Both Sides Now: Eye-Popping Artificial Intelligence Financials from Producers as Well as Suppliers, But…

Despite crashing public approval and incessant talk about a bubble, AI business results, and not only those from the gunrunners, are more stunning than ever.

In “Ads coming to ChatGPT for some US users as OpenAI seeks to generate new revenue” (Michael Sinkewicz, Fox Business, January 17^th), we saw something consistent with other online products, as “OpenAI plans to test advertisements in coming weeks for free and lower-tier subscribers ahead of (its) anticipated IPO.”  A company spokesperson said the pitches will appear “at the bottom of answers in ChatGPT when there’s a relevant sponsored product or service based on your current conversation,” they “will not be shown in accounts where the user is under the age of 18,” they “cannot appear near sensitive or regulated topics like health, mental health or politics,” and the company will “never sell your data to advertisers.”  Business as usual.

Riches for AI-making are relatively new, so there can still be headlines like “The A.I. Boom’s Uncertain Payoff” (Andrew Ross Sorkin et al., The New York Times, January 29^th), as “investors remain antsy about seeing results.”  However, “investors appeared worried” about the “payoff” from Microsoft spending “$37.5 billion in capital expenditures in its most recent quarter,” but less so about Meta planning on disbursing “$115 billion to $135 billion on capex this year,” and even Tesla, outside “making electric vehicles,” $20 billion.  The large questions still in the air are “how will investors feel if companies’ margins remain depressed for a long time amid all of their spending?” and “how will they ultimately measure success?”  Indeed, per “Microsoft Continues to Spend Big on A.I. While Profit Jumps 60%” (Natallie Rocha, The New York Times, January 28^th), despite the headline outcome along with a 17% quarterly revenue gain, its stock price later that day was “down more than 5 percent in after-hours trading.”

Next, we saw as “Nvidia’s Quarterly Profit Hits $43 Billion on Strong A.I. Chip Sales” (Tripp Mickle, The New York Times, February 25^th).  That was also $120 billion for the year.  The leading AI supplier - it “controls about 90 percent of the market for the cutting-edge semiconductors that power A.I. projects” - joined “only a handful of companies, including Alphabet, Microsoft and Apple” in profiting an annual $100 billion.  Yet, per the previous paragraph, its customers’ “spending is starting to unsettle Wall Street, and Nvidia’s share price has been relatively flat in recent months,” even if “demand for Nvidia’s chips was still growing at an astonishing rate.” 

Moving to this month and a new Nvidia competitor, “A.I. Chip Maker Soars 68% in Market Debut, as Tech I.P.O.s Ramp Up” (Natallie Rocha, also in the New York Times, May 14^th).  The company, Cerebras, which sold its first chips in 2019, made itself the largest IPO of the year so far, and closed that day with market capitalization of $67 billion.  It was profitable last year with $238 million net.  There will be more - which will survive?

We got another quarterly update, as “Nvidia’s Profit Hits $58.3 Billion as A.I. Boom Gathers More Steam” (Tripp Mickle, also in the Times, May 20^th).  That was over 35% higher, and as “Nvidia’s biggest problem appears to be meeting demand from its spendthrift tech industry customers,” there is plenty of room for the likes of Cerebras.  The stock market shrugged, though, as that day, Nvidia’s “share price fell 1 percent in after-hours trading, giving up most of its gains from earlier in the day.”

Since then, something has changed.  In “Why Memory Chips Are Dominating the A.I. Rally,” Andrew Ross Sorkin et al., still in the Times, reported that “Micron, Samsung and SK Hynix are now trillion-dollar companies,” all three of which among the firms that “dominate the memory chip supply chain.”  Those who invested in Micron or SK Hynix last year have been well rewarded, as their share values have increased threefold and tenfold thus far in 2026.  All of that meant “global stocks climbed to fresh records for a sixth straight session and US futures advanced as investors piled into tech shares” (Morning Briefing, Bloomberg, May 27^th).

