In these summer weeks between earnings reports and tariff cliffs, the Street turns to its current obsessions, things such as artificial intelligence agents and stablecoins.

An emerging fixation of the Street is the notion that “agents,” AI programs that plug into databases and other existing programs, will help companies better monetize AI.

Merrill Lynch’s software analyst Brad Sills on Tuesday took a shot at sizing the market for agents, which he concludes could be a “compelling” $155 billion total addressable market (TAM) by 2030.

“To our knowledge, this would make our estimated 2030 TAM the Street/third-party research high, and by a significant margin,” he writes. Other figures he’s seen from firms such as Boston Consulting Group, in the neighborhood of $52 billion, “are definitely too low,” he avers.

Elon Musk on Monday introduced a few of Tesla’s robo-taxis to Austin, Texas, a maiden voyage that has been anticipated since last fall’s unveiling of the self-driving technology.

The rides were offered to select “influencers,” according to Reuters’s Norihiko Shirouzu and Abhirup Roy, and the whole experiment involves “a small trial with about 10 vehicles and front-seat riders acting as ‘safety monitors’,” they relate.

The response from the Street is rather muted, and it’s clear there are many more fans of Google’s Waymo service than of Musk’s company.

Ben Rose of the boutique Battleroad Research estimates that Waymo is already making a quarter of a billion dollars in annual revenue on millions of self-driving rides in multiple markets. Google is probably also losing hundreds of millions of dollars on the service, he estimates, which shows you the power of being able to subsidize such a project.

Welcome back from the U.S. Juneteenth Holiday weekend. 

Stocks continue to make gains through Friday's close, June 20th, with the Nasdaq Composite Index closing the week up fractionally. 

The Nasdaq is now up 2% for the month of June so far, the Standard and Poor’s 500 Index up 1%. The TL20 group of stocks to consider is up 6% for the month. 

Both indices are up slightly, year to date, building on May's big gains.

Welcome to a quiet summer Friday of a holiday week, when yours truly seeks to catch up on some missed coverage of stocks earlier in the week.

First stop, a large initiation of coverage on chip equipment stocks, which have been very mixed this year, the group roughly flat. As you can see in the table, most are pretty far from their fifty-two week high prices, so the group has been generally out of favor.

David Dai of Bernstein Research on Monday took a fresh look at the matter and came up with what I think is an intriguing contrarian view: Dump ASML, buy BE Semiconductor.

Dai appreciates the strengths of both companies (both of which are in the TL20 group of stocks to consider), but he’s focused on what happens to two great franchises as we move out in time.

He thinks ASML becomes less important than it is now, and BE more so. He starts ASML at Neutral and BE Semi at Outperform.

Sometimes the zeitgeist wins out over feelings about individual stocks.

Shares of chip-maker Marvell Technology have been a dog for a while now, down thirty-two percent this year at a recent $75.47, and up just three percent in twelve months, as quarterly results have failed to show the kind of upside that investors wanted.

But all is forgiven this week as the company Tuesday gathered analysts for an upbeat discussion of “AI infrastructure,” chips for running artificial intelligence.

The analysts came away delighted, and the zeitgeist of the AI-chip trade is now working in the shares’ favor: the stock is up Wednesday by seven percent, bringing gains for the week so far to almost twelve percent.

Marvell, you may recall, competes with Broadcom in helping the tech giants make custom chips for AI, known as “ASICs,” with Amazon being its most prominent customer. 

Well, that was quick: I interviewed Fastly CEO Todd Nightingale last month, and late yesterday it was announced that he’s departing to become chief operating officer of Arista Networks. In his stead, Fastly’s head of products, Kip Compton, is taking over as CEO. 

This often happens to reporters: They get played, in a sense, when an executive has a big change coming up and is seeking to build momentum by making media appearances. Oh, well, live and learn!

Nightingale replaces Anshul Sadana, who left Arista a year ago after eighteen years at the company. 

Investors seem to hate both of these new appointments, bidding down Arista by almost five percent today, to $90.83, and bidding down Fastly by almost nine percent, to $6.60. 

