AI’s Big Guns Have a Serious Inflation Problem

AI cost inflation is finally going mainstream.

AI infrastructure costs just keep on rising. (…) The artificial-intelligence boom is also crowding out supplies of more conventional chips. (…)

Some of the most cash-generative tech firms and best-funded startups in history are frantically competing to secure enough hardware, fearing that otherwise they’ll be left behind in the race for superintelligence. And while these buyers are relatively price-insensitive, their chip suppliers tend to have dominant market positions with huge technical and financial barriers standing in the way of more competition. Controlling these bottlenecks is proving very lucrative. (…)

Hyperscalers’ latest earnings reports reveal the painful sting of inflation. Microsoft expects higher component prices to add $25 billion to its full-year capital spending bill, which now totals an eyewatering $190 billion. Meta hiked the midpoint of its forecast capex range by $10 billion, attributing it mostly to component costs, particularly memory chips. (…)

SK Hynix’s operating margins reached a record 72% in the latest financial quarter. Customers are “prioritizing securing volume over pricing,” the South Korean company says, unabashed. Samsung’s average selling prices for DRAM increased by more than 90% in the same three-month period compared with the quarter preceding it.

In aggregate, spending on various types of memory could account for 30% of hyperscalers’ capex in 2026, according to research firm SemiAnalysis. It was just 8% in 2024. (…)

[GPU] Rental prices are “increasing across the board,” says CoreWeave Inc., a prominent neocloud. (…)

Amazon expects Trainium to save it tens of billions of dollars yearly on its own spending. Claude and ChatGPT’s owners, Anthropic PBC and OpenAI, have signed multibillion dollar contracts for the chips with Jeff Bezos’s firm, although in the near term most supplies are sold out or reserved.

Among other promising innovations, Google’s TurboQuant compression technique could help curb memory spending, and Arm Holdings Plc expects its new CPU to cut the cost of a gigawatt of datacenter capacity by $10 billion or so. Meantime, the tech elite’s largess is having unwelcome spillover effects. (…)

And beyond the AI frenzy, manufacturers of smartphones, games consoles and PCs are struggling to secure memory chip supplies because their makers are prioritizing the more lucrative datacenter market and long-term contracts with hyperscalers. Consumer-tech product firms must either pass on price increases, lower device specifications or swallow a hit to their margins. Global smartphone sales are projected to decline around 13% this year, with budget handsets particularly affected. Nintendo Co. Ltd. has lifted the price of its Switch 2.

Semiconductor factories take years to build, so there’s no prospect of a swift supply response. It’s a notoriously cyclical industry and several companies suffered heavy losses not so long ago, making them wary of overexpanding.

When you factor in higher electricity prices caused by power-hungry data centers, AI will probably be quite inflationary for a while. (…)

Perhaps as a case in point, CoreWeave stock sunked 18% since May 6, after beating revenue estimates by 5% to +112%, showing a revenue backlog up 49% QoQ and contracted power reaching 3.5GW from 850MW at the end of 2025. Goldman Sachs suspects the “stock reaction is a function of the increase in capex reflecting increases in component pricing”.

From yesterday’s Daily Edge, in case you missed it:

David and I have been strong believers in AI. The February 9, 2026 post Railroaded? and the March 13 deep dive The AI Supercycle: A Deep Dive offered our supporting analysis.

Concluding Railroaded? I wrote:

• AI is truly transformational and will be quickly widely adopted for productivity and competitiveness imperatives.
• Unlike the railroads and the IT infrastructure booms, there is little front-loading “hoping/waiting for demand” investments, nor much leveraging at this time.
• Prices/costs are coming down fast, necessary to boost demand/usage along the way.
• Agentic AI will be huge.
• In the AI era, moat is crucial and bigger is better.
• Diversified revenue/cashflow streams are desirable.
• No or low indebtedness is preferable.

From a macro perspective,

• AI is clearly and significantly boosting GDP growth.
• AI is clearly inflationary in some sectors (construction, commodities, power), disrupting other sectors by hording resources.
• The massive hyperscalers expenditures (almost the size of the US defense budget) are redirecting their excess cash from the fixed income markets into the real economy with high multiplier effects. Will productivity gains offset demand-pull inflation?
• Will these financial flows (reduced corporate demand) impact US interest rates?

