Closing Bell Overtime: Tech Stocks Plunge Markets Search for A Bottom 11/5/26

OpenAI

Across the shows: 48 speakers in other episodes, 29 bullish and 40 bearish calls, past 30 days

• Dan Niles · Niles Investment ManagementBearish

  OpenAI's revenue growth is decelerating sharply, so the multiple investors pay for it is questionable.

  Niles cites OpenAI's sequential revenue growth falling from about 140% in Q2 to about 45% in Q3, with Q4 possibly at 20-25%. He argues that revenue growth across the AI sector is massively slowing, which he has discussed before. He asks what multiple can be put on that growth and what OpenAI will do next, implying the valuation case is weakening.

  “OpenAI growing a 140% sequentially in q two, then it's like 45% in q three. Q four, it might be like 20%, 25%.”at 12:32

Anthropic

Across the shows: 105 speakers in other episodes, 131 bullish and 88 bearish calls, past 30 days

• Dan Niles · Niles Investment ManagementBearish

  Anthropic may struggle to IPO at the valuation people expected or to raise as much as hoped.

  Niles questions whether Anthropic can go public at the valuation people had in mind and how much it can raise. He says those figures are going down, because the AI sector's revenue growth is slowing and the market has repriced in the past two days. The point is brief and not backed by hard numbers.

  “And is Anthropic gonna be able to get public at the valuation people thought, and how much can they raise? Right? Because those figures are going down.”at 12:55

VanEck Semiconductor ETF

Across the shows: 8 speakers in other episodes, 2 bullish and 7 bearish calls, past 30 days

• John Kolovos · Macro Risk AdvisorsBearish

  The semiconductor ETF has broken down on a double top, pointing to a drop toward about 375.

  Kolovos says the SMH broke support at the June and July lows with a breakaway gap, which completes a double-top formation. He reads this as a longer-term trend change in the semiconductor group. The implied downside target is around 375, roughly where the ETF traded during the March lows this year. He calls that the base case for now and says the market must reverse before the view changes. He also calls it a significant breakdown and notes the DRAM ETF may need another leg lower. He links this to the 2018 analogy, where semis peaked in June and declined while the S&P kept making higher highs.

  “When you look at the SMH, the downside implication of that is down to around three seventy five, which is approximately where it was trading around the March lows of this year.”at 46:10

US stock market outlook

• Dan Niles · Niles Investment Management

  Market still has too much froth; stocks are not cheap and the top has not been seen yet.

  Niles argues the market still shows too much positivity. He says the S&P 500 is only down about 6%, so there is a lot of froth to work off. Valuations on normalized earnings are not cheap, and he doesn't believe the top has been seen. He sees a possible near-term rally but expects the market to need a longer reset. Winding down of a $16 billion multi-strategy fund is a start toward a healthy reset.

  “I feel like there's still too much positivity in the market... there's still a lot of froth that you could say needs to get worked off.”at 7:15

• Lori Calvasina · RBC Capital Markets

  Without a Q3 or Q4 dip, history suggests only about 6% next-year returns, versus 16-18% after a dip.

  RBC looked back to 1970 and found that in most years there is a down quarter in Q3 or Q4, sometimes both. This year there was no down Q3. When the market does not dip in Q3 or Q4, the following year averaged only about a 6% return. When it does dip, the next year typically delivers a 16-18% move. Calvasina says the setup is not the same as the dip-to-rip narrative investors had hoped for.

  “most years in three q or four q, sometimes both, you get a down quarter. ... if you don't dip in three q or four q, you tend to only have, like, a 6% return in the next year. If you do, then you get, like, a 16, 18% type move.”at 15:46

• Lori Calvasina · RBC Capital Markets

  RBC has cut year-ahead earnings 10%, up from 5%, in case the capex cycle slows.

  RBC's rolling 12-month forward forecasts had already included a 5% earnings cut for lingering war impacts. Clients began worrying about the capex cycle over the summer, so RBC added another 5%, to a 10% total cut. That addition hedges the risk that the capex cycle is slowing and AI-related earnings are too optimistic. Calvasina also notes PE multiples compress when upward revisions decelerate, and falling rates ease that pressure.