The analysts covering the stocks seem to be more content with the switch. 

I tend not to write about healthcare or biotech names because I don’t have a background in health or life sciences, and I believe the factors that drive those companies and their stocks are very different from those that drive areas such as computer networking and semiconductors, where I’ve spent over thirty years reporting.

However, I need to make accommodations more and more as technology seeps into every field and every industry. And some of the tech analysts I regularly follow are themselves expanding their coverage into life sciences for the right names.

Case in point, my curiosity was peaked with the initiation of coverage Monday of newly public Hinge Health, an eleven-year-old startup in San Francisco that claims to be changing the nature of outpatient care using artificial intelligence.

The appeal of the stock for many is that it is a kind of software stock, but trading at a cheaper valuation despite high growth at the moment.

Everything was “solid” for Adobe in Thursday evening’s fiscal second-quarter earnings report — that’s the term all the analysts used.

But solid results are not convincing anyone who has been worried about Adobe’s future. The stock fell five percent on Friday, the third quarter in a row that the shares have declined on the report. Adobe is down fifteen percent in the past twelve months. And the three-year, annualized return is a paltry three percent.

As I explained in this week’s podcast, the Street is concerned with whether creative professionals, those who grew up on Photoshop and the rest, are enough to offset what is expected to be a gradual migration away from Adobe of the less-dedicated users to artificial intelligence tools, things such as ChatGPT.

The question seems to be whether Adobe’s power in the visual creation market will be eroded by AI. 

There are two versions of the database giant Oracle, the good one and the bad one. Thursday is seeing the good version triumph, the stock up thirteen percent at $199.74, following Wednesday evening’s fiscal fourth-quarter earnings report.

Price targets are zooming, with Brad Zelnick of Deutsche Bank having what I see as the highest target, at $240.

The bad version of Oracle, which we saw the prior two quarterly reports, is a version of Oracle with sales growth that is underwhelming. The good version of Oracle is a company with tons of potential, in the form of the key non-GAAP metric “remaining performance obligation,” or, RPO, a measure of its backlog.

The numbers, actually, were not bad, with the revenue upside the biggest in two years — but only after brokers had cut their own estimates in the past six months. The forecast of $15.05 billion, at the midpoint, is just fractionally above the average estimate for $14.991 billion.

Shares of Nvidia are down fractionally Wednesday as the Street seems relatively unmoved by the keynote in Paris of CEO Jensen Huang at his company’s first-ever Euro version of its “GTC” event, which had taken place in the U.S. in San Jose back in mid-March, and which was also received rather tepidly at that time. 

I think Nvidia is in some ways such a broadly understood story that these kinds of massive conferences by Huang don’t have the ability to surprise much. 

And, to be fair, a lot of what was announced sounded to me very familiar. For the actual news items, you can check out Nvidia’s official blog post. 

Among the points discussed by Huang were that the company’s top of the line machine, the “GB200 NVL72,” is being produced by Nvidia’s partners at a rate of 1,000 a week, which is a nice reassurance that business is moving along at a steady clip. 

Perhaps the most buzz=worthy quote from Huang was his enthusiasm about quantum computing. Huang said the field is “reaching an inflection point,” and that “We are within reach of being able to apply quantum computing, quantum classical computing, in areas that can solve some interesting problems in the coming years.”

The saying is “sell in May, and go away,” and they haven't done that yet. For the week ended June 6th, the Nasdaq Composite Index rose 2%, as did the S&P 500 Index, and the TL20 group of stocks to consider was up 4%.

Helping the positive sentiment was a bullish report from Broadcom, and very positive remarks on Micron Technology.

But there were also reminders that the tariff issue has not gone away. Shares of Ciena fell sharply following its earnings report on Thursday, June 5th. Ciena's CFO, James Moylan, who is retiring, said that Ciena's profit was partly impacted by having to absorb an expense related to tariffs totaling mid single digit millions of dollars last quarter. Bulls advised that you look past that to all the brilliant opportunities for Ciena with fiber-optics and data centers.