Only 4 months later, demand is even stronger than expected thanks to Anthropic’s Claude Code. Compute demand is even too strong, too quickly, forcing Anthropic to tame demand through higher prices and controlled access until more data centers add to compute supply.

The various expected roadblocks have materialized, e.g. power (Power Play Sept 23, 2024), equipment, workers and NIMBY protests, slowing supply growth but extending the buildout cycle.

Costs are rising fast, magnified by the US war on Iran and its effects on various supply chains. An increasing part of AI capex is scarcity-induced inflation, positive for suppliers’ margins dealing with scrambling, price insensitive buyers.

BTW, this is a dangerous situation for inflation. Consider:

• Exploding AI compute is fueling significant demand other than chips and memory for resources such as power (oil, gas, solar, nuclear), labor (construction, plumbers, electricians) and specialized hardware (turbines, transformers, cooling equipment). Time is of the essence so hyperscalers are totally price insensitive, a situation likely to persist through 2030.
• Hormuz is forcing countries and corporations to build reserves and hoard various critical resources and materials as protection from constrained supply chains. Hoarding of oil, gas, helium, fertilizers, chips, etc. is now required in an increasingly selfish, uncooperative environment. This will permanently raise resource costs across the world and reduce/eliminate operating efficiencies built during globalization.
• The Ukraine and Iran wars have significantly depleted ammo and equipment inventories which will need to be hastily rebuilt. At the same time, the world has also learned that alliances can be fickle, incentivizing many countries to find ways and means to protect themselves. Demand for military equipment will be firm and largely price insensitive for several years.

AI and productivity will not help offset the huge, world-wide demand for such a broad spectrum of resources and goods against largely price insensitive buyers.

This was already evident from April’s J.P. Morgan Global Manufacturing PMI:

• Global manufacturing growth accelerates in April despite sharp rises in prices and supply chain delays

The start of the second quarter of 2026 saw rates of expansion in global manufacturing output and new orders strengthen. However, price and supply chain pressures continued to build (…).

Manufacturing production increased for the ninth month running, with the rate of growth hitting a near five-year high. Expansions were signalled across the consumer, intermediate and investment goods sectors.

The increase in production was supported by faster growth of new business. (…)

The resulting input shortages and delivery delays led to a solid upswing in purchase price inflation, as demand exceeded available supply. Average input costs rose at the quickest pace since June 2022 and at one of the fastest rates in the 28-year survey history. (…)

Average output charges rose at the sharpest rate in 45 months, as manufacturers passed on part of the increase in costs to clients. Backlogs of work meanwhile rose for the third month running.

Superimposed on the above:

U.S. and Iran Are Locked in a Stalemate That’s Neither Peace Nor War The two sides are far apart from a deal but also don’t want to resume fighting

The U.S. and Iran are locked in a diplomatic stalemate over issues that have bedeviled the two sides for years, as the conflict settles into a gray zone that is neither war nor peace.

The cease-fire is entering its second month and, despite sporadic violence, has now lasted almost as long as the fighting which preceded it. There is little to indicate that either the U.S. or Iran is ready to compromise, but neither wants to start fighting again.

President Trump on Monday warned that the cease-fire with Iran is “on life support” and said he wouldn’t back off his goal of pressing Iran to abandon its nuclear program. Trump told reporters at the White House that Iran believed he would get tired or bored of the conflict, or feel pressure to end it because of rising energy prices.

“But there’s no pressure,” Trump said. (…)

• Iran Thinks Trump Is Bluffing The regime figures it can outlast the U.S. on the Strait and uranium.

The WSJ Editorial Board:

(…) By clinging to the cease-fire through it all, Mr. Trump sent the wrong signal. Iran’s regime clearly thinks it can outlast a President who no longer wants the fight. “They think that I’ll get tired of this or I’ll get bored, or I’ll have some pressure,” Mr. Trump recognized in his Monday remarks, “but there’s no pressure at all. We’re going to have a complete victory.”

