2026 Mid-Year S&T Review: Reversals, Contradictions, and Momentum
Last Updated:The last time we talked here – quite a while ago now, apologies for the delay – it was in the midst of the tariff tantrum unleashed after Liberation Day which saw wild market gyrations and even wilder prognostications about the implications that'd flow from the administration's new trade policy (otherwise serious people suggesting we'd see 7% CPI, a 3-4% hit to GDP, etc.).
(As an aside, that's not a bad post to revisit because it strikes at the heart of the misunderstanding of market participants regarding the approach the administration takes across the board – one that ratchets up to catalyze negotiations with the explicit intention of partially unwinding thereafter as Bessent made pretty clear at the time.)
Anyway, while vol is typically pretty muted in spring, it seems the administration wants to break that decades-old trend because, once again, we saw market participants downplay the odds of a significant escalation in Iran before events began to unfold – with the predictable result that the VIX exploded higher (and, as with the 2025 episode, otherwise serious people began to suggest some fairly unserious things such as $200/bbl oil).
The first half of 2026 has seen a series of narrative reversals – from further rate cuts to renewed rate hikes, from the unwind of the AI trade to the revitalization of the AI trade, from concerns about deflationary pressure to concerns about inflationary pressure.
And through this period pockmarked by reversals we've also seen some of the few themes that seemed to stand the test of time begin to crumble – like the hyperscalers having their worst collective month since META's (then FB's) IPO in 2012 (driven by concerns that these names that used to be defined by their capacity to print cash on demand are now planning to spend almost all their FCF on their respective AI buildouts).
Meanwhile, the S&P exclusive of AI names is around all-time highs (likewise the equal-weight S&P). Even as some market participants seem to have convinced themselves – not without the aid of the latest dot plots – that the Fed has adopted a hawkish posture and that we could be in for a series of rate hikes in response to rising inflation (with BofA anticipating three, which is an odd reaction function to have with breakevens and the long end of the curve well anchored – plus Warsh's disposition being generally agreed to sit toward the dovish end of the spectrum despite his recent comments at his inaugural press conference that were broadly interpreted to the contrary, but we'll see).
2026 Mid-Year Sales and Trading Questions and Answers
Since it's been a bit of time since my last post, I'll keep to the same format we've more recently used since everyone has seemed to enjoy it. Namely, talking through some markets-based questions that could come up in interviews (as always, the level of depth and detail in my answers will be a bit above expectations).
Through the remainder of 2026, what would be the case for rate hikes?
Do you view the current inflation we're seeing as transitory, or is there a risk that it could be stickier and thus a sign of concern?
The more things change, the more things stay the same. It was almost half a decade ago now that the debate began over the nature of the inflation creeping into CPI and PCE readings. Was it transitory or not? Should there be a monetary policy response, or should the inflation be looked through? After all, we went through a period of significant deflationary pressure in the immediate wake of the pandemic, so maybe inflation could run hot to average things out?
Of course, this debate continued (despite the ultimate monetary policy response being the fastest hiking cycle in four decades) until, at the beginning of 2025, it appeared that – whether one wishes to classify that inflation episode as transitory or not – inflation seemed destined to return to target.
That is, until Liberation Day occurred and a different debate began over the level of inflation that would feed through from increased tariffs and whether this was truly inflationary or a one-time price-level adjustment that’d largely be immune to rate tweaks.
The latter was always more correct than the former, as Chair Powell to his credit repeatedly stated, and despite the chaotic application of tariffs – which left no one quite knowing what the effective tariff rate was against anyone – what began to be observed was an easing of inflationary pressure as we entered 2026.
Just in time, then, for the Iran conflict to restart the debate: this time over how quickly we'd see inflationary passthrough from higher energy costs, and whether this would stoke underlying inflationary pressure throughout the economy or if this inflationary driver (the conflict) should be, as with the tariffs, looked through as well.
As is always the case when we talk about inflation, there are multiple factors in play. Right now, the dominant three – at least when it comes to figuring out how durable inflation is liable to be – are energy price passthroughs, AI-related price pressures (think: memory), and higher equity prices. (Equity price gains will provide a boost of about 0.1pp in May and 0.05pp in June to core PCE – one of those ironies of the core PCE composition.).
When it comes to energy (or commodity) passthroughs, the recent US-Iran MOU has taken the wind out of the sails of those that expected a significant lift to headline measures of inflation with higher energy prices then having downstream impacts that'd hit core and potentially be more durable.