Shares of Intel are higher by almost nine percent Tuesday afternoon, at $22.27, on no particular news — at least, not news that seems significant. 

The company said that Imperial College London has chosen Intel’s Xeon chip for its next supercomputer. Intel is apparently working with SoftBank on a new kind of memory chip for AI. And tomorrow, Wednesday, Intel’s CEO, Lip-Bu Tan, will discuss the roadmap for the company’s chips at an event called the “supplier summit.”

None of that seems eight-percent worthy. I’m left with a couple of random items that could have gotten someone excited. 

Will the third time be the charm for Oracle? 

The stock has lagged the software group this year, up only six percent, at a recent $177.15, albeit up a very healthy forty-one percent over the past twelve months. The underperformance is the consequence of two quarters in a row of disappointment from a revenue standpoint, the most recent one in March.

Oracle reports Wednesday evening, and a chorus of analysts arose Monday with some mixed views: Jefferies’s Brent Thill, BMO Capital’s Keith Bachman, Morgan Stanley’s Keith Weiss, and Citigroup’s Tyler Radke.

The most bullish is Thill, who has a Buy rating on Oracle stock, and who raises his price target to $200 from $190.

Thill writes that his survey of twenty of Oracle’s software resellers shows that two thirds “either hit or exceeded their plans” in the quarter, and almost half “cited q/q pipeline improvement.” 

“We need to change the mindset of forty years, fifty years of how data is maintained and stored. This is very hard.”

Things keep humming along at Broadcom, which on Thursday evening reported results that topped expectations, and for its outlook as well. 

However, the disappointment that has sent the stock down three percent in early trading, to $254.60, appears to stem from the company failing to talk about what’s in store for 2027. 

The bulls are putting that aside at the moment, figuring there’s plenty of time for Broadcom to talk about fiscal 2027 given that we are only halfway through fiscal 2025 (ending in October.)

The growth, meantime, is off the charts. Broadcom beat its own target for artificial intelligence-related chip sales last quarter, which had been forty-four percent growth, year over year, instead delivering forty-six percent, at a total of $4.4 billion. And for this quarter, that rate is going to speed up to sixty percent.

It was a report very much consistent with the positive trend seen in March, only more so..

Kudos to Raimo Lenschow of Barclays, who on March 30th took the side of database vendor MongoDB, arguing the shares were too cheap after what had been a twenty-five percent decline since the start of the year. 

Thursday, the stock is up twelve percent at $222.73, after the company’s earnings report Wednesday evening topped expectations. 

Lenschow takes a victory lap today, raising his price target to $270 from $252, noting that the principle element in the results was the company’s “Atlas” database, which makes up over two thirds of quarterly revenue, seeing its sales growth speed up for the first time in over two years, from twenty-four percent sales growth last quarter to twenty-six percent this time around. 

That’s a big relief for investors, as the product had been experiencing a profound slowdown for over two years now, the principle worry for the stock for a long time. 

Returns for investing, as opposed to random speculation, come over many years, as the best names are given a chance to work.

The TL20 group of stocks to consider is up 190% Thursday since its inception on July 15th of 2022, setting a new high for the group, surpassing the 187% achieved on January 22nd. That’s across five rebalancing events since the original group. 

That return is 2.7 times the return of the Nasdaq CompositeIndex in that same time. 

• The week that ended May 30th, an abbreviated week with Monday’s Memorial Day holiday in the US, was a good one for stocks, with the Nasdaq Composite Index rising 2%, lifted by most of large cap tech, with Oracle being the best performer, up 6% for the week, and other strong names such as Broadcom, Samsung, Texas Instruments, Meta, Nvidia, Apple, etc.

• We’re past the thick of earnings season, the tariff talk has cooled a bit, and tech has come roaring back. As just one indicator, the TL20 group of stocks to consider ended the week up four percent for the year versus the one-percent decline of the Nasdaq and a half-point gain of the Standard & Poor’s 500. 

Equity research types love puns, in good times, and in bad. 

The news Tuesday is very good for shares of Credo Technology Group, which is soaring by over eighteen percent, at $72.47, the exact opposite of the plunge in the stock back in March.

