|
Chip stocks rallied after July's crash. U.S. land clearing, as tracked by Vulcan, stayed below its spring pace. |
|
WHAT YOU'LL GET IN THE NEXT SIX MINUTES The three Vulcan signals that did not confirm the summer bounce in AI stocks, and the three things to watch after the November elections. |
The AI trade is sliding for the second time this year. According to Fortune, the Philadelphia Semiconductor Index (SOX) set a record on June 22 and then fell 21% in July, its worst month since October 2008. It bounced, closing September about 12% above its July 31 close. Now it has fallen again: Bloomberg reported that the SOX dropped about 10% over six straight sessions, putting more than $1 trillion of chip market value on track to vanish. The index closed October 14 near 11,700, roughly 20% below its record by our calculation.
According to news reports, the October sell-off was driven by financial concerns: a hot inflation report, worries about AI debt, and a report that OpenAI's revenue run rate was nearer $50 billion than the $68 billion some had circulated. In our view, none of that is a construction signal. Land clearing is.
While stocks recovered through August and September, Vulcan data showed land clearing had not followed. To compare months fairly, we use what Vulcan had recorded nine days after each month ended. On that basis, U.S. land clearing averaged about 2.0 GW a month from March through June. July came in at 1.0 GW and August at 0.5 GW. September improved to 1.3 GW but was still about a third below the spring average.
Figure 1. Chip stocks rebounded from late July while U.S. land clearing stayed below its spring pace. The bottom panel shows each month's U.S. land clearing as recorded in SynMax Vulcan nine days after the month ended, so every month is compared at the same age.
BEYOND CHIPS
Chips are only half the story. The sharpest losses are in the neoclouds, the companies that rent out AI computing capacity and borrow heavily to build it. On October 14, according to TipRanks, CoreWeave fell 15.5%, IREN 12.2% and Nebius 10.1%, against 4.3% for the SOX. CoreWeave closed near $66, about 56% below its 52-week high, after a chart of its 2032 bond yield nearing 15% circulated widely on social media (CNBC). By our calculations, as of the October 14 close, Nebius was about 38% below its June 22 high and IREN about 60% below its high.
Figure 2. Distance below the 52-week high at the October 14 close. The debt-financed neoclouds have fallen two to three times as far as chips. Nebius and CoreWeave closes are approximate.
The worry has moved from chips to balance sheets. Reuters described a growing crisis of confidence in an AI build-out financed with borrowed money, and Gartner has reportedly predicted that most neoclouds will lose their independence by 2030. Demand is not the problem. CoreWeave reports a backlog of about $104 billion. The question, in our view, is whether sites get built and energized fast enough to service the debt.
A neocloud pays interest from the day it borrows and earns revenue only when a site is energized. So every month a project sits behind a moratorium or a stalled clearing permit is, as we see it, a month of cost with no revenue. This is a SynMax hypothesis, not a finding about any named company.
Dollars bounced. The dirt didn't. Three Vulcan signals point the same way. We list them in the order they appeared.
SIGNAL 1
Land clearing is the first visible step of any data center project. It is the earliest physical proof that a developer is really going to build. Vulcan is updated weekly, so we can graph progress, or the lack of it.
Cleared U.S. capacity climbed steeply through the first half of 2026, adding about 25 GW. Since June 30, Vulcan has recorded only about 4.6 GW more, and the line has flattened.
Figure 3. Cumulative U.S. data center capacity land-cleared (GW), daily, January 2025 to October 9, 2026. The curve flattens from mid-2026.
SIGNAL 2
Vulcan also includes our friction tracker. It logs the moratoriums, approvals, denials, and protests happening across the country. In our data, as friction climbed through 2026, monthly land clearing fell.
We don't think friction is the only cause, but the timing lines up closely.
Figure 4. As friction rose through 2026, monthly land clearing fell. Orange: GW land-cleared per month (left axis). Green: active friction actions (right axis). Blue: new friction actions per month (right axis).
SIGNAL 3
You might have looked at construction and concluded everything was fine. For a while, that was true.
While new development paused, projects already underway kept moving. First structure starts (the first steel and concrete on site) kept accelerating through the second quarter (in the U.S., 6.7 GW in the first quarter and 8.9 GW in the second), which let some observers conclude everything was going well. But that healthy-looking number sat on top of a thinning pipeline.
In the U.S., about 50 GW of the roughly 70 GW cleared has now reached first structure, a ratio of about 70%. That is close to the early 2025 peak of 72.2%. A rising ratio means construction is working through the backlog of cleared land faster than new clearing replaces it. The backlog of cleared U.S. capacity not yet at first structure has shrunk from about 24 GW on June 30 to about 21 GW. At the third-quarter pace of first structure starts, that is roughly eight months of work. First structure starts have also eased, to about 7.7 GW in the third quarter from 8.9 GW in the second.
Figure 5. Cumulative U.S. capacity by first structure date (GW), daily, January 2025 to October 9, 2026. Growth is starting to bend.
Figure 6. U.S. first structure / land clearing ratio (%). At about 70%, it is near its early 2025 high.