We are now seeing companies creating AI products doing well, in some ways better than their suppliers.  That is a huge improvement.  Yet AI’s financial side, even if bilateral now, seems more like a protected cove than the state of the technology in general.  If American data center construction screeches to a halt or near-halt, and public dissatisfaction becomes more important in other ways, what will happen to AI profits, revenues, and market capitalization?  We will see, but it won’t be good.  AI may be soaring, but rough air may, as always, be just ahead.  Place your bets as you will, but don’t count on anything - we really don’t know artificial intelligence at all.

Artificial Intelligence Regulation - Activity on the Way to Something to Be Determined

How will AI be constrained?  That is the same open question it has been since the start of AI.  Yet things have been happening toward that.  What?

First, a late 2025 Donald Trump view: “Chasing an Economic Boom, White House Dismisses Risks of A.I.” (Tony Romm and Colby Smith, The New York Times, December 24^th).  At that point, the technology got “the administration’s unqualified support,” as that year “the president and his top aides have fully embraced A.I. and showered its leading corporate backers with money and regulatory” help.  Other major players, though, looking at possible lower product demand, job losses, or at least greatly decreased hiring, were not so sure.

In response to a widespread concern soon before then, a “Deepfake porn crackdown passes in Senate to allow people to sue” (Alex Miller, Fox News, January 13^th).  “The Senate quietly passed legislation… that would create stiffer penalties for explicit AI-manipulated images, known as deepfakes.”  The bill “is designed to beef up federal penalties against the creation, distribution or solicitation of “non-consensual digital forgeries,”” and is “geared to act as a companion to a previously passed bill targeting revenge porn.”  Even if passed into law, controversy on this topic remains, centering around the ease of producing such material, making it unlikely that here is the last word.

Do we want all AI-related laws to come from Washington?  That could be a presidential goal, as “White House Unveils A.I. Policy Aimed at Blocking State Laws” (Cecilia Kang, The New York Times, March 20^th).  “Dozens of states have passed laws in recent months to regulate A.I., which has created concerns about the technology’s potential to steal jobs, push up energy prices and threaten national security.  But President Trump has made clear U.S. companies should have mostly free rein in a global race to dominate the technology.”  That view is supported by the industry, as “Meta, OpenAI, Google and other A.I. giants have argued that a patchwork of state laws could slow down their progress” and “have repeatedly pointed to regulation as the biggest hindrance to the nation’s success in leading the world in A.I.”  Yet “the White House also called for provisions that protected children, including stronger parental controls and privacy protections.”

Appropriately, we also saw as “IBM CEO Arvind Krishna warns Washington must find ‘Goldilocks’ middle ground on AI regulation” (Kristen Altus, Fox Business, May 5^th).  Krishna said that, as “they were always going to regulate the use case of (AI),” “there is always a level of government oversight.”  However, “if it turns into a bloated bureaucracy, that would not be so good for us to win the AI race,” and “this is always the balance between innovation and safety.”  He thought it best with regulators “going to do their judgment quite quickly within a few days or a few weeks,” as that “serves everybody very well.”

With regulation in flux, it also makes sense that “Silicon Valley’s A.I. Lobbying Reaches a Fever Pitch” (Cecilia Kang, The New York Times, May 13^th).  As a component of “OpenAI’s increasingly aggressive push to sway A.I. policy,” the firm is opening “its first lobbying office in Washington,” which will be “part lab, part showroom.”  It joins Anthropic, which “opened its first office in Washington in April, as it battled with the Pentagon over the use of its technology.”  Already, “a quarter of the 13,000 federal lobbyists in Washington are involved in A.I. issues, up from 11 percent in 2023.”  The companies are hardly unified in their objectives, as while “OpenAI, Meta and Google have pushed for little or no regulation,” “Anthropic and others have supported new laws, pointing to the technology’s potential dangers.”  That same day, though, “OpenAI backs creation of global AI governance body led by the U.S. that would include China as a member” (Michael Sinkewicz, Fox Business).  Per an OpenAI vice president, “the proposed organization could resemble the International Atomic Energy Agency, which includes China and sets global safety standards for nuclear energy development.” 

