The week that ended October 10 ended on a sour note for markets with renewed U.S.-China trade tensions pushing the Nasdaq Composite Index down 2% and the Standard & Poor’s 500 index the same.

What continues to dominate the tech scene is discussion of the astounding slew of deals by OpenAI with just about everyone — Nvidia, Advanced Micro Devices, Oracle, etc. 

I mentioned on last week’s podcast that there was a danger of OpenAI being a kind of structurally weak point in the global economy as they do deals that have been regarded as “circular” — OpenAI gives buys from the vendors and also is in some way financed or linked financially to the vendors, so that money the reasoning for a deal is not really sound.

That subject is taken up by Richard Waters of the Financial Times in a long piece over the weekend, “How OpenAI put itself at the centre of a $1tn network of deals,” in which he quotes various investor types on the risks of the circular deals. One of the facts that is important that comes out of the piece is the potential for systemic financial risk as debt financing is put into data center projects that are propelled by AI.

This is a busy week for those corporate dog & pony shows that companies hold in between earnings season, with Dell and Autodesk both holding their annual briefing for analysts, and OpenAI holding its developer conference in San Francisco, all three of which generated a lot of coverage the past twenty-four hours. 

DELL STILL HAS TO PROVE AN AI PROFIT

With these analyst meetings, it’s all about keeping track of what was said before, at the last meeting, and what’s being said now.

In the case of Dell, the update from management is encouraging, especially as regards profitability, which has been a concern given the company’s rush into selling low-margin artificial intelligence server computers.

Do we need to start referring to OpenAI by the phrase “Too big to fail” like we used for the banks in the Great Recession of 2008?

The last podcast, September 22nd, followed the news Nvidia would invest a hundred billion dollars in OpenAI in order to help the latter to purchase Nvidia chips. 

The latest bubble-seeming incident is Monday’s announcement that OpenAI will purchase perhaps as much as $110 billion worth of GPUs from Nvidia rival Advanced Micro Devices. 

This deal comes with a strange extra clause: OpenAI will take a ten percent stake in AMD. For AMD, it means a huge lift in the company’s revenue potential. 

But now a lot is riding on a company that as The Financial Times points out, is on the hook for twenty-three gigawatts of capacity incoming years, over a trillion dollars in commitments.

Shares of Advanced Micro Devices are the latest to get the “OpenAI Bounce,” as I call it, surging almost thirty percent after the company said Monday morning it has struck a deal for OpenAI to deploy six gigawatts worth of AMD AI chips in return for a “performance-based warrant” for a hundred and sixty million shares of AMD stock, which at Monday’s price of $208.71, is worth about thirty-four billion dollars. 

This is the latest commitment by OpenAI to billions in chip capacity following its deal two weeks ago to buy a hundred billion dollars worth of chips from Nvidia, and a week prior to that, reportedly, with Oracle to use three hundred billion dollars worth of AI capacity in Oracle’s cloud.

This may be one of the most striking AI deals so far, in that the warrant is is “directly tied to increasing AMD stock price milestones, with the final tranche vesting at a price of $600 per share,” said AMD CEO Lisa Su on a conference call this morning. 

OpenAI, in other words, is on the hook to boost AMD’s share price to get value from the warrants it is receiving. 

You’re probably well aware that this time of year, early October, is the season for incessantly following the “lead times” of Apple’s iPhone. That’s the time it takes, on average, for a phone ordered online to be delivered. It’s a rule of thumb the Street uses to gauge how popular the latest iPhones are, in this case, the iPhone 17 models introduced two weeks ago.

Jefferies & Co.’s Edison Lee opines that things are going pretty well, but he still cut his rating on Apple stock on Friday to Underperform from Hold, after concluding that the new design introduced this time, the “Air,” is “the least popular by our tracking,” and that, “this new form factor has not helped AAPL driven iPhone sales.” 

More important, Lee is looking not this year’s crop, but next year, to a rumored “foldable” iPhone, something that Apple watchers such as Bloomberg’s Mark Gurman have been writing about for a long time now. 

It’s expected by some a foldable iPhone may show up next year, and if so, it would be a first from Apple in the foldable category, going head to head with models from Samsung Electronics and Google, among others. 

It would appear there’s great appetite among retail investors for curated stock picking, especially when it comes to artificial intelligence.

The Dan IVES Wedbush AI Revolution ETF, owned by Wedbush Securities, which you can trade under the “IVES” ticker, informed by the equity research of longtime Wedbush analyst Dan Ives, hit three quarters of a billion dollars in value a week ago after starting with just a million dollars four months ago.