The problem is that he is under pressure, and everyone knows it. Why else is the President now talking about pausing the gas tax? Mr. Trump is right about the regime’s perception of him, but he’ll have to prove it wrong. (…)

This is a regime that thinks it can absorb economic pain from the U.S. blockade longer than Mr. Trump can tolerate higher prices for oil and petrochemicals. Mr. Trump will have to persuade Tehran’s leaders they’ve underestimated him—and the pain.

While most pundits talk of a potential inflation flare, the biggest danger is in severe and persistent shortages of materials, resources, and components that will broadly impede production across the world. To wit:

• Wood Mackenzie expects aluminum prices to rise to around US$3,500 per tonne in 2026 due to supply deficits.
• Semiconductor makers face supply risks as Qatar helium supplies may take a while to return to pre-war levels, potentially hampering AI chip production.
• The effective closure of the Strait of Hormuz has jeopardized over 25% of the world’s helium supply (primarily from Qatar). Helium is indispensable for semiconductor lithography and cooling; without it, chipmakers face capacity constraints.
• Reports emphasize that the war has exposed a “security of supply” crisis for raw materials needed for the green transition, such as nickel and lithium. Morgan Stanley predicts an 80,000-ton shortfall in lithium carbonate equivalent (LCE). This is directly increasing the cost of U.S. power grid upgrades and electric vehicle production.
• Roughly 50% of the world’s seaborne sulfur trade transits through the Strait of Hormuz. China, the world’s top sulfuric acid producer, implemented export bans in May 2026.
• Sulfur is the primary ingredient for sulfuric acid, used to make phosphate fertilizers.
• Sulfuric acid is essential for leaching metals like copper, nickel, and cobalt.
• In Chile, sulfuric acid prices have more than doubled, threatening the production of copper needed for global electrification.
• High-purity sulfuric acid is a critical “workhorse” chemical in chip manufacturing. Analysts at The Soufan Center warn that the sulfur shortage is now a direct threat to the global defense industrial base.
• While the United States produces about 90% of its sulfur domestically, the current global “sulfur shock” is beginning to impact the U.S. economy through price contagion and supply chain friction.
• Shortages of yttrium and scandium—vital for defense tech and aerospace—are worsening, with some suppliers reportedly turning away clients as of late February 2026.
• A key pain point is yttrium, used in coatings that keep engines and turbines from melting at high temperatures. Without regular application of these coatings, engines cannot be used. Some coatings manufacturers are also now starting to ration material, according to company executives and traders. Executives at two North American firms that buy yttrium to make coatings told Reuters they have needed to temporarily pause production due to shortages. One is also now turning away smaller and offshore customers in order to conserve supply for larger clients, which include certain engine makers.
• U.S. semiconductor makers are running low on scandium, putting production of next-generation 5G chips at risk, said Dylan Patel, founder and CEO of research firm SemiAnalysis. (…) scandium plays small but important parts in fuel cells, specialty aluminium aerospace alloys and advanced chip processing and packaging. Major U.S. semiconductor manufacturers all rely on scandium for making chip components that “go into essentially every 5G smartphone and base station”, Patel said.

U.S. Adds 115,000 Jobs in April With Solid Hiring Across Sectors

The U.S. job market blew past expectations again in April, buoyed by gains across industries including retail, transportation and warehousing, and healthcare. The results were a sign that the labor market remained resilient so far in the face of the Iran war.

The American economy added 115,000 jobs in April, the Labor Department said Friday, far exceeding expectations.

That was down from a net gain of 185,000 in March. But it was much better than the 55,000 jobs that analysts polled by The Wall Street Journal had expected to see for April.

The unemployment rate stayed unchanged at 4.3%, as economists had expected. (…)

In the first four months of the year, monthly payrolls have averaged 76,000, up from an average of about 42,000 during the same period last year. (…)

The Friday report should put the focus squarely on inflation data when it comes to determining where the Fed—now firmly on pause—goes from here. While the labor market has steadied, inflation is now drifting up instead of down due to the effects of tariffs and the Iran war, which has sent gasoline prices soaring.

