Why Chip Stocks May Peak After Fundamentals, Not Before, Says Truist

The common belief is that chip stocks peak about two quarters before the business does, so you sell when the news still looks good. Truist‘s William Stein argues the history points the other way. His analysis finds semiconductor stocks have typically topped about a quarter after revenue growth peaks, which matters if you are trying to judge where the semiconductor cycle sits right now.

The question is urgent. The Philadelphia Semiconductor Index (SOX), a benchmark that tracks major chip companies, fell about 29% between 22 June and 29 July 2026. It then rebounded about 27% into early October and has since turned lower again.

The cost of getting this wrong runs both ways. Sell too early and you miss what Stein calls a “SuperCycle.” Hold too long and you ride a genuine cycle top down.

Here is how to tell a valuation correction from a real peak, along with the strongest evidence for and against Truist’s view.

What does Truist’s “stocks peak after fundamentals” finding actually claim?

The standard rule says stock prices lead fundamentals. Markets look ahead, so chip shares should roll over well before revenue growth does. Most investors act on that idea without testing it.

Stein tested it. In a note reported by Investing.com on 20 October 2026, Truist found that chip stocks have, on average, peaked about one quarter after revenue growth peaks.

Truist’s core finding Semiconductor stocks have historically topped out around one quarter after fundamentals peak, not two quarters before.

The current numbers make that finding concrete. Truist’s analysis rests on three cycle markers:

• 74%: consensus year-over-year revenue growth expected in Q4 2026
• 15th: the number of straight quarters of improvement that figure would represent
• 29%: the growth rate at which demand typically peaks, usually in the seventh quarter of a cycle

This cycle has already outlasted and outgrown the typical pattern by a wide margin. Stein suggests the Q4 2026 figure could mark the fundamental peak. If the historical lag holds, the stock-peak window opens around Q1 2027, and he expects gains through at least that quarter, possibly longer if growth keeps accelerating.

If 74% really is the top for fundamentals, the risk window for chip stocks sits ahead of you, not behind you. That changes how you should read the recent sell-off.

One caveat applies. The full Truist note is not publicly accessible, so these figures come from secondary reporting. The thesis also breaks if revenue estimates stop rising before Q4 results arrive.

Is this really a SuperCycle, or 1998-2001 all over again?

The parallel is uncomfortable. Chip stocks are rising on a new technology, investors are paying stretched multiples, and growth rates look extraordinary. In May, Reuters reported that the SOX weekly relative strength index (RSI), a momentum gauge that flags when prices may have risen too far too fast, hit its most overbought reading since the March 2000 peak.

Then the numbers start pulling the comparison apart.

What is different this time

The demand shows up as actual purchases of accelerators, networking gear, memory and capacity. World Semiconductor Trade Statistics (WSTS) now forecasts global sales of about $1.655 trillion in 2026, roughly 108% growth. The Semiconductor Industry Association (SIA) reports sales above $1 trillion through August, with the full year projected above $1.6 trillion.

The buyers are different too. Hyperscalers, the giant cloud operators that run vast data centres, fund much of their spending from internal cash flow. Reuters cited a BofA estimate of hyperscaler capital expenditure (capex), meaning spending on long-term physical assets, approaching $1.5 trillion by 2027.

Hyperscaler capital expenditure of roughly $130 billion in Q1 2026 alone, with full-year guidance near $725 billion, is the demand base that Truist’s revenue growth estimates ultimately rest on.

Truist also points to an out-of-phase industrial recovery. Industrial demand is recovering later than AI demand, which stretches the cycle’s length.

What has not changed

Marija Veitmane of State Street Global Markets said in July that the cycle could become “a lot longer,” but that cyclicality won’t disappear. A longer upswing still ends.

Concentration has also sharpened. WSTS expects memory sales to rise about 302% this year, so memory is doing much of the heavy lifting. A powerful aggregate number does not mean every chip stock you own shares the same runway.

What did the summer SOX pullback tell us about cycle timing?

If stocks lead fundamentals, the summer collapse should have been the warning shot. It worked better as a stress test.