Since inception, through Friday morning, the fund is up thirty percent, nicely topping the seventeen percent of the Nasdaq Composite.

“We expected it to be very successful but this is a lot quicker than I anticipated,” says Cullen Rogers, portfolio manager for the IVES. “I think there’s a confluence of events there, which is the general market performance, and AI buzz, but, the retail acceptance of it has been fantastic.”

Enthusiasm for quantum computing comes and goes — this year, it’s way up. Stocks of the most prominent names, IonQ, D-Wave Quantum, and Rigetti Computing, have, on average, doubled in price.

But, the fundamental challenge for all quantum companies hasn’t changed in two decades. They are all trying to get to “scale,” and it’s not clear when, if ever, any of them will.

I’m emphasizing this point because the scaling issue is fundamental to all computing; without scaling, nothing is meaningful.

Scaling means that as a computer gets more complex — basically, bigger — the amount of work the computer can do increases in proportion. The main example is the integrated circuit, which became more and more powerful over six decades, increasing dramatically what could be done, to the point that you now have a supercomputer in your pocket with today’s phones.

Quantum is not there yet. All the companies have shown interesting machines, but none have proven they can scale those machines the way the traditional semiconductor scaled.

What’s the value of a company with no revenue?

Barclays analyst Christine Cho writes Monday that it all depends on the headlines.

Cho initiates coverage this week of Oklo and NuScale, two nuclear power hopefuls, giving Oklo an Overweight rating, and giving NuScale an Equal Weight rating. Her main argument is that the news cycle is what will drive the stock prices.

“Generally,” Cho writes, “we think that the macro news, such as policy or trade updates we get from the Administration (which tend to be more positive than not), and headlines around how the world is short power, will be the largest drivers to stock price reaction while announcements for any binding agreements should also act as a positive catalyst.”

A week ago, we heard about Nvidia’s plans to dump a hundred billion dollars into OpenAI, one of Nvidia’s biggest customers, in return for an enormous commitment to buy data center chips over some unspecified period of time. It seemed to me, and to a lot of others, a bit of “circular” business reminiscent of the DotCom days, and a red flag in that respect.

But the market is undeterred. Nvidia dipped a bit following that report, but is higher Monday afternoon at $182.09 amidst continued confidence in the AI trade.

One undaunted enthusiast is Blayne Curtis of Jefferies & Co., who on Monday reiterates his Buy rating on the stock and raises his price target to $220 from $205.

Curtis offers one of the first increases that I’ve seen in estimates based on the OpenAI deal. 

The SPAC is back!

So-called “special-purpose acquisition companies,” or, SPACs, also known as blank-check companies, were the rage until the 2022 plunge in the market. A blank-check firm issue shares to the public, then uses the money raised in that offering to go on the hunt for some private company to buy, thereby using a back-door process to take companies public that might not have otherwise come to market.

SPACs never went away, but enthusiasm died down a bit. Now, it seems there’s a resurgence of SPACs around artificial intelligence. 

Bloomberg’s Bailey Lipschultz relates Friday that “The SPAC ecosystem is growing again,” noting that “So far this year, nearly 100 blank check firms have pooled $19.3 billion, surpassing the total volume for 2023 and 2024 combined, according to data by SPAC Research.”

The poster child for the SPACs, Chamath Palihapitiya, saw his latest blank check, amusingly named American Exceptionalism Acquisition Corp., listed on NYSE under ticker “AEXA,” make a strong debut Friday morning, rising over eight percent. The offering, underwritten by Santander, raised $300 million to go after “businesses operating in the energy production, artificial intelligence, decentralized finance and defense industries.”

“The number seven [vendor] buying the number two, the dynamics of this are intriguing to see how you make it successful.”

Everything that sounds questionable about the artificial intelligence trade in recent days — the decision by Nvidia to invest $100 billion into OpenAI for example — sounds more more convincing, more realistic coming from Micron Technology CEO Sanjay Mehrotra. 

Tuesday evening, Mehrotra extolled the benefits of AI. “Artificial intelligence is also a powerful productivity driver for Micron,” he said. “We are using AI throughout the company across product design, technology development, manufacturing.”

As one example, “In manufacturing, we have driven a five-times increase in [semiconductor] wafer images analyzed in the past year,“ he said, improving the amount of defect-free chips Micron produces.

Mehrotra has good reason to trumpet AI: it is dramatically lifting the business. He presided overly a quarterly earnings report by that was pitch perfect: The company not only exceeded expectations, it exceeded the already high preview of its results it had offered last month. And demand for both DRAM chips and NAND flash memory are rising faster than previously expected this year.