While there was no doubt that a more permanent closure of the Strait of Hormuz would eventually lead to significantly higher oil prices, the on-again-off-again nature of the “negotiations” – alongside a significant use of global reserves and some apparent China demand destruction – kept oil constrained. And now, with the MOU, most strategists – and the futures market – are predicting a steady glide path down to a level that represents a slight premium to where we were when this all began.
(Even if tensions flare again – which, for what it's worth, I think market participants are significantly underweighting – I wouldn't anticipate a significant response in futures as everyone has trepidations about getting in the way of a market that could whipsaw on a singular Tweet or Truth.)
Right now, most see Brent sharply dropping to the low-70s and staying there through 2026-27, which would take around 0.2pp off headline PCE and 0.05pp off core PCE (a sign of how minimal the oil passthrough effect is in the US context).
Beyond oil, the same theme remains across nitrogen, refined products, etc. with the expectation being that a continued downward drift in pricing will occur and that by the time the passthroughs are in the data at the peak they'll reflect a 0.40pp uplift in core PCE readings before abating.
What wasn't on anyone's bingo card – or, if it was, they've enjoyed a nice 200%+ return – surrounds the spike in memory prices that has continued through the first half of 2026 (with some types of memory up 10-15x through Q1).
While computer software and accessories make up a pretty small component of both PCE and CPI, when even a driver of a small component has such a significant increase that tends to move the needle for the top line reading (software and accessories inflation was running 4-5% month-over-month in recent months).
While prices have begun to cool, the implications of a level set higher still haven't been felt in phone and computer price changes (which are beginning to occur now, so will feed through to inflation measures in the months to come). All told, GS sees around a 0.4pp PCE uplift by year-end (an estimated 0.25pp right now, with a 0.6pp uplift at the peak).
When the impact of energy passthroughs and AI pressures are stripped out, the inflation backdrop looks remarkably benign for the amount of volatility we've seen over the last 18 months – importantly, housing inflation, while volatile, has settled into a lower range and, most importantly, wage inflation has remained subdued for most (as I harped on through the 2022-2024 period, elevated wage inflation is always the key measure to watch when it comes to durable inflation – and it's the reason GS was directionally more correct than others on their inflation calls through that period).
As GS noted in a recent report, if one squints there appears to be an uptick in broader inflationary pressure – but when you decompose out those components impacted by energy passthrough it appears much more as a blip than the beginning of a trend (which is also matched by what can be seen in alternative data sources).
Given all this, it does appear that the major drivers of inflation readings being well above 2% – energy passthroughs and AI-related impacts – are on pace to wind down by the beginning of next year and that we'll return to a slightly above target level of inflation. In other words, neither appears to be sticky although both have the capacity to be sticky if the Strait remains open in one moment and closed the next for an even more prolonged period and if AI bottlenecks continue to be squeezed ever tighter (there’s also an open question on if memory prices will be fully fed through to consumers now, or if some companies will absorb higher memory prices to retain market share while they wait to see if memory prices collapse as they historically do after rapid runups). Or, of course, there’s little doubt we’ll see more durable inflation, and a Fed response, if there's a sudden surge in growth and employment that begins a new cycle of wage inflation but there's little sign of that right now.
Through the remainder of 2026, what would be the case for rate hikes?
After the insurance cuts of 2025, the consensus view was that rates would stay flat through 2026 – with perhaps an additional cut or two to bring rates back to neutral (or what most Fed members believe is the neutral rate). Then, as the Iran conflict began, the market began to develop a bifurcation: with, on the one hand, some believing there would be additional cuts as GDP took a hit and, on the other hand, some believing the Fed would begin to engage in some emergency hikes to dull the inflationary impact of the conflict (through the aforementioned energy passthroughs).
Right now, the primary outlier calling for hikes – despite, or perhaps because of, the MOU – is BofA who's calling for three through the remainder of 2026 (at the September, October, and December meetings). The call is rooted, as one would expect, in the US exceptionalism theme: that the labor market showed remarkable resilience through the conflict and is poised to accelerate hiring through the remainder of the year. However, because of how relatively tight the labor market still is, wage inflation will begin to accelerate and to get ahead of it the Fed will begin tightening further (taking rates from slightly restrictive to more truly restrictive).
BofA is also an outlier (although not as much as with their rates call) on their inflation outlook – with the thought being that while there is passthrough from energy and AI, that's a sideshow. What matters is what's underneath, which has trended lower and which they believe will begin to pick up. And it's this fact, alongside potential wage inflation from a tighter labor market, that creates the conditions for a more sustained ~3% core PCE which Chair Warsh has made clear he does not want to become an embedded assumption in markets (i.e., that the inflation target might be 2% on paper, but is 3% in practice).