WHAT HAPPENS NEXT
November is around the corner. In our view, some of these moratoriums may be lifted after the elections, given the economic value data centers can bring to the regions that host them. This is a SynMax opinion, not a prediction or a forecast of any outcome.
Why would communities change course? We think much of the friction comes down to local choices, and those can change. Our Placing the Bust paper made the case with real data from Loudoun County, Virginia, the densest data center cluster on earth.
|
6.2 GW operational in Loudoun today, up from 261 MW in 2010 |
-30% residential property tax rate, from $1.145 to $0.805 per $100 (2016 to 2026) |
~35x property tax offset vs. the electricity increase, by our own arithmetic |
Loudoun's residents pay more for electricity. They also pay far less in property tax. Mangum Economics estimates that without data center revenue, the median home would owe roughly $5,800 more per year. The Dominion increase approved by Virginia regulators is about $165 per year at its peak. By our arithmetic, the two numbers are not close.
Figure 7. Loudoun County: capacity up, resident taxes down. Source: SynMax Vulcan / IIR; Loudoun County. Adapted from Figure 15 of SynMax's Placing the Bust.
The lesson from Loudoun is that friction is mostly a governance problem. Where the local bargain is structured well, data centers fund schools, roads, and lower tax bills. Where it is structured badly, communities push back. That may be why opposition clusters in some places and not in Loudoun.
|
THE CAVEAT Even with every permit granted, transformers, substations, and interconnection queues run on multi-year lead times. Loudoun shows the constraint can be solved over time. It does not show that it is dissolving now. In our paper's data, 166 of 174 friction signals were still tightening. |
YOUR PLAN
If a turnaround comes, we expect Vulcan data to show it in roughly this order. This is our view, not a prediction. Watch for all three.
|
01 Land clearing Does the weekly pace turn back up? This is the first signal and the earliest to move. |
|
02 Friction Do moratoriums lift and denials fall in the states that matter most? |
|
03 First structure ratio Does the ratio roll over from about 70% as new clearing refills the pipeline? |
We'd treat one signal turning as a headline and all three as a turnaround.
NEXT STEP
Through the third quarter, the dirt did not confirm the bounce. If a turn comes, we expect it to show up there as well. Those are the three signals we are watching, and the order we expect them to move in.
|
Talk to David Bellman about Vulcan See how Vulcan tracks land clearing, friction, and first structure across the data center build-out. Email: dbellman@synmax.com |
SOURCES
Fortune (Aug. 2, 2026): SOX June 22 record; July decline of 21%, worst month since October 2008.
Bloomberg (Oct. 14, 2026): SOX down about 10% over six sessions; more than $1 trillion of market value at risk.
Nasdaq, Investing.com, Barchart: SOX daily closes and the 14,655 record.
CNBC (Oct. 14, 2026): CoreWeave bond yields and neocloud sell-off.
TipRanks (Oct. 15, 2026): One-day moves for CoreWeave, Nebius and IREN on October 14.
Yahoo Finance (Oct. 14, 2026): CoreWeave about 56% below its 52-week high; IREN quote data and 52-week high.
Weiss Ratings (Oct. 14, 2026): Nebius 52-week high of $299.86 set June 22, 2026.
Reuters (Oct. 15, 2026): CoreWeave debt and AI build-out financing concerns.
24/7 Wall St. (Oct. 15, 2026): Reported Gartner prediction on neocloud independence.
Financial Times, via secondary reports: OpenAI revenue run-rate figures.
Mangum Economics; Virginia SCC: Loudoun property tax and electricity estimates, as cited in SynMax's Placing the Bust.
SynMax Vulcan: Datacenter land-clearing and first-structure milestone dates and capacity, latest snapshot as of October 9, 2026 (U.S. units only in Figures 1, 3, 5 and 6); friction data (Figure 4).
IMPORTANT DISCLOSURES
Not investment advice. This material is provided by SynMax Inc. for general informational and educational purposes only. It is not investment, financial, legal, accounting, or tax advice, and it is not an offer, solicitation, or recommendation to buy, sell, or hold any security or financial instrument. SynMax is not a registered investment adviser, broker-dealer, or research analyst and has not considered the objectives or circumstances of any reader.
Data and sources. Market data and company information come from the public and third-party sources listed above. They are believed to be reliable, but SynMax does not guarantee their accuracy or completeness. Stock prices and index levels are as of the October 14, 2026 close and will not be updated after publication. Some chart values are approximations, and percentage declines marked as ours are SynMax calculations. SynMax Vulcan data are subject to revision, and land-clearing figures for recent months are preliminary and typically rise as dates are added to the database later.
Opinions and forward-looking statements. Statements about what may happen, including possible changes to moratoriums after the November elections and the future direction of land clearing, friction, or first structure, are SynMax's opinions and hypotheses. They are not predictions or guarantees, and actual results may differ materially. A relationship between Vulcan data and market prices does not establish causation, and past data relationships may not continue.
No affiliation. Company, product, and index names are the property of their respective owners and are used only to identify them. SynMax is not affiliated with, and is not endorsed by, any company named.
Investing in securities involves risk, including the possible loss of principal. Consult a qualified professional before making investment decisions.