With the labor market giving the Fed cover to wait, the next move on the policy debate is when and how to tilt toward a neutral bias that suggests a rate increase could be as likely as a rate cut. The answer could turn almost entirely on the inflation numbers. The Labor Department will report the latest data on the consumer-price index next week. (…)

A key measure of underemployment also rose—a sign more people are taking on part-time roles because they can’t find a full-time job.

The U-6 rate, which includes people working part time who would prefer a full-time job and people who are so discouraged they have stopped looking for work, rose to 8.2% in April from 8% in March. It is at the highest level since December and more than a percentage point above where it was just before the pandemic.

Closely watched by the Fed, U-6 suggests many Americans are struggling with underemployment. The number of people working part time who would have preferred full-time employment increased by 445,000 to 4.9 million in April, the Labor Department said. (…)

The household survey is not quite as hopeful as the establishment survey, the latter being down for the 4th consecutive month during which time it declined by 343k monthly jobs on average to reach its lowest since December 2024.

Smoothing out the monthly establishment data, the 3-m pace of hiring is now 48k vs the 6-m average of 55k.

A slow May, or revisions, could reset the “resiliency” narrative.

The U-6 unemployment rate is still creeping up, as are the number of part-timers, as Americans find ways to make ends meet. At 8.2%, the U-6 is well above its Q4’2029 level of 6.8% while part-timers are up 16.5%.

Hourly earnings rose 3.6% YoY, in line with its 4-m average. But MoM, growth has slowed from +4.3% annualized in January to +2.6% in March and to +1.9% in April.

Indeed Job Postings peaked in early March and are still weak through April 30, suggesting the next JOLTS report will be weak.

Canada: April’s jobs report reinforces the weak start of the year

From LSEG IBES:

440 companies in the S&P 500 Index have reported earnings for Q1 2026. Of these companies, 83.2% reported earnings above analyst expectations and 13.2% reported earnings below analyst expectations. In a typical quarter (since 1994), 67% of companies beat estimates and 20% miss estimates. Over the past four quarters, 78% of companies beat the estimates and 17% missed estimates.

In aggregate, companies are reporting earnings that are 8.1% above estimates, which compares to a long-term (since 1994) average surprise factor of 4.4% and the average surprise factor over the prior four quarters of 7.1%.

Of these companies, 78.2% reported revenue above analyst expectations and 21.8% reported revenue below analyst expectations. In a typical quarter (since 2002), 63% of companies beat estimates and 37% miss estimates. Over the past four quarters, 73% of companies beat the estimates and 27% missed estimates.

In aggregate, companies are reporting revenues that are 2.1% above estimates, which compares to a long-term (since 2002) average surprise factor of 1.3% and the average surprise factor over the prior four quarters of 1.9%.

The estimated earnings growth rate for the S&P 500 for 26Q1 is 28.6%. If the energy sector is excluded, the growth rate improves to 30%.

The estimated revenue growth rate for the S&P 500 for 26Q1 is 11%. If the energy sector is excluded, the growth rate improves to 11.6%.

The estimated earnings growth rate for the S&P 500 for 26Q2 is 22.9%. If the energy sector is excluded, the growth rate declines to 19.6%.

Revisions are exceptionally strong:

 

Amid a highly uncertain world, many more companies are offering guidance than at the same time last quarter (86 vs 59), slightly more negative (+51%) than positive (+33%).

Trailing EPS are now $286.72, up 4.1% from Q1 and +12.9% YoY (S&P 500 Index is up 24.2% YoY). Full year estimates: $336.49e (+24.0%). Forward EPS: $347.01, up 7.7% from Q1 and +28.8% YoY). Full year 2027e: $386.7 (+14.9%).

Since 2019 (pre-pandemic) S&P 500 EPS would be up 106% by the end of 2026. Doubling in 7 years!

That’s only happened in 6 periods since the Great Crash of 1929, the last ones in 2015-17 and 1998-2001.

But hang on here!

US public companies are required to report the quarterly mark-to-market value of their equity holdings in their net income.