As won't come as a surprise, I tend to side a bit more with GS (although I always found their call for a possible July move, which they've now downplayed, a bit at odds with their own interpretation of the data). In a recent desk note, Anshul Sehgal, Co-Head of FICC, made the case that – unless there's another exogenous shock – inflation should come back closer to the mid-2s: not just because oil prices are lower, but because “there is no wage price spiral,” labor is marginally hurting right now, and savings rates are still reasonably high. In other words, not worried about the economy, but not seeing inflation as a pressing issue where the Fed might need to restart the hiking cycle – hence taking down their probability of a July hike.
This is all reinforced by the recent payroll report which took some air out of the balloon with the three-month average NFP ticking down to a steady 111k instead of the somewhat hot 188k it was for the prior three months (before the latest report). So, we have muted vacancies, muted job openings, muted wage inflation, and payroll growth that's steady but not spectacular. Hardly a recipe for a preemptive rate hike, and the market seems to slowly be coming around to that consensus (here's the latest hikes priced in by upcoming meeting date).
Much of the incoherence in rates chatter as of late stems from Chair Warsh's earlier June FOMC comments and the dot plots that most interpreted as hawkish – but in his more recent Sintra speech he touted improved inflation expectations (that have stayed well anchored and inched down). Which I think is fairer to read as almost taking credit for his tougher rhetoric leading to lower inflation expectations and thus negating the need to contemplate hikes.
The one notable element of Warsh’s time in Sintra – which most interpreted to be quite dovish overall – came from an offhand comment made when the BoC's Macklem remarked that all developed economies face similar inflationary issues (rooted in energy price passthroughs). To which Warsh felt the need to state the obvious: that the Fed must also contend, unlike all other central banks, with an unprecedented level of capex aimed at the AI buildout and the inflationary pressures that stem therefrom (perhaps signalling that while he's fine looking through energy ebbs and flows, he's less comfortable with the potential inflation consequences of the AI buildout).
As an aside, there's been much ado in the media about the market pricing in Fed hikes, but as Barclays noted in a recent report that stands in contrast to an uptick in short-end bond fund inflows (which you wouldn't expect to see in the run-up to a hike – historically flows weaken as the market begins to anticipate one).
Barclays' determination – and it's sensible enough – is that the broad market isn't fully buying the rate hike narrative and certainly isn't buying into the idea of an actual rate hike cycle beginning (one that'd see two or three hikes in quick succession over three or four meetings).
Instead, investors are being drawn in (not for the first time) by the slightly higher yields on offer in the US as the market begins to digest the possibility of hikes. Although, as perhaps a prelude to a theme we'll discuss in the future as significant long-duration supply comes online toward the end of the year, the appetite for long-duration funds has been notably weak despite the fact that the 30Y Treasury has quietly hit post-GFC highs.
What do you think the largest risks to earnings are this quarter, and do you see equities higher or lower?
No different than when we discuss the largest risks to an upside surprise on inflation, there are a myriad of possible answers here – and, also as with inflation, it tends to be that most risks in a certain moment are correlated.
When one looks at analyst estimates, they're projecting EPS growth at an eye-watering 22% this quarter – which, to most, will look almost absurd since it's not off a low base (in other words, not relative to a quarter that saw abnormally low growth). However, it's actually a step down from the realized EPS growth of last quarter, which was an even more eye-watering 27%.
And, of course, the reason for this is the reason behind much of the consternation of the last few years in markets: AI spend and, more specifically, how much longer it can possibly continue. Because if one looks at the median S&P company, EPS growth is expected to be 9% (still nothing to sneeze at, but it illustrates how much EPS growth is being driven by a few names concentrated within the same theme – 10 stocks alone contributed 70% of EPS growth last quarter, with just Micron and NVIDIA making up 40%).
So, as with the last half dozen quarters or more, the largest downside risk to earnings is a slowdown in projected capex spend which would create a kind of domino effect in which the reduced capex would lead to sharp revisions to EPS growth estimates for the semiconductor, AI infra, and memory names that have driven EPS estimates (and earnings) higher. For example, last quarter the hyperscalers alone raised capex guidance by $100b which drove higher EPS revisions across a wide set of semi, AI infra, and memory names.
However, it's on balance less likely that we'll see similar sharp (upward) revisions this quarter. Because for the last few quarters the hyperscalers have begun to be punished by the markets for their aggressive capex spend (as previously mentioned, spending almost all their FCF) and the risk is that the hyperscalers take a look around, realize the market would reward them for more constrained capex spend, and do so.