With a pun at the ready, TD Cowen’s Josh Buchalter declares that “CREDOn’t have anything to pick on.” 

In the March earnings report, the problem was that among Credo’s three biggest customers, Amazon, Microsoft, and Elon Musk’s xAI, Amazon made up an amazing eighty-six percent of revenue last quarter, raising the fear of customer concentration. 

This time around, Credo not only beat expectations, as it usually does, it said Amazon dropped from eighty-six to sixty-one percent of revenue.

Buchalter and others rejoiced.

“Our business plan is to drive growth both from this broadening customer footprint, and from our largest customer, in parallel.”

The White House’s budget proposal, which goes by the much-disputed name of the “One Big Beautiful Bill,” has some areas of concern for renewable energy names such as First Solar, whose stock is already down fifteen percent since the start of the year.

But, have no fear, writes Mizuho Securities’s Maheep Mandloi, who reiterates his “Top Pick” designation on First Solar on Monday, after concluding that the bill is “less dire than consensus fears.” 

The fear, as mentioned by The Washington Post’s Theodoric Meyer on Monday, the bill “would scrap credits for [renewable energy] projects that don’t start construction within 60 days of Trump signing the bill into law,” an effort to swiftly roll back tax credits that were given to the group by Democrats in 2022. 

For the past year, Nvidia has been going after Arista Networks in the ethernet networking market, leveraging its dominance in artificial intelligence in order to sell its own flavor of ethernet networking, called Spectrum-X.

One happy possibility is that there doesn’t have to be one winner, but rather both Nvidia and Arista will thrive because the whole market will grow “dramatically” in the next five years, according to Amit Daryanani of Evercore ISI.

Daryanani on Wednesday offered a thirty-six page primer on what’s called the “back-end” AI networking market. The back end of AI means the connections that happen between groups of GPU chips, either along the rear of a cabinet full of racks of the chips, or between the individual racks, or even long distance, between AI data centers.

Two of the most talked-about items from Wednesday’s earnings are, of course, Nvidia and Salesforce, with their shares Thursday going in opposite directions, Nvidia up four percent at $138.52 and Salesforce down five percent at $261.91.

Price targets are rising for Nvidia today, and some of the numbers in the forecasts are getting to be simply staggering. Merrill Lynch’s Vivek Arya, reiterating a Buy rating, and raising his price target to $180 from $160, models profit more than doubling from this year’s projected $4.37 a share.

“Our FY28E (effectively CY27E) pro forma EPS of $7.23 is aligned with $277 billion in data center sales and $325-$350 billion TAM [total addressable market] (at 80-85% NVDA share),” writes Arya. “If we use a higher industry TAM assumption of $450-$500 billion and same NVDA share, it paves the way for $10+ in CY27/FY28E pro forma EPS, conceptually.”

Update: Sure enough, there was not much upside to Nvidia’s report, with the smallest revenue upside in two years, 1.7%. And the sales outlook for this quarter, $44 billion to $45.9 billion, actually came in below the consensus estimate, $45.8 billion, for the first time in over two years.

But the stock is up five percent in early trading, and price targets are rising to the $210 mark. Were it not for U.S. restrictions on the company’s sales into China, Nvidia would have booked another eight billion dollars in the quarter. So, the happy result is that things turned out just fine, with revenue from the data center, i.e., chips for artificial intelligence, rising by seventy-three percent, to $39.1 billion. 

That’s a pretty staggering amount of growth at that scale in dollar terms. You’ll get no complaint from me: this business continues to perform very, very well. 

I’ll be back in a little while with commentary from the Street.

Nvidia is one of the TL20 stocks worth considering, and it is up an amazing 755% since I put it in the first cohort, in July of 2022.

• Welcome to an abbreviated trading week, with markets closed for the Monday Memorial Day holiday in the U.S.

• The preceding week, the week ending May 23rd, was a poor one for stocks. The Nasdaq Composite Index declined 2% and the Standard & Poor’s 500 declined 3%.