In early 2026, the mark-to-market rule has already caused massive, multi-billion dollar swings in the reported net income of several S&P 500 giants, effectively detaching their “headline” earnings from their actual operational performance. E.G.:

• Alphabet (GOOGL): Reported a massive $37.7B gain in other income for Q1 2026, primarily from net unrealized gains on its non-marketable equity securities (including stakes in Anthropic and SpaceX). This non-operating tailwind added roughly $2.35 to its diluted EPS, making its 81% net income jump look much larger than its 30% operating income growth.
• Amazon (AMZN): Its Q1 2026 net income included $16.8 billion in pre-tax gains from its investment in Anthropic.
• Meta recorded $6.9B in “income tax benefits net of investment losses”.

Goldman Sachs notes that while the aggregate S&P 500 earnings growth was reported at roughly 25% so far, “normalizing” GOOG and AMZN results would bring the underlying earnings growth was closer to 16%.

We are not done normalizing. Many other companies yet to report will also include gains in their Anthropic/OpenAI/SpaceX/etc. holdings as of Q1, but also going forward when more gains will likely accrue.

Be careful looking at some P/E ratios! The “E” may not be sustainable operating earnings.

Based on recent filings, GOOG owns 14% of Anthropic, AMZN between 15-21%, MSFT 5-10% and NVDA 3%.

Alphabet and Amazon used Anthropic’s February 2026 Series G funding valuation of $380B for their Q1 mark-to-market.

Last week: Anthropic weighs deal for near $1tn valuation as revenue surges

The new round is expected to value Anthropic at about $900bn pre-money and to raise as much as $50bn, said three of the people. They added it was likely to close within two months. OpenAI was valued at $852bn post-money in March after it closed a record funding round of $122bn.

FYI, based on my available info:

• There are 12.1B GOOG shares outstanding, meaning GOOG’s 14% stake would be worth $125B at $900B valuation or $10 per GOOG share.
• There are 10.7B AMZN shares outstanding. Its ~18% approximate share would be worth $162B at $900B valuation or ~$16 per AMZN share. AMZN also owns ~6% of OpenAI, worth some $51B or ~$5 per AMZN share.
• There are 7.4B MSFT shares outstanding. Its ~7.5% approximate share would be worth $68B at $900B valuation or ~$9 per MSFT share. MSFT also owns ~27% of OpenAI, worth some $230B or ~$31 per MSFT share.

Back to earnings:

Q2 estimated growth is 22.9%, thanks to Energy (+100%), Materials (+30.6%) and Tech (+55.4%).

All 7 other sectors are seen up 8.7% on average vs +9.3% on April 1 and 10.0% on Jan. 1. The term “broadening” market may be too broad.

Mind you, 8-9% profit growth is good, but we need to monitor the slowdown given the rising squeeze inflation is having on consumers’ real spending power.

Weekly payrolls were still up 4.0% YoY in April but PCE inflation was up to 3.5% in March, rising fast from +2.8% in since last November.

Putting aside hours worked (cancelling every second month), payrolls growth has been slowing every month since January, increasingly relying on wage growth (black) which is also slowing fast as seen above.

About the broadening market:

Big Tech’s heft is obscuring emerging weaknesses elsewhere.

According to a recent UBS analysis, 42 stocks are driving the bulk of the S&P 500’s returns; typically, around 100 do.

The index is up 12 per cent since the end of March on the back of AI-fuelled tech blue-chips, which also include Alphabet, Microsoft, Apple, Meta Platforms and Broadcom. The equal-weighted version of the benchmark, which in effect turns down the volume on those giant stocks, is up half as much.  (FT)

 

Topdown Charts Professional

• Never mind valuations, Big Mo is the game:

• The below helps explain the above…

 @MikeZaccardi

• … and some of the below:

Goldman Sachs

• Positioning at extremes!

Topdown Charts Professional

Ed Yardeni explains what’s currently happening:

(…) late last year, a growing concern that the AI boom was turning into an AI capital spending arms race among the Mag-7 was heightened by Michael Burry’s warnings that the hyperscalers’ massive AI capex might prove unprofitable for various reasons.

But those concerns have diminished in response to significant beats by the hyperscalers during the Q1 earnings season in April. Their cloud earnings continue to soar, confirming that rapidly growing demand for “compute” might justify all the AI capital spending after all.