The Reuters “just over 20%” reading and the 29% figure measure different dates. The deeper fall came later, and it matches the Truist note.

Reported triggers stacked up quickly:

• Concern that AI capex may not deliver adequate near-term returns
• High valuations and profit-taking
• Higher bond yields and oil prices
• Tariff and export-control concerns
• Competition from Chinese memory makers

The sharpest catalyst was Broadcom. It guided AI semiconductor revenue to about $16 billion, against roughly $17.2 billion expected, and did not raise its full-year AI forecast.

Throughout all of this, industry revenue forecasts kept climbing. Stein called the July decline a buying opportunity, citing strong feedback from AI-infrastructure buyers. He also noted that pullbacks this large which did not become downturns are uncommon, and he sees upcoming earnings as a probable stabiliser.

Earlier SOX drawdowns beyond 20% in 2011, 2015 and 2018 were each a positioning correction inside a wider bull market, which is the precedent bulls cite for the summer decline.

The lesson for you is direct. Price weakness alone does not prove a fundamental peak, and a 29% drop with rising estimates is a different event from a 29% drop with falling ones.

How do you judge cycle timing, and which names does Truist favour?

Truist’s average is not the only history. In 2000, 2018 and 2021-22, chip shares turned before fundamentals visibly rolled over, with inventories, trade restrictions or rates shifting first. The useful response is a set of checks rather than a bet on one historical average.

A framework you can apply

1. Forward estimate revisions. Rising revenue and earnings forecasts support the cycle; sustained cuts are an early warning.
2. Capex growth against AI revenue growth. UBS estimates hyperscaler capex growth could slow from about 76% in 2026 to 25% next year and 6% in 2028.
3. Memory prices and inventories. The forecast leans so heavily on memory that a pricing reversal could cut growth rates fast.
4. Customer concentration. A pause by one or two hyperscalers could hit single-customer suppliers hardest.
5. Orders and lead times. These often turn before reported revenue does.
6. Correction or peak? Ask whether fundamentals are still improving while prices fall.

That last point carries the most weight. A high price-to-earnings (P/E) ratio, which compares a share price with company profits, is weaker evidence of a top than a sustained fall in orders, memory pricing and hyperscaler capex together.

Where Truist’s Buy ratings fit

Stein’s Buy ratings fall into two groups. The AI beneficiaries are Nvidia, AMD, Broadcom, Monolithic Power Systems and TTM Technologies. His conviction has limits: he cut his Broadcom target to $520 while keeping a Buy.

The second group is cyclical recovery names: Vishay, Arrow Electronics, Avnet, Diodes and Sensata. Truist describes these as businesses with more fragile models, tied to the later industrial upswing.

These two groups sit at different points on the risk spectrum. AI leaders depend most on hyperscaler capex holding up, while the recovery names depend on industrial demand arriving on schedule. These are analyst views, not recommendations.

For readers wanting a structured exit plan, our dedicated guide to semiconductor cycle investing sets out a graduated de-risking sequence for memory, foundry and equipment names.

Why the peak call matters less than the evidence that follows it

Truist’s thesis is credible. Fundamentals remain exceptionally strong, and the historical lag places the stock-peak window around Q1 2027 or later. The summer drawdown still showed that prices can break from fundamentals for weeks at a time.

The deciding evidence is whether revenue estimates and hyperscaler capex keep rising after the correction. If they do, the sell-off looks like a valuation reset. If they stall, the stocks-lead-fundamentals pattern of 2000, 2018 and 2021-22 becomes the better guide.

Over the coming weeks, watch three things: earnings and guidance from chipmakers, memory pricing, and hyperscaler capex plans for 2027. Those signals will settle the question faster than any peak forecast. This is analysis of published views, not personalised advice.

This article is for informational purposes only and should not be considered financial advice. Investors should conduct their own research and consult with financial professionals before making investment decisions. Past performance does not guarantee future results. Financial projections are subject to market conditions and various risk factors. Forward-looking statements are speculative and subject to change based on market developments and company performance.