Shares of Micron Technology are down fifteen percent in early trading Thursday morning, at $88.12, the response to a disappointing revenue forecast for the current quarter offered Wednesday night. The fall was doubtless made worse because the report followed Wednesday nearly three-percent sell-off of the broader market following “hawkish” states from the U.S. Federal Reserve Board.

This is a blip, however, as reasons for the downbeat outlook are hardly epic and all too familiar: There continues to be a build-up of DRAM and NAND inventories among customers of Micron’s in the smartphone and personal computer market, and the automotive and industrial equipment markets for chips continue to grind through a long period of recovery before their appetite for new chip purchases turns around.

“What we're essentially doing is we're attempting to repeat a playbook that we executed with our first large customer at these others.”

Semiconductors stocks will look a lot next year like this year, at least for the first six months, according to Merrill Lynch’s Vivek Arya, who on Monday offered up the firm’s 2025 outlook on chips.

Arya writes that the year is set to be one of “two varying trends,” with the first half of the year all about cloud computing and artificial intelligence, and the second half producing a rebound in the more mundane business of chips that are “cyclical,” namely, industrial and automotive applications.

That focus on AI will probably yield a lot of gains for existing winners, especially as Nvidia is Arya’s biggest proposed gainer, with a potential upside in 2025 of forty-two percent.

I mentioned on Monday that it’s worth considering alternatives to Nvidia in artificial intelligence chips, the most prominent two options being Marvell Technology and Broadcom.

Shares of Broadcom soared on Friday as investors got a new round of enthusiasm for that AI opportunity. The stock closed up twenty-four percent on Friday, for a gain of a hundred and one percent this year.

The company’s earnings report Thursday evening was good even if it didn’t present stunning upside. Fourth-quarter revenue was actually slightly below consensus, despite rising forty-four percent, and the forecast for this quarter’s revenue was merely in line with Street expectations.

But the stunning part was the two hundred and twenty percent growth in revenue from AI chips for the full year, to over twelve billion dollars.

Some recent developments suggest the beleaguered software group of stocks may be coming back into favor, as I noted a couple weeks ago following positive responses to both Snowflake and Elastic. 

The interest in software is showing up again this week, with initiations and upgrades and downgrades on a raft of names.

NOT EVERYONE’S A WINNER

First, the bad news: Some software stocks may be even less desirable in a rebounding software market. 

Wednesday, JP Morgan’s Pinjalim Bora rejiggered ratings on three software stocks, C3.ai, which just reported positively on Monday; PagerDuty, which sold off last month on its rather mixed earnings report; and Sprinkler, which also had a downbeat report last week. 

In earnings news Thursday, shares of Adobe are sinking while Ciena is surging, as the payoff from artificial intelligence is playing out very differently for the two.

Both stories, however, require a leap of faith.

Adobe shares are down over thirteen percent after its forecast offered Wednesday evening for this quarter came in below expectations: Adobe sees revenue rising eight to ten percent, the Street was looking for eleven percent. 

The big picture is that the business is not speeding up from AI as investors had hoped it would, something that had been dismissed in September by the bulls as merely a blip but now cannot be ignored. 

Tuesday’s assessment of Oracle was wild: numerous bullish analysts raised their expectations on the stock following the earnings report Monday evening, even though they cut their estimates for this year and for next year by a few hundred million dollars.
The analysts’ enthusiasm is in contrast to the general investor disappointment, as the stock sold off by seven percent Tuesday.

The disconnect is that the sell-side analysts believe that, no matter what’s happening at the moment, Oracle is a company reborn, with higher revenue growth ahead. Investors are saying, Show me.

The week that ended December 6th was a good one for stocks. The Nasdaq Composite Index rising 3%, the Standard & Poor’s 500 up 1%, and the TL20 group of stocks worth considering up 5% for the week.