A fine idea, and whether you consider that regulation or a device to keep regulation at bay is up to you.  So, control of artificial intelligence is still sketchy and makeshift, but, in the pieces above, we may be able to see more clearly what its future could resemble.  As always, stay tuned.

Different Views on Artificial Intelligence’s Effect on Getting Jobs, All From the Past Five Weeks

Within AI, the press subjects vary.  Lately, I’ve seen nothing about problems with ever-larger data feeds, and while a general backlash reaction to AI is gaining strength and attention, little specific has been out lately about its previous hot topic, data centers.  Even rundowns on specific AI achievements and its huge philosophical issues have been relatively scant.

The oldest here was “Time to ditch AI anxiety – experts say there’s a lot less to fear than we think” (Simon Constable, Fox Business, April 11^th).  As you will see, not all agreed with this thesis.  Constable started with acknowledging the amount of change, and concomitant tension, AI has wrought, and with “data from Challenger, Gray and Christmas” saying that “AI was directly involved in firing 54,000 people during 2025” (though that is hardly a huge number), moved to saying that “last year, approximately 280,000 new jobs in Gen-AI were created for people, according to Electro IQ Job Creation Stats,” that the “28.3% of the working-age population” that “used generative artificial intelligence” had become more productive, and that, ultimately, “creativity comes to life because people working with AI need to do the thinking.”  A short but clear argument.

As a solution to positions disappearing, “The AI revolution threatens office jobs, but revives demand for skilled trades” (James Altmire and Riley Burr, Fox News, April 12^th).  This is clearly happening, especially among younger career-choosers, who are also rightfully attracted to earlier paychecks, remarkably high compensation, modern-day work conditions, and, with all of that, higher prestige.  The subtitle “trade careers requiring manual dexterity, problem-solving and emotional intelligence remain beyond AI’s reach” describes the other side.  “Even if AI is able to automate some of the more routine tasks in the workplace, tradespeople are further insulated from AI-driven job displacements because of the unique need for human touch in these roles… AI algorithms may help diagnose issues, but human experts must then step in… with careful judgment, manual dexterity and complex problem-solving.”  There are reasons why trade positions have long had apprenticeships, which tells us that abstract learning is not sufficient preparation for them.  Nor will it be.

I didn’t like seeing the headline “Silicon Valley Is Bracing For a Permanent Underclass” on my May 3^rd New York Times, and that feeling didn’t improve as I read the story.  What was author Jasmine Sun trying to say?  “Most people I know in the artificial intelligence industry think the median person is screwed” – is that the median West Coast programmer, stuck in a career field endangered long before AI came to prominence?  “Anthropic chief executive Dario Amodei” made self-serving “pronouncements about a white-collar blood bath” – so what?  “You feel it in the fretting of recent college graduates who apply to hundreds of jobs without landing a single interview” – so what else is new?  From a quoted “23-year-old start-up founder and Stanford dropout… There’s only a matter of time before GPT-7 comes out and eats all software and you can no longer build a software company,” and another future tool “can perform all physical labor as well” – the first is called creative destruction, and the second can most charitably be called wildly unlikely.  It may well be that Silicon Valley will become less prominent – change does happen, and it is not all favorable to everyone – but the country in general will not succumb to a failure of “society’s ability to cushion A.I.'s disruption.”  We are long past any need for reality-unsupported, obsolete-anyway screeds like this.

Perhaps the New York Times editor had a sense of humor, when, on the same date as above, there also appeared Ezra Klein’s “Why the A.I. Job Apocalypse (Probably) Won’t Happen.”  Despite unexamined-sounding predictions such as Microsoft AI’s CEO saying “that most white-collar work will be fully automated by an A.I. within the next 12 to 18 months,” “the microdata isn’t matching the anecdata:  The unemployment rate was 4.3 percent in March 2026; in March of 2020, it was 4.4 percent.  Average hourly earnings are stable.  Claude Code is a marvel, yet demand for software engineers is booming.”  Additionally, “A.I. will make knowledge plentiful,” “the more automation there is, the more people value a human’s touch,” “computers can do much that humans once did, but they didn’t put humans out of work,” as “the ability to do more made people realize there was more to do,” and, personally, “the better my A.I. has gotten the more I’ve wanted from the human beings around me – and from myself.”  A future of mass idleness indeed does not seem reasonable, so this piece is much better.

