AI Boom Concentrates Software Control and Costs

Berkshire Hathaway has invested more than $900 million in Bending Spoons as the Italian company buys software businesses whose valuations have fallen amid the artificial intelligence boom. Berkshire disclosed another $330 million investment on Wednesday. Bending Spoons’ acquisitions have brought layoffs, departing staff and higher subscription prices at several companies it owns.

Who Controls the Software

Bending Spoons launched in 2013 and is based in Milan. At its Nasdaq debut in July, the company owned older internet brands including AOL, Vimeo, Eventbrite and Evernote. Friends Luca Ferrari, its CEO, Matteo Danieli and Francesco Patarnello founded the company. After their mobile app Evertale failed, they used the $40,000 they had left to build an operating company rather than “products nobody wanted,” according to the IPO prospectus.

The company announced in August that it would acquire workflow software provider Airtable for about $1.3 billion, nearly 90% below Airtable’s late-2021 peak valuation. On Monday, it said a “large portion of the pre-acquisition team” would leave with “substantial separation packages.” Bending Spoons has also laid off staff at Evernote, Vimeo, WeTransfer and other portfolio companies.

Last month, the company agreed to buy collaboration software vendor Miro for $1.36 billion. Miro had once been valued at $17.5 billion. Bending Spoons says it aims to generate profits from acquired businesses rather than pursue an eventual exit, and it makes hands-on changes to technology, operations and personnel.

Customers and Workers Face the Changes

Bending Spoons raised Evernote’s annual personal subscription price from $70 to $130 in April 2023, months after it bought the company. It later charged $100 for its starter plan and $250 for Evernote Advanced, compared with $170 for the former pro plan.

Mizuho analyst Lloyd Walmsley said his initial reaction was, “Oh man, I guess I have to switch to OneNote,” but he stayed with Evernote. The product provided value, he said, and switching costs were high. Walmsley said Bending Spoons makes changes at a depth and scale that traditional private equity firms don’t typically pursue.

Former Gainsight CEO Nick Mehta said “Bending Spooned” had become a Silicon Valley verb for companies acquired for pennies on the dollar. “Unfortunately, a lot of my friends in that game are stuck,” he said. “Investors want out but there are no buyers. Anything pre-ChatGPT is legacy now.” Mehta said private equity firms were limited because “they got their nose bloodied.” Thoma Bravo and Vista Equity Partners didn’t respond to interview requests. Bending Spoons said it hadn’t been outbid for a target in “several years.”

The AI Spending Bet

AI agents powered by large language models are pressuring software-as-a-service businesses in a shift referred to as the “SaaSpocalypse.” Shares of Intuit, HubSpot and Adobe were each down at least 30% that year. Private SaaS companies that had attracted venture capital struggled to raise money, while public markets showed little appetite for initial public offerings. Anthropic and OpenAI were moving toward trillion-dollar valuations.

A separate Fortune report said Amazon, Alphabet, Microsoft and Meta were projected to spend more than $720 billion in capital expenditures that year, most of it on AI infrastructure. Alphabet President and Chief Investment Officer Ruth Porat said Alphabet intended to spend “significantly more” in 2027, with projected capital expenditures of $195 billion to $205 billion, compared with $91.4 billion in 2025.

Private credit funds including Blue Owl, Apollo and KKR increasingly financed AI development. By some estimates, outstanding private loans to AI-related companies had climbed above $200 billion. Alphabet reported negative free cash flow in the second quarter for the first time since its 2004 initial public offering. At the end of June, long-term debt stood at about $98 billion, more than four times the year-earlier amount.

Jefferies’ global head of equity strategy, Chris Wood, called massive U.S. capital destruction the most likely long-term outcome of the AI boom. He argued that cheaper Chinese open-source models could undermine the economics of U.S. investment. Porat rejected that argument: “It's illogical to say we're in a capacity-constrained position because there's such strong demand for the insight in the ability to improve business and government, and yet you can't drive a return on it.”