It was a good performance from mega cap technology companies. The top gainer for the week was Tesla, rising 13% amidst enthusiasm for the potential payoff to Elon Musk and company from the incoming administration of U.S. president-elect Donald Trump. Second highest, 11% for Broadcom. Taiwan Semiconductor up 10%. Salesforce up 10%. Amazon up 9%. Meta up 9%. ServiceNow, a TL20 company up 7%, and Arista Networks, up 7%.

As I noted in a think piece Monday, Broadcom and Marvell Technologies could be a good alternative to Nvidia for investors looking for AI exposure. And both companies are also an increasing threat to Advanced Micro Devices, according to Merrill Lynch’s Vivek Arya, who cut his rating on AMD stock Monday to Neutral from Buy.

Arya warns that AMD is behind Nvidia in terms of the total GPU buffet, and now Broadcom and Marvell’s work on custom chips for Amazon and others could limit AMD’s AI chip sales.

“AMD's pipeline remains 1yr+ behind NVDA's (which is accelerating) and lacks a competitive networking (switching, optics) portfolio,” writes Arya.

It can be treacherous to bet against the shares of Nvidia, now the world’s greatest chip maker.

For example, the GraniteShares 2x Short NVDA Daily ETF, run by the New York-based ETF company GraniteShares, which promises to give you two-times the inverse of whatever NVDA stock does, has lost ninety-four percent of its value this year as Nvidia stock soared a hundred and eighty percent.

You would have done much better betting on GraniteShares’s other speculative instrument, the GraniteShares 3x Long Nvidia Daily ETP, which is up almost eight hundred percent. (Although, read up on why the ETP is not quite what it appears.)

If, however, you are concerned about Nvidia’s rising stock valuation, it’s worth considering alternatives.

It turns out that betting on artificial intelligence is a bet that can flip wildly from quarter to quarter.

Both Dell Technologies and Hewlett Packard Enterprise had high expectations heading into this earnings season. But Dell’s report last week left investors somewhat disappointed with the level of AI server sales, while Hewlett on Friday is rising by over nine percent after its report Thursday evening prompted a newfound faith in its AI abilities.

The jump in the stock is especially striking given that CEO Antonio Neri had to hold off on forecast revenue for the new year because the company is in the midst of buying networking firm Juniper Networks.

The gains may be short-lived, as the AI business is a precarious one, with lower margins for both Dell and Hewlett as the technological innovation is increasingly dominated by their supplier who is also their competitor, Nvidia.

We are once again in the dregs of earnings season, as some smaller, more obscure names turn in results.

We heard from Credo Technologies on Monday and Pure Storage on Tuesday, who each had news of stunning progress.

And Wednesday night, we got a glimmer of hope from Verint Systems, the cloud-based “customer-engagement” software maker, and a disappointing forecast from Synopsys, the chip intellectual property vendor that hasn’t yet shown the kind of interesting prospects I discussed with CEO Sassine Ghazi in October.

Verint is an example of how trading the earnings bounce can be tricky.

Update: Pure Storage and Zscaler news and views.

Bandwidth chip maker Credo Technology is a small-cap that has been riding the rise of artificial intelligence networking all year long, with numerous analysts singing its praises (see the sidebar.) It tends to take wild swings on its earnings reports, and Tuesday is a big, big up day for the stock, rising forty-five percent.

The surprise in Monday evening’s earnings was a very large one: the revenue outlook for this quarter, $115 million to $125 million, came in forty percent above consensus. That’s quite an achievement considering that all of the analysts covering the stock have a Buy rating on it.

Credo has products that overlap, to an extent, with what Astera Labs does, and Astera has been an absolute darling this year, but so has Credo, up two hundred and seventy percent versus Astera’s gain of a hundred percent since its March IPO.

Shares of Intel are up almost a point on Monday after the company announced that CEO Pat Gelsinger, who took the job almost four years ago, is retiring, to be succeeded by a team of CFO David Zinsner and current head of product Michelle Johnston Holthaus as co-CEOs as the company searches for a replacement.

It’s hard to see this moment as other than negative, and, in fact, Bloomberg’s Ian King, Liana Baker, and Ryan Gouldreport that Gelsinger report what I suspect, namely, that Gelsinger “was forced out after the board lost confidence in his plans to turn around the iconic chipmaker,” citing multiple unnamed sources.

Gelsinger was “given the option to retire or be removed,” they relate.

Sure: If things were going great, Gelsinger would not be retiring now, just shortly before all the work he’s done is supposed to return Intel to preeminence in chip manufacturing. It’s also questionable that he would retire without having already begun a selection process for his replacement.

“It's been four years since I got here, and I think we've created a fair amount of value here for our shareholders.”

There was a lot of optimism going into Tuesday evening’s reports from Dell and HP — too much so, it turns out.

Both stocks are down by about thirteen percent on Wednesday. Which is interesting because the results for Dell were not actually bad at all — take it from a bear on Dell’s stock.

“Results were fine,” writes Krish Sankar of TD Cowen, who has a Hold rating on the shares.

The main blemish was that the company’s sales of artificial intelligence-based servers was down nine percent from the prior quarter, he notes, which is a result of delays with Nvidia’s “Blackwell” chips (even though Nvidia claims the chips are fine.)

But more important than sales are the stats below the revenue line, notes Sankar, and those numbers are strong: “We believe AI server backlog figure of $4.5B (up 17% Q/Q) in CQ3 and orders (up 11% Q/Q at $3.5B) are a better representation of actual demand than revs./shipments.” He also notes that the total sales “pipeline” for AI stuff has risen by fifty percent, and “we estimate it now stands slightly below $20B.”

We’re not done with earnings yet. Today brings Nutanix, Workday, sensor-chip maker Ambarella, CrowdStrike, Analog Devices, Hewlett Packard and Dell

Before we talk about Analog Devices, which reported in the morning, let’s consider the set-up for HP and Dell. Both are reporting after the closing bell today.

Their artificial intelligence businesses will, of course, be front and center. And both have been strong performers this year, with Dell stock up over eighty-eight percent and HP up over thirty percent.

The tea leaves have been looking good for both. Evercore ISI’s Amit Daryanani offered his view on Friday that the computer market should benefit from a trend to “replenish” the computer inventory that’s in the distribution channel, a trend Daryanani thinks should continue next year, “driven by an improving global economy.”

“We’re going to win the enterprise, and we’re going to win across the platform.”

I’ve been writing for a long time that investors were going to come back to software stocks after dumping them in 2022. The first of those hopeful articles was in December of that year. But I was early: Software has continued to be a tough investment group for the past two years.

Hope springs eternal, and last week’s surge by Snowflake and Elastic has the Street percolating with enthusiasm.

Are these two harbingers of something real?

They are certainly a sign that an appetite exists for good software names when those companies stop stumbling and deliver the goods. Jefferies & Co.’s Brent Thill notes over the weekend that valuations for software are up thirty-six percent since the start of October, and the “rebound has been driven by better-than-expected results like SNOW and ESTC last week.”

Daniel Patrick McGahn, CEO of AMSC, also known as American Superconductor, talks about how the company came back from the brink of disaster in 2011, how he has transformed the company from a component maker to a systems maker, and the future of electric ships, the energy grid, and AI data centers.

Wednesday evening was a story of potential turnarounds.

Snowflake was the big star of Wednesday’s reports, which also included Nvidia’s blowout numbers, and an upbeat report from cyber-security firm Palo Alto Networks.

Shares of Snowflake closed up Thursday by thirty-three percent, at $171.35, after the company Wednesday afternoon reported the most upside in revenue in two years, and offered a forecast for this quarter solidly higher as well.

You’ll recall that the concern for Snowflake going into this report was whether the company’s new approach to selling its software would boost growth. The numbers have convinced the Street that the approach has worked.

Late Wednesday evening, some men were sitting around the Russian baths in Wall Street’s financial district talking about Nvidia.

I overheard one of the men impressing his friends with his investing acumen and his knowledge of artificial intelligence.

“My guy got me into it,” meaning, Nvidia stock, “at $40,” the man said. “Before the split?”, his friend asked, referring to this year’s ten-for-one split. “Pre-split,” the man said with emphasis. That would be a nearly ten-thousand percent return over the past eight years. His friends were suitably impressed.

Chip giant Qualcomm held its “analyst day” event on Tuesday in New York, the kinds of meetings with the Street where there is a feel-good vibe about future potential, and a selection of financial targets. The targets seemed to have underwhelmed most observers, and the stock rose just fractionally.

Qualcomm has been for years now trying to focus investor attention on the parts of its business that are not selling chips into smartphones, namely, automotive chips and the Internet of Things. Some of what management said on Tuesday was a repeat of what was said during the previous analyst day event in September of 2022, so it may not have been enough to move the ball forward.

• The week that ended November 15th was a poor one for stocks, with The Nasdaq Composite Index falling 3%, the S&P 500 down 2%, and the TL20 group of twenty great companies whose shares are worth considering, falling worse than the market, down 5% on weakness that included Nvidia.

• While larger names such as Applied Materials struggled, some small cap names such as cybersecurity firm CyberArk really delighted folks.

• A very prominent name that topped the list of gains last week was Bloom Energy, makers of alternative energy products, fuel-cell products. it was highlighted in an interview that I did a year ago with Paul Wick, the fund manager of the Columbia Seligman Group of Funds.

This week’s earnings reports bring three heavy-hitters on Wednesday evening: Nvidia, Snowflake, and Palo Alto Networks — all three of them being part of the TL20 group of stocks worth considering.

Nvidia’s report comes amidst rising estimates and price targets, but also newly reported problems with its forthcoming “Blackwell” GPU chips for artificial intelligence.

The shares are down two points Monday, and are down four percent in the past week, as Qianer Liu and Anissa Gardizy with The Information reported over the weekend that Nvidia “has asked its suppliers to change the design” of integrated computer “racks” it sells with the Blackwell chips because of “overheating problems,” citing unnamed Nvidia employees, customers and suppliers.”

Thursday saw two upbeat presentations by chip equipment giants, but neither presentation did much for the near-term outlook for an industry that’s been in a slump.

In the morning, ASML Holding, the monopoly producer of lasers for etching circuits, held its “analyst day” event, a gathering of stock analysts to hear the company’s promises for the years to come. In the evening, Applied Materials reported its fiscal fourth quarter.

Both companies, and peers such as Lam Research, KLA Corp., Tokyo Electron, and Teradyne, are dealing with uncertainty about how much their customers such as Taiwan Semiconductor Manufacturing will spend. The official view, from the industry trade group, SEMI, is that the total market for “wafer fab equipment,” or, WFE, will rise almost three percent this year, and fifteen percent next year.

This has been another good earnings week for small, young companies, echoing the surge in shares of AppLovin last week.

This week’s stars include CyberArk Software, the thirteen-billion-dollar maker of security and “DevOps” software, whose shares are up five percent on Thursday following Wednesday evening’s report; dLocal, the three-billion-dollar online payments purveyor, surging fifteen percent on Wednesday’s report; Paymentus Holdings, the four-billion-dollar maker of cloud-based billing software, which soared twenty-seven percent on Wednesday following Tuesday night’s report; and Rocket Lab, the nine-billion-dollar maker of launch systems for putting satellites in space, which shot up twenty-eight percent on Wednesday following its own bullish report on Tuesday.

“The grid is being asked to do things it just wasn't inherently designed to do … We’re asking it to do basically everything in our lives.”

Thought for the day: Investing in gold was better than equities for almost a quarter century stretching back to 2000. That’s the main takeaway from a sixty-four-page report published by Deutsche Bank economist Jim Reid on Tuesday.

Reid looks at real equity returns dating back to the start of 1999 and finds that gold’s return of 6.8% beats the return of 4.9% of U.S.-listed equities — the first time gold’s been better in all the quarter centuries going back to 1800. (The actual start of the quarter century was on January 1st of 2000, notes Reid, so there’s actually one more year for equities to pull it out.)

It’s nicely laid out in one chart: