AI news · Saturday, July 18, 2026
Big Tech AI spending surges as infrastructure costs outpace real growth
Companies like Google, Amazon, Microsoft, and Meta have committed to spending over $700 billion this year on data centers, but rising prices for everything from memory chips to power connections are inflating that bill significantly. Morgan Stanley estimates the cost of building one gigawatt of AI capacity has jumped about 20%, meaning a larger portion of that massive capital expenditure is simply covering inflation rather than genuine expansion. Investors are bracing for upcoming earnings reports, where the focus has shifted from revenue to whether these companies can actually show a return on their astronomical infrastructure bets.
Warren Buffett recently flagged this "game" of spending billions on chips and data centers, noting that these outlays are eroding the cash flows that previously funded stock buybacks. While some analysts see a potential for a bullish market, the high cost of this AI buildout remains a major, unresolved risk. This intense focus on AI infrastructure has also trickled down to real estate and local planning, as data centers push into residential areas.
While architects are proposing creative ideas like heating public bathhouses with waste heat from servers, others argue that we need to rethink whether we even require 24/7 AI availability if the cost is community disruption. Meanwhile, companies are often using the 'AI' label to mask standard restructuring, even as the real costs of token usage and internal security infrastructure swallow up projected savings. The human toll of this shift is visible in the job market, where many workers, like a former PayPal consultant who applied to over 6,000 jobs after a layoff, are finding it tougher than ever to land roles despite an increasingly AI-focused corporate environment.
Even in personal lives, the rise of AI is bleeding into domestic agreements, with couples now discussing 'AI infidelity' clauses in prenuptial contracts to define boundaries for emotional relationships with chatbots. In San Francisco, these tensions are hitting the streets as Mayor Daniel Lurie calls for tougher regulations on robotaxis after autonomous vehicles repeatedly stalled during citywide emergencies, causing traffic gridlock. Regulators are also cracking down on more predatory uses of the tech, with Apple and Google removing 'nudify' apps after city officials demanded action against tools used to target women and girls.
Finally, on the research front, China’s Moonshot AI has sparked a new round of debate about global competitiveness with its Kimi K3 model, which some claim (claimed) rivals the frontier performance of top American systems.
The quick hits
- Big Tech plans to spend over $700 billion on AI infrastructure this year, though rising costs for hardware and power mean much of that money is covering inflation rather than new growth.
- The job market remains brutal for many, with some workers reporting thousands of unsuccessful applications despite the AI-driven corporate push for efficiency.
- San Francisco's mayor is pushing for mandatory performance regulations for robotaxis after repeated incidents where autonomous vehicles paralyzed traffic during city emergencies.
- Couples are starting to include 'AI infidelity' clauses in prenuptial agreements to set boundaries for relationships with chatbots, signaling that the definition of cheating is evolving.