Final Thoughts: What to watch if you're looking for a peak

If you’re in that camp, you’ll probably say that demand will continue to surprise to the upside. You might also say that fall outages could be longer than usual because some spring outages were cut short by mills looking to cash in on a good spot market. You might also note, as I have in recent Community Chats, that this is the strongest market since 2021 by some metrics.

You’re not wrong. Here are just a few examples from SMU data. Our lead time for hot-rolled coil is slightly above seven weeks on average, something we haven’t seen since 2021. (Our historical data on lead times is here.) Sheet inventories are the lowest they’ve been since May 2021. (That data series is here.) And we haven’t seen 43-44% of survey respondents reporting improving demand since 2021. (That’s here.)

Meanwhile, people have underestimated how long this current upcycle will last and how high prices might go for most of 2026. (That said, it seems sentiment has changed over the last month. More people now appear to be in the “stronger for longer” camp. That might be its own indicator. But I digress.)

My point is this: We’re getting into rarified air with HR prices a little over $1,100 per short ton (st). More than a few of you seem pretty sure we’re headed to $1,200/st or higher. And I’m guessing even some of the most bullish bulls out there might be leery of restocking at these levels.

Again, it’s been a great year for steel, which is awesome. But let’s say you load up at $1,200/st and then the market turns. It might be a not-so-great year-end.

Rocking and rolling like ’21. Yes, but…

The steel market might be as good as it’s been since 2021. And this might be a long-overdue manufacturing boom. Or at least a boomlet. Even so, it’s not on par with what we had in 2021.

Take lead times: Seven weeks is well above a more typical average of four to five weeks. But do you know where HR lead times were in June 2021? Nearly 11 weeks on average. I realize some US mills are out roughly that far. But they’re the outliers. Others are at least starting to catch up.

It’s a similar theme when it comes to mill negotiation rates. Only 13% of survey respondents said mills were willing to negotiate lower spot HR prices in our last survey. That’s historically low. But throughout most of 2021, that figure was in the single digits. And some weeks it was zero.

Yep, you read that correctly. There were weeks in 2021 when even the contrarians among our survey respondents said there were no deals to be had.

Is supply starting to catch up to demand?

I wouldn’t be shocked if we start to see lead times plateau. On a very basic level, domestic mills are producing more steel now than they were just a couple of months ago.

US mills produced 38.9 million tons of steel through the end of May, according to American Iron and Steel Institute (AISI) figures. That’s up from 36.5 million tons through the first five months of 2025. That amounts to roughly 480,000 tons per month more this year than last. If that pace were to continue for the balance of the year, it would amount to an additional 5.76 million tons.

To put that in context, look at capacity figures for U.S. Steel’s mills. It’s a little smaller than adding a new Big River Steel (annual capacity of 6.3 million tons per year) but a lot bigger than a new Edgar Thomas Plant (annual capacity 2.9 million tons per year).

I realize that’s apples for oranges. AISI figures include everything – not just sheet and plate but also rebar, beams, wire rod, etc. And, yes, tariffs have structurally changed the US landscape to the point where we probably don’t need as many imports as we used to. Even so, that’s no small increase.

Meanwhile, imports are also ticking upward. The US imported 1.81 million metric tons (mt) of steel in May, according to license data from the US Commerce Department. That’s not a big gain compared to 1.79 million tons in April. And, yes, we’re still below levels we saw in the first half of 2025. That said, we’re up nearly 27% from 1.43 million tons in December 2025.

In other words, imports have been creeping up steadily along with US prices – and even with 50% Section 232 tariffs still firmly in place. (Whether they will remain so is a matter of debate as USMCA negotiations heat up.) And I don’t know why imports wouldn’t continue to go up if lead times remain extended and if people think US prices will continue to rise.

Also, certain mills seem to be struggling to catch up to demand. And at this point, I know some buyers are wondering whether they can. It seems reasonable that some of those buyers might turn to imports in part to guarantee supply.

Boom or boomlet

If you think prices will remain high for the rest of the year, I think the big question is whether demand will become as strong as it was in 2021. I know a lot of people who think demand is the best it’s been in years. I don’t know many who think it rivals the post-pandemic boom.