The week ended September 19th was a good one for tech, the Nasdaq Composite Index rising 2%, the Standard & Poor’s 500 rising 1%, and much of it had to do with the big announcement by Nvidia that it will take a four-percent stake worth $5 billion in Intel, its arch-rival, and that the duo will make custom chips for the data center and for laptop computers based on Intel’s x86 technology.

Intel ended the week up 23%, giving it a return this year of 43%. Nvidia ended the week down fractionally, but its return this year is still very good at 37%.

However, the spending by Nvidia is raising questions. The Intel deal was followed on Monday, September 22nd by Nvidia’s announcement it will make an investment of a hundred billion dollars over some unspecified period of time in OpenAI, makers of ChatGPT, contingent on the latter buying and deploying at least ten gigawatts’ worth of Nvidia chips.

Nvidia is making news as much with its largesse these days as its chips.

Monday, the company announced a plan with ChatGPT creator OpenAI for the latter to deploy “at least” ten gigawatts of Nvidia’s chips, in return for which Nvidia will invest a hundred billion dollars in the startup, 

The news follows Nvidia’s announcement last week it will take a ten-billion-dollar stake in chip rival Intel.

Nvidia shares were up one percent on the news, at $178.97.

To put that in perspective, New York City pulls five gigawatts of power on an average day, according to those in the data center business who track these kinds of mammoth usages. Ergo, the two companies are planning to deploy two very large metropolitan areas’ worth of AI chips.

The deployment of ten gigawatts constitutes “millions of GPUs for OpenAI’s next-generation AI infrastructure,” said the two companies in prepared remarks. 

Intel shares are up twenty-three percent Thursday afternoon after it announced with Nvidia that the latter will take a five-billion-dollar stake in Intel, equivalent to four percent at Wednesday’s close, and the two will develop custom chips for artificial intelligence servers and a new breed of laptops.

The stock bounce gives Intel a fifty-three percent gain for the year, topping Nvidia’s thirty percent to date. Nvidia shares rose four percent, while ARM Holdings, perceived as put out in the cold by this, is down four percent. Advanced Micro Devices, now the odd one out for the x86 world, declined by one percent.

The Street reaction has so far been very upbeat for both companies, although I suspect that the balance of the account is squarely in Nvidia’s favor.

IonQ, one of the three publicly traded “pure plays” in quantum computing, along with Rigetti Computing and D-Wave Quantum, last Friday held its “analyst day” to brief sell-side analysts on its technology roadmap and financial goals. 

Most of the points made by CEO Niccolo de Masi and his team were the same points that de Masi made in my interview with him last month; you could say that our chat was a preview of what was heard last week. 

You can catch the whole day’s presentation in the archived recording on the IonQ investor Web site. You can also download the deck of accompanying slides. 

The talks were very well received, with the shares closing up by eighteen percent on Friday. At a recent $68.27, IonQ shares are up sixty-three percent this year.

At a high level, the company reiterated its outlook for this year’s revenue, $82 million to $100 million. And de Masi again compared IonQ to IBM as the only company that he contends comes even close to competing with IonQ. Analysts so far have not made any dramatic changes to their estimates for IonQ’s financials.

We are not in a bubble of artificial intelligence, but that doesn’t mean things can’t get perilous in the AI trade.

So implies Bernstein analyst Mark Newman in an initiation of coverage on Monday spanning a hundred and twenty-three pages in which Newman reflects on the AI payoff for Apple, Dell and other vendors of IT hardware such as server computers, disk drives and smartphones.

Newman thinks the opportunity for hardware vendors is “massive,” though he also sees the risk there could be a period of “digestion,” which brings to mind thoughts of the DotCom collapse.

“We see sustainable increase in spending as AI expands the addressable market for IT and creates a potentially massive opportunity for IT hardware to return to growth,” writes Newman.

Newman’s report is filled with lots of great detail, reflecting his many years covering the hardware beat. His contribution is to put some serious numbers around the AI trade and where it is going.

Newman has three main points.

We have come to the end of earnings season, with some significant news from two of the last companies to report, Oracle and Adobe, really the first companies of a new earnings season, they just report very early.

Oracle’s stock soared by 36% on Wednesday after its Tuesday-evening report contained the stunning disclosure that it had more than quadrupled its annual backlog of business, its “remaining performance obligation,” or, RPO, promising sharply increasing cloud revenue for the next five years.

Oracle is up 75% for the year through Friday’s close. The pop in Oracle stock gave the newly rebalanced TL20 group of stocks to consider a forty-percentage-point gain since the rebalancing, now up 273% for the year through Friday’s close.

Almost immediately, questions arose about that backlog. 