More additive AI effects are described in “How AI exposure is reshaping jobs in creative fields” (Eric Revell, Fox Business, May 4^th).  The technology is integrating here, instead of taking over.  While some roles, such as dancing and acting, don’t work as much with AI, for music directors and composers “a substantial portion of their tasks involve composition or production that AI tools may draft or modify.”  A recent study “found little evidence that generative AI has broadly reduced artists’ earnings.”  There, though, could be jobs lost by people not able to keep up with those using AI for assistance.

So, should we be concerned if “Congress Is Doing Little to Prepare for Potential A.I. Job Losses” (Ben Casselman and Tony Romm, The New York Times, May 5^th)?  “The federal safety net isn’t ready for such a shock,” presuming that displacement would be sudden and severe, but would it be?  If “unemployment insurance and other safety net programs are long overdue for an overhaul,” it could be right to do that, but the authors’ comparison with when “over just a few weeks in the spring of 2020, more than 20 million Americans lost their jobs” is weak.  Accordingly, we can go from there.  Although evidence for expecting massive AI-caused cuts is missing, the chance is more than zero, and it may be time to prepare.  Let us consider that, while still not giving in to unjustified reactions in any artificial intelligence area. 

April’s Jobs Report No Success After Employment Gain – Latent Demand Little Dropped per AJSN of 16.5 Million

One of many problems of emphasizing a single result in the monthly Bureau of Labor Statistics Employment Situation Summary is that it may not be representative of the outcomes as a group.  April’s edition, which came out this morning, was a prime example.

The number of net new nonfarm payroll positions almost doubled the published estimate I saw, reaching 115,000 instead of 67,000.  From there, although seasonally unadjusted joblessness fell from 4.3% to 4.0%, the adjusted version did not worsen at 4.3%, unadjusted unemployment was seasonally off 465,000 to 6.77 million, and long-term unemployed held at 1.8 million, it was a negative month.  The count of adjusted jobless rose 200,000 to 7.4 million.  The two figures showing how common it was for Americans to be working or just short of that, the employment-population ratio and the labor force participation rate, each dropped 0.1% to reach 59.1% and 61.8%.  The number of people working part-time for economic reasons, or holding short-hours positions while seeking thus far unsuccessfully full-time ones, soared 400,000 to 4.9 million.  Average hourly private nonfarm payroll earnings again gained less than inflation, 3 cents per hour, to $37.41.  Despite April being a seasonally stronger month than March, the unadjusted count of employed moved up only 17,000, to 162,781,000. 

The American Job Shortage Number, the seasonally unadjusted metric showing how many new and unadvertised positions could be suddenly filled if all knew they would be easy to get, lost about one third of a million, less than seasonal expectations, to the following:

 

The inputs, except for classic unemployment which deducted over 500,000, offset one-third of the latter’s subtraction, with most coming from increased numbers of those not looking for the previous year, those wanting to work but unavailable for now, those not wanting a job, and those in school or training.  The share of the AJSN from this official joblessness shrank from 39.3% to 37.0%. 

We did not make progress in April.  As well as in the results above, the year-over-year comparison showed the AJSN up 423,000, with more from nearly all of the secondary categories above as well as 169,000 from unemployment, which was almost 200,000 lower in April 2025.  More people are falling behind in pay.  More people – 650,000 additional over the month – have left the labor force, with 422,000 more claiming no work interest.  The number of Americans going part-time for economic reasons, an indicator of hardship while working that is not reflected in other statistics, is again pushing post-pandemic highs.  It is good to get more jobs, but this time that wasn’t enough.  The turtle stayed right where he was.