawkishness trade most clearly.
The pullback makes sense tactically. If investors believe the Fed has already reached its most hawkish point, the front end should lead yields lower as the probability of another hike gets priced out.
But this is exactly where we disagree.
The data does not yet justify declaring that Fed hawkishness has peaked. Inflation pressure is no longer just an oil story. Core PCE has not convincingly broken lower. Supercore remains hot. Consumption has not clearly weakened. Labor demand is not loose enough to give the Fed an easy exit.
More importantly, our model still shows that the 2-year real rate is not restrictive enough, even after its recent pickup (see below).
It is currently sitting just below 0.5%, still below the level reached during the March conflict peak.
Model note: For our short real-rate proxy, we use the 2-year Treasury yield minus Cleveland Fed expected inflation. We prefer this to relying only on short-term inflation swaps because swaps can move sharply with oil and headline energy repricing. That makes them more reflective of market sentiment than the broader inflation backdrop.
Cleveland Fed expected inflation incorporates Treasury yields, inflation data, inflation swaps, and survey expectations. That gives us a wider inflation-expectation lens.
Based on this framework, the 2-year real rate has picked up recently. But it still does not look restrictive enough to validate the market’s peak-hawkishness trade.
If near-term inflation expectations stay elevated, the real policy stance is less tight than the market wants to believe.
So yes, the 2-year nominal yield can retrace as a peak-hawkishness trade. But we would not treat that as confirmation that the Fed is done.
Further downside should require actual evidence of weaker inflation, softer growth, or a looser labor market — not just the market deciding that peak hawkishness is behind us.
4. U.S. 10-Year Treasury Yield
The 10-year is being pulled between two competing narratives.
On one side, the market is buying the disinflationary-AI story: higher productivity, cheaper compute, automation, lower marginal costs, and eventually less inflation pressure.
That helps explain the long end’s retracement, especially as energy inflation expectations eased after the mid-June US-Iran memorandum.
We are not convinced this narrative dominates yet.
In the near term, AI still looks more like a capex and infrastructure shock than a pure productivity shock. The demand impulse from power, memory, data centers, cooling, networking, construction, and financing remains alive.
So while AI may be structurally disinflationary over time, it can still be cyclically inflationary today.
That is why we think it is too early to aggressively buy duration.
If the market keeps believing AI is already disinflationary, the 10-year can stay capped or retrace further. But if growth stays resilient and capex keeps spreading through the economy, it should eventually resume moving higher.
5. Yield Curve
Since the FOMC meeting, the curve has started pricing the peak-hawkishness narrative. That creates room for a limited bull steepening if the front end leads yields lower.
The key question is whether this stays a positioning-driven repricing or turns into a deeper easing-cycle trade.
For now, we think it is more likely to remain limited unless the data clearly validate a disinflationary or slowdown thesis.
6. U.S. 5-Year Treasury Yield
The 5-year sits between the Fed-sensitive front end and the growth- and inflation-sensitive long end.
It expresses whether policy is restrictive enough over the medium term.
The market’s argument is simple. Inflation expectations have come down, partly because of oil and partly because of the AI-disinflation belief. If nominal yields do not fall as much, implied real yields look more restrictive. That supports the peak-hawkishness narrative.
We disagree on the same grounds as the 2-year.
Shorter real yields still do not look restrictive enough if near-term inflation expectations remain elevated.
So the 5-year is not giving a clean “Fed is done” signal. It is showing the market’s confidence that inflation will fade. We think that confidence is fragile.
If the disinflationary-AI story is being priced too early, the belly of the curve may eventually have to reprice higher again.
Note on key levels: The key levels across asset classes remain unchanged from our previous edition. For the full breakdown across rates, USD, and indices, please refer to the Pre-PCE Regime Update here: https://macroinsightsqueen.com/regime-update-pre-pce-edition/
Bottom Line
The market is pricing peak hawkishness, a softer inflation path, and an AI story that is ultimately disinflationary.
We agree all three are possible. We just think the market is pricing them too early and too cleanly.
Our framework is simple. AI can be structurally disinflationary in the long run but cyclically inflationary in the near term.
The market can be right about the long-term productivity story while still being wrong about the near-term rates path.
The key issue is that our model-implied short real rate still does not validate the market’s peak-hawkishness trade.
The 2-year real rate has picked up recently, but it remains just below 0.5%. It still does not look restrictive enough if near-term inflation expectations stay elevated.
That means the real policy stance may be less tight than the market wants to believe.
So yes, a limited bull steepening is possible. Yes, the dollar may face temporary pressure if front-end yields fall. And yes, the 10-year can stay capped if the market keeps buying the oil-disinflation and AI-disinflation narrative.
But unless incoming data show genuine disinflation, weaker capex, softer consumption, and a looser labor market, we still think the broader regime remains mid-cycle reacceleration with overheating risk — not a clean return to Goldilocks.
That makes this Thursday’s NFP the next key test.
A stronger-than-expected print would likely challenge the peak-hawkishness trade and bring rate-hike fears back into the market. That could trigger another round of equity retracement or consolidation, especially in the more rate-sensitive parts of the market.
But we would still view that as part of the broader buy-the-dip setup laid out earlier in the letter.
The clearer winner from a strong NFP would be the U.S. dollar.
If labor demand remains resilient, short real rates still do not look restrictive enough, and the Fed remains relatively more hawkish than its peers, the market will have a harder time defending the idea that peak hawkishness is already behind us.
🦊👑 MIQ Team
info@macroinsightsqueen.com