• Mega-Cap tech generally traded down, though Alphabet was a bright spot, thanks to its “Google I/O” conference, and the discussion of how it will be spreading its “AI Mode” text summaries throughout its paid search business. That was well-received by analysts.

• The proximate cause for stocks’ decline was the budget bill passed by the U.S. House of Representatives, which observers expect will drive the U.S. national debt from a hundred percent of the economy to a hundred and thirty-four percent in a decade from now.

Informatica is a company that has had a different lifetimes. It was founded in 1993, taken private in 2015 by a private equity group, and then spun public again in October of 2021.

And on Tuesday, the company began a new chapter, as software giant Salesforce said it agreed to acquire the company for eight billion dollars, or $25 per share in cash. The stock had already jumped seventeen percent on Friday as Bloomberg reported on the potential for the deal.

It’s not surprising to me that Informatica is being taken out. For one thing, the deal has been rumored for over a year now. For another thing, Informatica has minimal revenue growth, as it has a large component of legacy software contracts that have been in long-term decline. That leaves Informatica as less desirable to software investors than peers with still high growth such as Snowflake.

You can read about the whole history of Informatica in my many interviews with CEO Amit Walia, the most recent of which was last summer.

For Salesforce, which also has minimal revenue growth, the deal is seen by the Street as an attempt to bolster “agentic AI” that has yet to really “move the needle” for Salesforce.

Shares of chip giant Analog Devices sold off by nearly five percent on Thursday, a striking decline given it followed the seventh quarter in a row the company’s results have topped consensus, and the fifth quarter in a row its outlook was better than expected — in both cases, by the most upside in years.

I wrote in February that the company was finally seeing a recovery from roughly two years of malaise in its markets for chips for a diverse set of products in the industrial and automative markets. The numbers Thursday morning confirmed that. 

It’s a simple story: chips are cyclical, they go through periods of boom and bust, and so when a market has burned down inventory and is ready to buy more, that’s a very clear positive for a chip-maker such as Analog Devices.

This has been a busy week for gigantic conferences from tech giants, including the keynote by Nvidia’s Jensen Huang at Computex on Monday, and Microsoft’s BUILD developer conference, also Monday, Dell’s Dell World show in Las Vegas on Monday and Tuesday, and Google’s “Google I/O” show on Tuesday.

In the case of the Google I/O show, the main event were many revisions to Google’s basic search function.

Google announced it is making the “AI Mode” available to all U.S. users of search, and clearly the intention is to spread it globally. This is where Google has a tab that populates with answers rather than the traditional links.

Along with spreading AI mode, the company is adding in all sorts of “agentic” functions, such as searching your gmail, or providing a lengthy research report, rather than a link, in response to queries. There’s also a hook into commerce as well, as the bot will direct you more readily to online product listings, and use your picture to show you how you might look in apparel selections.

As I wrote in February, before we all became completely distracted by talk of the U.S.’s tariff moves, the software industry has been climbing out of a multiyear slump, with some of the best names experiencing rising rates of sales and faster acquisition of customers.

One of the names I highlighted then, Snowflake, again reported encouraging results on Wednesday evening. And as in February, its shares popped in late trading, rising eight percent, the third quarter in a row the shares have jumped on results.

As in February, the metrics that matter for Snowflake showed momentum.

How bad is the outlook for Tesla? Rather scary, just take it from a bull on the stock.

Adam Jonas of Morgan Stanley, who rates the shares Overweight, with a $410 price target, about nineteen percent above Tuesday’s close of $343.82, on Tuesday writes in his latest missive that China “may have already won the EV battle.”

He thinks Tesla could be clobbered by the latest entrant, the forthcoming SUV from Xiaomi, the scrappy Chinese handset maker. Xiaomi’s first SUV, the “SU7,” was a stunning vehicle, with looks of a Porsche, and the price of a Volkswagen iD3, a compact EV, notes Jonas.

The SU7 sold a hundred and twenty thousand units in its first thirty-six hours, he notes.

This new model, the “YU7,” pictured above, which is expected to be unveiled on Thursday, “will replicate that success,” according to Jonas’s colleague, Andy Meng.