Adobe shares are rising three percent in early trading at $361.90 following Thursday evening’s fiscal third-quarter earnings report, and, if it holds, this would be the first time in five quarters the stock hasn’t sold off. 

The drama of Adobe is by now well understood. The worry on the part of investors is that the company will see its franchise eroded as more and more average users, people not trained on its Photoshop and Premiere tools, turn to generative artificial intelligence to do their design work.

The stock has been an absolute dog as a result, down forty percent this year and the same amount over a twelve-month period. 

The praise keeps coming today for Oracle as its shares soar by thirty-five percent to $325.04 after its stunning surprise Tuesday night in which it announced an enormous backlog of business of nearly half a trillion dollars.

One new convert Wednesday is Merrill Lynch’s Brad Sills, who Wednesday raises his rating to Buy from Neutral, and hikes his price target to $368 from $295, and writes that the backlog of business is “exceptional,” and “suggests that Oracle has become a key AI compute platform.”

Profit is an open question, as I noted last night, but is not a deal-breaker for Sills.

“Although profitability of AI workloads remains a key debate, it is clear that Oracle is capturing share in the large and rapidly growing market for AI infrastructure (we estimate that the AI applications industry alone will represent $155 billion by 2030),” he writes.

“Oracle is clearly leveraging a number of advantages in its cloud software/hardware businesses to attract the largest of the AI enterprises, including OpenAI, xAI, Meta, NVIDIA and AMD). These visible reference customers should help position Oracle to capture share of AI compute.”

When I added Oracle to the TL20group of stocks to consider last week, I was not expecting it would have a blowout earnings report. Frankly, I was not thinking about Tuesday evening’s report, which crept up on me.

But here we are, the stock is up twenty-six percent in late trading at $305.12. (Update: The stock jump is above thirty percent Wednesday morning at $317.30.)

It’s a bit of an odd report because it contains a big surprise. The revenue and profit were just okay, as is often the case with Oracle. Revenue growth of twelve percent slightly missed expectations, actually, and profit was lower than expected as well.

What was jolting was that Oracle’s backlog of business, known as “remaining performance obligation,” or, RPO, more than quadrupled, year over year, and more than tripled from the prior quarter, to just under a half a trillion dollars versus expectations for $149 billion, mostly for its Oracle Cloud Infrastructure, its cloud computing service.

Several billion dollars’ worth of new contracts for cloud computing had come during the quarter, CEO and CFO Safra Catz indicated. “We have signed significant cloud contracts with the who's who of AI, including OpenAI, xAI, Meta, NVIDIA, AMD and many others.”

Catz had until this time been a little bit of the girl who cried wolf: she talked a lot about big business coming down the line from AI but it kept being another quarter out. Well, it has now come to pass and people such as myself who rolled our eyes are rather stunned.

Monday brought the initiation of coverage of Bullish, the seven-billion-dollar Bitcoin trading exchange that went public on August 13th, and which has subsequently dropped by twenty-six percent.

The shares fell four percent Monday to $50.12 as two of the three lead underwriters for the IPO, JP Morgan and Jefferies & Co., came out with a Hold rating, while Citigroup started it at a Buy.

Bullish is closer to the financials industry than to tech per se. It makes the majority of revenue in fees and in the narrow difference in price that is the spread between bid and ask in the sales of crypto-currency it facilitates. It is, however, a tech business in the sense that any market maker brings to bear technology that is supposed to lower costs to transact and improve price discovery, etc., and it appears Bullish has very good software and systems for doing all that.

It has a smaller, second business lead by CoinDesk, the well-regarded crypto news site, which Bullish bought in 2023, which sells data, brand advertising and tickets to its popular crypto conference. 

Bullish, which was brought public with the help of special-purpose acquisition company, a “SPAC,” a blank-check firm, Far Peak, is among the biggest IPOs this year, at $1.2 billion in proceeds from the offering, all of which the company took possession of in crypto-currency!

It’s smaller than CoreWeave’s $1.4 billion IPO, but ahead of software maker Figma, which raised just under $400 million, and Circle Internet, another crypto venture, which raised $254 million.

Friday was a nice day for two of the TL20 stocks to consider, software maker Samsara, which soared seventeen percent, and chip giant Broadcom, which jumped nine percent, both propelled by their respective earnings reports on Thursday evening. 

I had had a reservation about Broadcom for a long time until I put it into the TL20 in February. My reservation was that CEO Hock Tan spent billions on software firms that were tired assets, which was to my mind a way of milking fading assets for not good purpose. 

It’s good that I got on board because all that is rapidly becoming irrelevant. With surging growth in Broadcom’s artificial intelligence chips, Broadcom is poised to see AI revenue go from a third of its total revenue today to seventy percent in two years time. 

Broadcom is becoming an AI-chip company overall, in other words, the same transformation that happened to Nvidia in the past several years. 

Thursday is an interesting contrast of old and new tech. Shares of Ciena, the last remaining big fiber-optic networking firm, is soaring by eighteen percent, to $116.29, while Figma, the software maker that went public on July 31st, is down nineteen percent, at $54.56.

Ciena’s revenue, almost four percent above expectations, was the largest quarterly upside in two years, prompting Evercore ISI’s Amit Daryanani to title his report “Shock & Awe.”

The key point is the company’s forecast for 2026 revenue growth of “high-teens,” notes Daryanani. Both the cloud computing operators and Ciena’s traditional telecom company customers are buying big, he notes. The “webscale” part of the business that serves the cloud saw sales rise by more than ninety percent. But phone companies are also buying, invest in building out telecom infrastructure for data centers and for “AI-to-edge” computing.

This week was the sixth rebalancing of the TL20 group of stocks to consider, and it was the biggest change so far since the TL20 was inaugurated in July of 2022. 

The group has been changed from a market-cap-based weighting to a “capped” weighting in order to alter the influence that different stocks have on the group. 

Specifically, Nvidia drops from being almost half the value of the group to just twenty-five percent.  

Nvidia’s 982% return since the start of the TL20 (through the close of August 29th) has helped the group handily beat all benchmarks, including the Nasdaq Composite, the S&P 500, Cathie Wood’s ARK Innovation ETF, the ARKK, and the iShares Expanded Tech-Software Sector ETF, the IGV.

But that soaring performance made the TL20 way too dependent on Nvidia.

Shares of Alphabet are up eight percent Wednesday, at $229.46, after U.S. District Judge Amit P. Mehta, who a year ago found the company guilty of violating Section 2 of the Sherman Act in order to maintain its monopoly in mobile search, on Tuesday avoided imposing the strongest remedies requested by the U.S. Department of Justice, such as forcing Alphabet to spin off Google’s Chrome browser unit or share data about its advertising business.

This is taken as the best-case outcome for Alphabet, though it is likely to be appealed. 

The larger question is just what is the competition in the market. Mehta indicated that he was measured in imposing remedies because Google faces more competition in search services than it has in decades, especially from artificial intelligence-driven chat bots such as privately-backed Perplexity.

So, is that true? Well, maybe not so much in practice.

It’s time to put less weight on Nvidia, as marvelous as the company is. The giants of the AI trade still have very good growth ahead of them but probably less upside in price. Corning is one name with a favorable valuation being added to the group.

Plus ça change plus elle reste la même.

A year ago, I wrote of the earnings reports of Dell Technologies and Marvell Technology, “Marvell and Dell boosted by AI,” and although both stocks are down today, both are still all about the artificial intelligence trade.

The worries about both of these companies caused broad weakness in the market Friday, helping to push down shares of Nvidia and others.

My sense is that the blemishes for both companies are really specific to where those two companies are at, their unique challenges and opportunities, and are not reflective of some kind of overall breakdown in the AI equipment market.

Allow me to explain each in turn.

Wednesday is a nice night for four members of the TL20 group of stocks worth considering. Shares of Snowflake and Pure Storage are both surging in late trading following upbeat reports. While Nutanix and Nvidia are both trading down, the results and outlook are also strong.

The bottom line is that all four of these businesses are doing very, very well.

First, Nutanix. I spoke following the report with Nutanix CEO Rajiv Ramaswami. He first rattled off the numbers that matter to investors.

“Solid top line growth, solid profitability, 2,700 customers added over the year, I think that's quite substantial,” says Ramaswami. He cites also that revenue rose eighteen percent for the full year, and “annualized recurring revenue,” or, ARR, rose up seventeen percent, while free cash flow rose “significantly” to twenty-nine percent.

The company beat by a wide margin its own forecast offered a year ago of fourteen percent revenue growth and twenty-three percent free cash flow margin.

The idea that commercial software companies will be destroyed by artificial intelligence because AI will automatically create code is a theme I’ve returned to a many times this year. 

It’s a fear that hangs over software investing. Even cool-headed investors such as Paul Wick of Columbia Seligman are paying attention, as Paul mentioned in my interview with him last month.

This week, three prominent software analysts offered their thoughts on what the threats are and what might happen to the industry.

KeyBanc’s Jackson Ader offers the first of a series of promised notes. He identifies three threats. Two are concrete, namely, that AI can both lower barriers to entry and displace commercial software altogether. A third thing, harder to conceive, is “the collapse of the app,” the blurring of software functionality into one giant app: