This is the web version of Data Sheet, Fortune’s daily newsletter on the top tech news. To get it delivered daily to your in-box, sign up here.
The history of modern computing breaks down neatly into four eras: the mainframe; client-server, better known as personal computers and the software that runs on them; the Internet and its companion, the World Wide Web; and our current age dominated by smartphones and cloud computing.
If Simon Segars, CEO of an obscure but giant semiconductor company called ARM, is right, the coming wave of connected devices powered by artificial intelligence will be a fifth distinct and equally meaningful technology epoch.
A soft-spoken Englishman, Segars has a powerful claim to define the times. ARM, which he joined in 1990 shortly after it was formed as a joint venture between now-defunct computer maker Acorn and Apple, has licensed the technology that’s gone into 130 billion chips. The high-powered, small-size chips soared with the smartphone, and now are in gadgets of all sizes. Segars oversaw the 2016 sale of ARM to Masayoshi Son’s SoftBank and now sits on the parent company’s board. (ARM is based in Cambridge, England, where it started; Segars runs it from Silicon Valley.)
Segars says ARM’s revenue growth has slowed as the smartphone market has matured. Profitability has taken a bigger hit because SoftBank is investing aggressively in the connected device/AI future he foresees. For all his experience in the tech world, Segars affects a sense of wonder over working with Son, a “lateral-thinking person.” Says Segars, in an interview Monday in San Francisco: “He’ll say, ‘Well what about this …?’”
What’s fascinating about this fifth wave of computing progress is that it’s not yet clearly defined. Mainframe and client-server computing are old news. The Internet-plus-browser revolution has largely played out. A mere decade ago the smartphone and cloud services were exciting ideas rather than discrete industries. The “Internet of Things” powered by AI is all the rage, but it’s hard to pin down exactly what products describe it without making long lists of applications.
ARM, having dominated the current technology phase, needs to ensure its chips are in cars, sensors, 5G networks, data centers, and everywhere else. It will be feat for the ages if it succeeds.
My shot. When does a trend become a fad? Well, according to a report from PwC “everyone is talking about blockchain, and no one wants to be left behind.” Globally, 84% of executives surveyed said their companies were actively involved with the technology first popularized via bitcoin.
The story of tonight. As part of a deal to work jointly on self-driving vehicles, Toyota will invest $500 million in Uber at a valuation of about $72 billion. Toyota will use Uber’s autonomous driving technology to power a fleet of minivans, a segment of the market that “is not where Uber has a long-term interest in participating,” Jeff Miller, head of business development for strategic initiatives, tells the New York Times.
History has its eyes on you. A federal court on Monday blocked Texas-based nonprofit Defense Distributed and its founder, Cody Wilson, from posting 3D-printed gun blueprints online. District Court Judge Robert Lasnik, who had imposed a temporary restraining order on Wilson that was about to expire, mandated a preliminary injunction that blocks online distribution while the legal proceedings are ongoing. That could be years, as the case wends its way through the courts.
Tomorrow there’ll be more of us. Last year, the number of women and minority students not including Asians taking advanced placement tests for computer science more than doubled. This year the increase was strong, but not quite as strong. Out of almost 136,000 test takers, 38,195 were women, up 39% from last year, and 7,301 were black or latino, up 44%.
The world was wide enough. Moore’s Law may still be around but it’s getting harder and harder to keep pace. GlobalFoundries, one of the largest semiconductor makers in the world, said on Monday that it was dropping out of the race to build chips at a scale of seven nanometers. That leaves only Samsung, Taiwan Semiconductor, and Intel in the race.
Guns and ships. To kick off its annual customer conference, VMware announced it is acquiring CloudHealth Technology, which helps big companies operate on cloud servers run by multiple companies at the same time. The move comes as some companies don’t want to depend only on Amazon’s AWS or Microsoft’s Azure platform, but want to spread their business across both, for example.
The world turned upside down. The president expanded his war on the press to Google on Tuesday, aiming a pair of early morning tweets at the search giant. Trump accused Google of hiding good news about his administration in search results. A quick check by the Washington Post found the allegation was false.
(Headline reference explainer, like you need one.)
FOOD FOR THOUGHT
On Friday, we offered a piece critical of the stock buyback boom. Unfortunately, I accidentally included the wrong web link. The correct link to the essay by William Lazonick and Ken Jacobson is here. And now for the flip side. Finance professor Alex Edmans at the London Business School wrote a piece last year arguing that the evidence against buybacks is weak and that financial theory suggests the tactic should be beneficial to the overall economy:
The idea that buybacks (or, for that matter, dividends) stifle investment is “partial thinking.” It considers investment only in the company in question and ignores the fact that shareholders can reinvest the cash returned elsewhere. And this represents a second advantage of buybacks over dividends. In a buyback, investors choose whether to sell their shares back. They will likely only do so if they have alternative investment opportunities; no rational investor would sell their stock and just horde the cash. Dividends are paid out to all investors, even those who have no good alternative investment opportunities and who may indeed allow the cash to sit idle. In this way, repurchases are targeted: they return cash to shareholders with the best other uses for it.
Indeed, the fundamental premise implicit in many buyback critiques—that more investment is good and less investment is bad—violates a basic idea in Finance 101. Investment only creates value if its returns are higher than the other projects shareholders could invest in. It takes no skill to simply spend money. Responsible companies don’t invest willy-nilly; they invest when opportunities are good, and show restraint when opportunities are bad. A restriction on repurchases could take us back to the 1970s, where CEOs simply wasted free cash on building empires—RJR Nabisco being a prime example—rather than paying it out to be allocated elsewhere. Repurchases allow shareholders to reallocate funds to young, high-growth firms that are screaming out for a cash injection.
IN CASE YOU MISSED IT
How the U.S. Cities With the Fastest Download Speeds Stack Up Globally By Nicolas Rapp and Brian O’Keefe
BEFORE YOU GO
Remember the story last year about the company that was asking employees if they wanted subcutaneous microchips implanted? Well, they’re still at it. Tech services firm Three Square Market in River Falls, Wis. says about 80 workers have opted for the RFID chip treatment. The chips let the employees buy things at the company cafeteria and log into their computers. “It’s just become such a part of my routine,” says software engineer Sam Bengtson, who uses his chip 10 to 15 times a day. Welcome to the future?