  “We had previously been baking in sort of 5% cut to year ahead earnings, and then we basically took that to a 10%. ... in case people are right that the capex cycle is starting to slow”at 16:38

• Lori Calvasina · RBC Capital Markets

  Market is only in early pain; a true 10% drawdown depends on whether investors fear recession.

  Calvasina's framework treats a 5-10% decline from the high as the start of a true drawdown. The S&P was down about 5.6% from its all-time high, so she says the market is only in early stages of feeling pain. Reaching about 7,000 on the S&P would put it near a 10% decline. The key question is whether investors fear a recession. If they do not, the Iran war bottom at about 9.1% could be a reference point, implying the decline is near its end.

  “we were down 5.6% from the all time high. In my framework, five to 10% is where you really start to have a true drawdown. ... The question people have to ask is, do we fear a recession or not?”at 18:40

• Keith Lerner · Truist Advisory Services

  Risk-reward for US stocks is improving over six to twelve months after a deep tech valuation reset.

  Lerner sees the market in a 'buyer strike' with selling still under way, and he wants stabilization first. He argues that the tech forward P/E has fallen from about 32 a year ago to about 17, a contraction of roughly 45-50%, which he calls the deepest PE contraction this decade and a move from euphoria to fear. He says earnings momentum is still the strongest in the market outside energy, though he acknowledges earnings are at risk. He also notes that since this bull market began, tech has roughly doubled the S&P's returns, with six pullbacks of more than 10% and an average pullback of 17%. Today's roughly 17% drawdown is therefore in line with this bull market's average correction. Because historical rebounds after such corrections have averaged over 40%, and fear can flip to greed quickly, he thinks a little good news could go a long way. He is not calling a bottom, but he believes risk-reward has improved considerably over six to twelve months.

  “when you have historic selling pressure, you have valuations that have been reset tremendously ... And I think as you look out six to twelve months, the risk reward has actually improved quite a bit.”at 40:37

• John Kolovos · Macro Risk Advisors

  Crash risk is elevated; the S&P 500 is likely to fall to about 7,000 and stay capped below 7,725.

  Kolovos says the S&P 500 is testing its 200-day moving average and that more damage is likely. Once the June lows break, the next major support is around 7,000, the highest level since January. He expects the tape to struggle beneath 7,725 for the foreseeable future. He does not see today's drop as a run-of-the-mill pullback. His work shows financial conditions are still very tight, and that raises the chance of a market crash. He draws a close parallel to 2018: a midterm election year, a new Fed chair being tested, and semiconductors peaking in June and falling steadily while the S&P made higher highs. He thinks the market is forcing the Fed's and possibly the administration's hand. For a bottom, he wants to see a real capitulation, with VIX near 30-40, a spike in new lows, and a higher pull-call ratio.

  “I think the biggest macro risk here is, at least according to my work, that financial conditions are still very very tight. And when that happens, the chance of a market crash are very elevated.”at 48:08

AI chips & semiconductors

• Dan Niles · Niles Investment Management

  Memory stocks are not cheap: they trade over 5x book and would bottom near 1x in a downturn.

  Commentators calling memory stocks cheap are wrong in Niles's view. They trade at more than five times book value, while in a down cycle memory stocks bottom at about one times book. He says they are currently over-earning, and end-demand can change very quickly. He also says valuations on a normalized earnings basis, not peak 80% operating margins, are not cheap for these stocks.

  “They're trading at over five times book value. In a down cycle, they bottom at one. They're over earning.”at 11:06

• Lori Calvasina · RBC Capital Markets

  Semis earnings upward revisions are decelerating; more pain likely needed before the group bottoms.

  RBC downgraded tech on October 14 after being overweight since midsummer. It pulled back for two reasons: broader tech valuations looked pricey, and the rate of upward earnings revisions for semis and semi equipment in the Russell 3000 was decelerating. Calvasina says that deceleration seems very early and more pain is likely needed. Fewer upward revisions tend to compress PE multiples.

  “the other reason was that rate of upward revisions on the semi and semi equipment group ... was starting to decelerate. That seems very early days in terms of that deceleration. We're probably gonna need to see some additional pain there.”at 17:46

• Keith Lerner · Truist Advisory Services

  Semiconductors are ground zero of the selloff and must turn first before the market finds a floor.