(…) Here is a quick review of how the AI-11 fits into the AI supply chain:

(1) Foundry and lithography (TSMC, ASML). TSMC fabricates the leading-edge logic for everyone. ASML owns the EUV lithography chokepoint.

(2) Logic and custom silicon (AMD, Broadcom, Intel). AMD is taking a significant share in AI inference. Broadcom is the custom ASIC partner for hyperscalers and the incumbent in networking silicon. Marvell rounds out its custom silicon, networking, and optical connectivity offerings. Intel is the foundry comeback story with CPU exposure to the AI server cycle.

(3) Memory (Micron, SK Hynix, Samsung). Micron, SK Hynix, and Samsung supply the high-bandwidth memory that is the actual bottleneck for AI training. SK Hynix leads the HBM market globally.

(4) Enterprise NAND and storage. SanDisk has emerged as the pure-play beneficiary of NAND and enterprise SSDs. Western Digital provides the HDD complement.

Every dollar of hyperscaler capex for AI infrastructure flows through this supply chain before reaching a server rack.

David and I have been strong believers in AI. The February 9, 2026 post Railroaded? and the March 13 deep dive The AI Supercycle: A Deep Dive offered our supporting analysis. 

Concluding Railroaded? I wrote:

• AI is truly transformational and will be quickly widely adopted for productivity and competitiveness imperatives.
• Unlike the railroads and the IT infrastructure booms, there is little front-loading “hoping/waiting for demand” investments, nor much leveraging at this time.
• Prices/costs are coming down fast, necessary to boost demand/usage along the way.
• Agentic AI will be huge.
• In the AI era, moat is crucial and bigger is better.
• Diversified revenue/cashflow streams are desirable.
• No or low indebtedness is preferable.

From a macro perspective,

• AI is clearly and significantly boosting GDP growth.
• AI is clearly inflationary in some sectors (construction, commodities, power), disrupting other sectors by hording resources.
• The massive hyperscalers expenditures (almost the size of the US defense budget) are redirecting their excess cash from the fixed income markets into the real economy with high multiplier effects. Will productivity gains offset demand-pull inflation?
• Will these financial flows (reduced corporate demand) impact US interest rates?

Only 4 months later, demand is even stronger than expected thanks to Anthropic’s Claude models. Compute demand is even too strong, too quickly, forcing Anthropic (and some others) to tame demand through higher prices and controlled access until more data centers add to compute supply.

The various expected roadblocks have materialized, e.g. power (Power Play Sept 23, 2024), equipment, workers and NIMBY protests, slowing supply growth in the short-term but extending the buildout cycle.

Costs are rising fast, magnified by the US war on Iran and its effects on various supply chains. An increasing part of AI capex is scarcity-induced inflation, positive for suppliers’ margins dealing with scrambling buyers, largely price insensitive.

BTW, this is a dangerous situation for inflation. Consider:

• Exploding AI compute is fueling significant demand other than for chips and memory for resources such as power (oil, gas, solar, nuclear), labor (construction, plumbers, electricians) and specialized hardware (turbines, transformers, cooling equipment). Time is of the essence so hyperscalers are totally price insensitive, a situation likely to persist through 2030.
• Hormuz is forcing countries and corporations to build reserves and hoard various critical resources and materials as protection from constrained supply chains. Hoarding of oil, gas, helium, fertilizers, chips, etc. is now required in an increasingly selfish, uncooperative environment. This will permanently raise resource costs across the world and reduce/eliminate operating efficiencies built during globalization.
• The Ukraine and Iran wars have significantly depleted ammo and military equipment inventories which will need to be hastily rebuilt. At the same time, the world has also learned that alliances can be fickle, incentivizing many countries to find ways and means to protect themselves. Demand for military equipment will be firm and largely price insensitive for several years.

AI and productivity will not help offset the huge, world-wide demand for such a broad spectrum of resources and goods against largely price insensitive buyers.

This was already evident from April’s J.P. Morgan Global Manufacturing PMI:

• Global manufacturing growth accelerates in April despite sharp rises in prices and supply chain delays

The start of the second quarter of 2026 saw rates of expansion in global manufacturing output and new orders strengthen. However, price and supply chain pressures continued to build (…).