This would be sensible enough but when one looks at the level of spend (this isn't a Metaverse-style experiment) and the significant equity and debt raises that have been done to facilitate this spend, it's more likely that the hyperscalers are committed come hell or high water to their capex plans (and, if there's overcapacity, will tell themselves they can just sell excess capacity like Meta has floated).
The other large downside risk – as always – is more macro in nature. In recent quarters, especially since the US-Iran conflict began, there's been a steady rise in input costs that by and large seem to have been able to be fed through to keep profit margins on track. However, as we discussed in relation to the inflation backdrop, what we haven't seen is wage pressures begin to build and with the latest jobs data we don't have quite as tight of a labor market as some thought a month ago (although by any measure it's still pretty tight).
Therefore, a natural consequence here could be that the pricing power of companies will be chipped away by input costs rising at a level that can't be fully passed through to consumers due to the lack of wage inflation we’re seeing (with wage inflation representing the other side of the coin to consumer buying power).
When it comes to price action in equities over the next quarter, the simplest frame would be that if the current inertia holds, even if at a slower pace, then one would expect equities to continue their upward march. However, the most liable crack that could form would be if capex spend began to be revised lower (even if not until 2027) which would lead to EPS growth revisions across the largest EPS growth engines of the S&P (i.e., Micron and NVIDIA) which would then lead to a larger scale sell-off across semiconductor, AI infra, and memory names.
And even if the hyperscalers were behind the reduced spend – which could lead to the market rewarding them, potentially – with how much growth is tied to AI names outside the hyperscalers, and how crowded that trade is with sentiment at extremely stretched levels, it's unlikely that positive hyperscaler price action could offset the declines seen in names that are direct or indirect beneficiaries of their huge spend.
Another way to see this more granularly is through the correlation between hyperscalers and memory names, which has completely broken down over the last quarter as market participants have begun to realize that the hyperscalers are a significant funding arm of the semis/memory trade (which makes the hyperscalers, trading on an NTM P/E basis at 10-year lows, a hedge against the semis/memory trade – but it also means that unless this correlation reverses then if the hyperscalers begin to move it'll be at the expense of the area of the market that's most driven recent moves higher).
Summary
In the posts I wrote in 2022-24 I frequently (somewhat) joked that the only theme that really mattered was rates – because it seemed like all asset classes, at one level removed, ebbed and flowed on where rates were projected to be (especially in the time when everyone waited with bated breath for rates to begin to be cut and had their hopes routinely dashed as rate cuts were pushed back further and further).
While rates still drive a significant amount of discourse – especially over the last few months – there's no question that other themes like tariffs and AI impacts operate on somewhat separate tracks and that has perhaps made some market participants pine for the days in which a correct call on rates, no matter the asset class one dealt with, was all one needed to be concerned about.
And it seems like both of these themes will be with us for some more time. While we've discussed AI at length, it's worth a brief mention on the state of play on tariffs because things are still as uncertain as they were after Liberation Day when it comes to the effective rate (albeit the stakes are lower now and market participants have learned to love, or accept, the bomb to an extent).
After the SCOTUS decision, the administration enacted Section 122 tariffs which, for simplicity's sake, we can think of as a broad-based 10% tariff with a few carveouts here and there. Since, by statute, Section 122 tariffs are time-limited (five months) they're set to expire on July 24.
The administration made it clear from the outset that they understood Section 122 tariffs would expire, and would thus run a parallel process to put in place more durable (from a time and legal perspective) tariffs through Section 301 (that require an investigation to demonstrate need) set between 10-12.5% depending on the country.
However, overhanging this is the possibility that the administration could seek to use Section 301 tariffs not to backfill the Section 122 tariffs (that limited the total tariff level) but instead to backfill the effective tariff rates on countries (generally speaking) in place before the SCOTUS decision. Which would mean a significantly higher effective tariff rate – probably somewhere within the 15-20% range. This latter possibility doesn't seem priced into markets much at all, but the administration quietly launched a separate Section 301 investigation on industrial capacity that seems like it could lead to more bespoke tariff levels in addition to the kind of base set by the earlier 301 investigation (to roughly replicate the Section 122 levels).
Regardless, as I've mentioned a few times over the last 12-18 months, while the theme dispersion we're seeing can create some level of disorientation whipsawing markets, and theme rotation, all decompose to volatility, and for sell-side desks volatility generally means higher desk activity and higher desk revenues (which is exactly what we've seen of late). Thus, all the vol should be welcomed with open arms – at least for now.