  Lerner says semiconductors led the market down and remain its epicenter. He points out that on the previous day, although the tech sector looked strong, semis were down beneath the surface, apart from some hyperscalers. Over the last twenty-five years of cycles, he has looked for signs such as green appearing in unexpected areas, like a semiconductor name, as evidence of relative-value buying. Because the group leads both the decline and any recovery, he says it needs to show leadership on the upside before the market can be considered on firmer ground.

  “I do think semis are ground zero. They led the way, down then.”at 41:16

AI infrastructure & data centers

• Dan Niles · Niles Investment Management

  Post-midterm pushback on data center construction is a durable shock, unlike reversible tariffs or war.

  Texas, Georgia, Ohio and Pennsylvania all flipped in the midterms, and those states account for about 50% of incremental US data center capacity coming online. Senators calling right after the midterms to stop all new data center construction will constrain buildout. Niles contrasts this with tariffs or the war, which one decision can reverse. Midterm outcomes cannot be re-voted, so he expects this to take time to sort out. He cited an Australian data center developer whose stock fell 40% after announcing community-driven delays before recovering when two other firms invested.

  “when you have senators coming out right after the midterms and saying, we need to stop all new data center construction, this is what you're gonna get.”at 6:17

• Dan Niles · Niles Investment Management

  Neoclouds with leveraged balance sheets and no other business will mostly go to zero; hyperscalers survive.

  Neoclouds are the overflow capacity for hyperscalers such as Microsoft and Google, and they are the first to suffer when those customers cut back or when neoclouds cannot get capital or permits. Niles says they have leveraged balance sheets, massive negative cash flows and no other businesses, so they will likely go to zero, most of them. He argues the industry is overbuilding, as in past industrial booms such as railroads and the internet. Microsoft, Amazon, Google and Meta will survive because their cash-generating core businesses (software, search, ecommerce, social) can absorb the bust. He also noted that if neoclouds cannot deliver capacity, hyperscaler cloud results such as Azure and Google Cloud could miss their numbers.

  “The neo clouds have leveraged balance sheet, massive negative cash flows, no other businesses.”at 9:44

AI bubble & tech valuations

• Dan Niles · Niles Investment Management

  Current selloff resembles the 1997-98 funding crisis: semis fall 40%, finish both years up, then rip.

  Niles says the market is in a 1997-98 type funding crisis, comparing it to the Asian currency crisis of 1997 and the Long-Term Capital Management collapse of 1998. Semis fell about 40% from peak to trough in both those years but still finished each year up, and then ripped higher in 1999. He thinks the current episode is like that, though he says the other side of the story is that revenue growth is slowing sharply, so the multiple investors will pay is uncertain. He has not seen the top.

  “I think we've seen a 1997, 1998 type of funding crisis... Semis went down... they went down 40% from peak to trough. Both years, semis finished the year up.”at 12:03

• Keith Lerner · Truist Advisory Services

  Tech valuations are resetting rather than bubble-like; current P/E sits below where the ChatGPT-era bull market began.

  Lerner acknowledges the question of whether this is a bubble and that it requires a judgment call. He compares the current cycle with the late 1990s, when P/E expanded more than 300% over four years before the top. By contrast, the P/E for this bull market is below where the ChatGPT bull market began. His base case is that data center spending will likely slow down, but demand remains there over time, so valuations will reset. He expects the market is pricing in a large earnings reset, and while he expects some slowing, he thinks it will not be as severe as what is priced in. He calls this the next thing to watch.

  “the PE for this bull market is actually below where the CHAT GPT bull market began.”at 44:09

US rates & the Fed

• John Kolovos · Macro Risk Advisors

  The Fed should not hike; a hike into this selloff would be a disaster.

  Kolovos argues that market stress is forcing the Fed's hand. Financial conditions remain very tight, and his 2018 analogy has a new Fed chair being tested. He thinks the market will push the Fed toward not hiking. He says a hike would be a disaster, given the cyclical weakness in semiconductors and the risk of a market crash. Bond yields are well off their highs and oil is falling, so he sees the market rather than the data as the main pressure on policy.

  “So I think the market, like I said earlier, I think is really gonna be forcing the Fed's hand here to not hike. If they hike, it'll be a disaster.”at 48:50