Short Takes #27: The Full Consequences
Jamelle Bouie | So Much Money Building AI Infrastructure | Ticketing Driverless Cars
…
We cannot predict the full consequences of what we do, and so we should choose carefully and deliberately as we navigate the world. We should be modest in our ambitions, aware of our own fallibility and mindful of the way things can go wrong.
| Jamelle Bouie
…
It’s clear to me that many of our so-called leaders — political and economic —do not operate by Bouie’s admonition.
I’d really like readers to sign up for a paid annual subscription, so for the present time, I have dropped the annual subscription to $30. Note that I’ve also raised the monthly subscription to $10 per month from $6 per month. Give annual a try. The biggest value is years of posts behind the paywall, and of course, seeing new posts in their entirety.
So Much Money Building AI Infrastructure
Karen Weise adds up the hyperspending of the hyperscalers in A.I. Spending Sets a Record, With No End in Sight
In the first three months of the year, the four companies [Amazon, Google, Microsoft and Meta] reported in their financial results, they plowed a total of $130.65 billion into capital expenditures, largely spending on data centers that power A.I. That figure — which was another record — was more than three times what the Manhattan Project cost to develop nuclear bombs and 71 percent higher than what the tech giants spent in the same quarter a year earlier.
All of the companies said they would be spending even more, totaling roughly $700 billion this year. Meta, for one, raised its spending forecast for 2026 to between $125 billion and $145 billion, up from its previous prediction of $115 billion to $135 billion. Google also boosted its projection, to at least $180 billion, and said its spending would be “significantly” higher next year.
Om Malik adds his observations about how much more is off the balance sheets:
What is also true is that funding is increasingly off the balance sheet, that supplier relationships are being prepaid, that lease commitments are being deferred for as long as accounting rules allow, and that a meaningful portion of the AI revenue and AI investment gains are flowing in a circle through the same small set of AI labs.
The platform shift is real. AI engineering is real. So is the financial engineering.
Let the good times roll.
Our economy is structured — by policy, not just market forces — so that these companies have this much money to bet on the AI lottery. Meanwhile, the federal debt just surpassed 2025 GDP, in part because these companies (and their billionaire owners) don’t pay very much in taxes.
Ticketing Driverless Cars
It had to happen, I guess. Orlando Mayorquín reports:
Police officers in California will soon be able to issue traffic tickets to driverless cars like Waymo robotaxis and require their manufacturers to move them out of the way during emergencies.
The state’s Department of Motor Vehicles adopted the new rules for autonomous vehicles this week, in accordance with a 2024 law that imposed more regulation on the technology.
The rules, which go into effect July 1, are designed to address some of the challenges that have vexed local governments and residents in places where driverless carmakers, like Waymo, have expanded their fleets.
What about driverless police cars issuing tickets?
AI Layoffs Starting To Hit India, Too
Steven Lee Myers, Paul Mozur, and Saumya Khandelwal looked at AI impacts in India:
For a quarter century, India has made itself the world’s back office, providing an educated, English-speaking work force to do tasks more cheaply than in the United States or Europe. The industry today employs more than six million people and is worth nearly $300 billion, more than 7 percent of the country’s gross domestic product.
Now, A.I. threatens to do to India what its outsourcing model did to the rest of the world: replace hundreds of thousands of office workers.
India grew a massive workforce as a low-cost alternative to office workers in the West. Now, we are seeing the impact on new grads, with the same calls to ‘upskill’ as we are seeing here:
The tremors are already being felt. Tata Consultancy Services, one of India’s largest employers, has shrunk its work force to 580,000, a decline of more than 20,000 from a peak in 2022, when it hired 100,000 new workers in one year alone.
Its main rival, Infosys, has also slowed hiring, while dozens of smaller start-ups laid off workers across the country in 2025, according to Inc42, a digital economy news outlet in India.
Graduates of the country’s universities and technical colleges are finding fewer openings, forcing them to scramble to “upskill,” an increasingly popular term in the context of learning the A.I. technology that is reshaping the industry.
Maybe these giant service companies will start building their own data centers to lowball the cost of running AI models? At some point, it comes down to the cost of the chips and electricity, and with the growing concerns about data centers in the US, it might be more politically attractive to locate them on the other side of the planet.