Manufacturing production increased for the ninth month running, with the rate of growth hitting a near five-year high. Expansions were signalled across the consumer, intermediate and investment goods sectors.

The increase in production was supported by faster growth of new business. (…)

The resulting input shortages and delivery delays led to a solid upswing in purchase price inflation, as demand exceeded available supply. Average input costs rose at the quickest pace since June 2022 and at one of the fastest rates in the 28-year survey history. (…)

Average output charges rose at the sharpest rate in 45 months, as manufacturers passed on part of the increase in costs to clients. Backlogs of work meanwhile rose for the third month running.

As Ed Yardeni illustrates, the recent jump in the ISM-Prices Paid Index almost guarantees higher PPI inflation …

… which would translate into higher consumer inflation …

… potentially scaring central bankers:

Oil Market in ‘Race Against Time’ on Hormuz, Morgan Stanley Says

(…) “The United States’ 3.8 million barrel-a-day increase in exports and China’s 5.5 million barrel-a-day cut in imports have shielded the rest of the world from 9.3 million barrel-a-day of tightness — a very significant amount,” the analysts said in a section headlined “A race against time.” (…)

“The path matters: a reopening in June with US and Chinese buffers still partly intact is the base case; a closure that runs into late June or even July is the regime in which Brent flat price has to do work it has so far been able to avoid,” they said, referring to futures for the global crude benchmark.

In the bank’s still-current base case, Dated Brent — a physical marker — is seen at $110 a barrel this quarter, $100 in the following three months, and $90 between October and December, with forecasts unchanged. In the bull case — based on a longer closure — prices were seen at $130 to $150. (…)

“Even if the strait reopened tomorrow, the time required to restart fields, repair refineries and reposition tanker tonnage means the market is on track to lose another billion barrels over the balance of 2026,” they said.

But for Ed Yardeni, the race is against earnings:

Raising Our 2026 S&P 500 Target Range Due To Earnings-Led Meltup

We are raising our year-end S&P 500 target from 7700 to 8250. We’ve been bullish on earnings but not as bullish as the recent consensus of industry analysts. We’ve never seen consensus earnings expectations rise so quickly for the current and coming years as they have in recent months. The result has been an earnings-led meltup in the stock market.

Our 2026 and 2027 EPS estimates have been $310 and $350, respectively, since late last year. Those were bullish estimates back then. Consensus EPS estimates have rocketed above our targets in recent weeks. They are currently $336.49 (up 22.0% from last year!) and $386.70 (up 14.9% from the 2026 consensus estimate).

We are raising our EPS estimates to $330 this year and $375 next year (Our $375 year-end estimate for forward earnings is conservative, in case the analysts are too exuberant). We are sticking with our forward P/E range of 18.0-22.0, resulting in a year-end range for the S&P 500 of 6750-8250, assuming (as we do) that forward earnings per share will be will be $375 at the end of this year. The latter is already at $354. (…)

Our outlooks for EPS and RPS imply that the S&P 500 forward profit margin will rise to 15.0 this year and 16.3 next year. These forecasts are a bit higher than the current consensus. (…)

We are now also raising our subjective probability of a continuation of the Roaring 2020s to 80% from 60% simply by merging it with our meltup scenario (previously at 20%). We are doing so because we believe that any meltdown will be a buying opportunity and won’t trigger a recession or bear market similar to the 1999-2000 Tech Bubble and Tech Wreck. We are sticking with 20% odds of a recession that causes a bear market. (…)

What could possibly go wrong? The war in the Middle East isn’t over, though that hasn’t kept stock markets from soaring around the world in April and so far in May. That’s because oil prices have remained around $100 per barrel. The shock waves from the war could still hit the global economy. Another round of fighting could be even more troublesome, as it could result in stagflation. A more persistent inflation problem would force central banks to raise interest rates. The Bond Vigilantes would likely push bond yields higher in this scenario.

Nevertheless, for now, we are sticking with our 10,000 target for the S&P 500 by the end of 2029. It might arrive ahead of schedule.