Take automotive, the most concentrated market for steel in much of the US. It was the posterchild of a surging post-pandemic market in 2021. CRU, SMU’s parent company, presents a cautious outlook for the North American auto industry. It’s a good read. But it does not read like a 2021 forecast.

There are, of course, some caveats. You could say the trend lines showing higher US production and lower US sales are a recipe for trouble. You could also make another case: that those trends instead reflect automakers pivoting to make more vehicles in the US in response to tariff policies.

My point is this: Whichever side you’re on, demand for new cars, trucks, and SUVs is not so hot – and production so limited (remember the chip shortage?) – that people are willing to pay above list price. In 2026, in contrast, the word of the year seems to be “affordability”. (Honestly, I can’t stand that term. “Too expensive” works just fine.)

What about construction? No one is going to question it if you say you’ve got strong demand related to data centers. And as the technology proves itself, barriers like local opposition and power limitations might slow the buildout, but not stop it. And extending the cycle is not a bad thing. (Side note: If you think data/AI are in bubble territory, then that’s a whole different game. Then we’ve got not only a stock market bubble but also a bubble in the market for steel and other materials. That, however, is a little beyond the scope of SMU to assess.)

I think the bigger concern is what happens to the broader construction market if inflation stemming from the Iran war remains elevated. The market was hoping for rate cuts this year. Maybe some of you have heard of a new Wall St. buzzword: “NACHO”. It’s short for “Not a Chance Hormuz Opens”. In other words, what happens to construction (outside of data centers and projects related to them) if rates don’t come down?

Does CSP carry more weight now?

Such concerns aside, I’m guessing spot prices will continue to tick upward in the short term. Sure, scrap might be sideways in June. But the relationship between scrap and steel isn’t very tight when the market is strong.

Steel prices tend to take on a momentum of their own in an upward market. And I’m guessing Nucor will continue to inch spot prices up by $5/st or $10/st every Monday. Why? I was going back through SMU’s price announcement calendar. Nucor hasn’t been sideways with its consumer spot price (CSP) since early/mid-January.

Meanwhile, no other US mills have announced an increase since last fall. Cleveland-Cliffs, which had been aggressive with price hikes in the past, hasn’t rolled one out since June 2025. What would happen if Nucor kept CSP unchanged? My guess is it would take on more significance now than a year ago.

What to watch in SMU data

I’ll be watching for any signs of lead times flattening out or declining. I’ll also be watching for any signs of service center inventories ticking back up again. (We’ll post updated lead times on Thursday. And we’ll release May inventories to our premium subscribers next week.)

It’s hard to say whether we might see a peak coming into view this month, next month, or sometime in the fall. But it’s fair to say we’ll probably see it first in lead times or in service center inventories – and well before we see prices inflect.

With that in mind, let’s go back to where we started, with 2021 data. History doesn’t repeat. But it sometimes rhymes.

Negotiation rates remained in the single digits until September 2021. In other words, people didn’t report mills as being more willing to negotiate lower spot prices until around the same time that prices started falling.

HR lead times, in contrast, flattened out at around 11 weeks in late May/June-July 2021 before notching sharp declines in August – well before HR prices started to fall in September 2021.

What was the earliest sign of a peak? Service center’s sheet inventories bottomed out at 41.4 days of supply in March 2021 before gradually ticking higher. In other words, they moved even before lead times did. I’m biased, of course. But that is what I would call an edge.

Upgrade to premium!

I’ve said it before, and I’ll say it again: Proprietary data like our service center inventories is a great reason to consider upgrading your subscription from executive to premium. If you’d like to, let us know at smu@crugroup.com.

NexGen Leadership Awards

You know what’s coming at you faster than the SMU Steel Summit on Aug. 24-26 in Atlanta? The deadline to nominate young executives for our NexGen leadership award. In fact, the deadline is June 8. Yep, today, if you’re reading this on a Monday.

The award will presented live at Steel Summit. So if you know a star at your company who is 38 or younger, give them a chance to shine at the event. You can nominate them here.