What Happens to VCs’ Forgotten Startup Toys?
The gifting season is upon us, with kids and adults alike tearing into their new toys and gadgets while must-have presents from just a few years before sit forgotten. Venture capitalists can probably relate.
In the tech world, we’re wrapping up year two of VCs splurging on shiny new artificial intelligence startups. AI startups to automate coding, customer service and sales are under every tree, and maybe a really good boy or girl will get inference chips in their stocking.
Meanwhile, VC attics are getting stuffed with startups that are now collecting dust, like the proptech and B2B startups investors binged too hard on in 2021, and well as many of the fintech, edtech, e-commerce and telemedicine bets they once lavished with love.
Venture capitalists like to justify their forgotten toy collections by saying that the potential payoffs from wins on the next big thing are so huge that they can’t waste time or money spiffing up older startups that are slowing or never got off the ground to begin with.
“Companies that aren’t working are just getting no attention from investors—it’s a sad reality,” a venture capitalist at a top-tier firm told me last week.
The most notorious shiny toy fan of all might be SoftBank’s Masayoshi Son, who became the poster child of the spray-and-pray approach to backing startups after the 2017 to 2019 investing splurge by his $100 billion Vision Fund. Blowups at WeWork, Katerra, Brandless and many others followed. So did tears (actual tears) and an exodus of SoftBank general partners.
Last year, Son turned his attention away from the Vision Fund and its successor, to focus on the Japanese conglomerate’s most valuable holding, chip designer Arm, which was going public. Vision Fund general partners couldn’t even get him on the phone.
Now, Masa’s money has started sloshing around Silicon Valley again.
A lot of it is going to AI, including a commitment to buy $1.5 billion worth of shares from OpenAI employees. SoftBank has also taken smaller stakes in Glean, Perplexity, Wiz and Databricks recently. On Sunday, Bloomberg went deep on Son’s latest “singular obsession”— to create a challenger to Nvidia. Ever the showman, Son has also said he would spend $100 billion on AI projects in the U.S., even though he doesn’t really have that kind of money.
To his credit, Son has been on the AI train longer than many power players in tech. His biggest win so far has been Arm, which helped improve his reputation after the Vision Fund disasters.
And SoftBank is still sitting on stakes of companies that have aged well, including banking startup Revolut, which has become one of the world’s most valuable private firms, and ByteDance, which could be another big win if it ever resolves TikTok’s political drama and goes public.
But those make up just a small part of SoftBank’s $56.7 billion portfolio of stakes in 382 private companies. Some of its other high-profile bets include Gopuff, the rapid delivery service that keeps laying off workers and burning through cash; Nuro, which recently ditched its ambitions to develop its own autonomous delivery vehicles; and Flexport, the logistics company that founder Ryan Petersen has been trying to lead out of a freight recession.
How SoftBank and other investors deal with their collections of out-of-favor startups—if they pay attention to them at all—will have major implications for dealmaking in the coming year.
More of those once-hot companies will eventually disappear entirely, starting with the ones that have significant capital needs to keep running. This year saw the deaths of doctor-in-a-box startup Forward and high-tech lettuce grower Bowery Farming.
Others may decide to wave the white flag and sell at near distressed prices, as long as VC boards greenlight deals that wipe out common shareholders, and maybe even some early backers. Some investors may try to orchestrate soft landings or make their portfolio companies merge. Many will probably not care and focus on winning the next AI deal.
For now at least, a lot of the AI hype seems to be holding up. OpenAI’s “12 Days of Shipmas” wowed researchers, while startups like Glean and Perplexity have real revenue.
But some AI startups could start looking dusty in a hurry—just ask Inflection, Character.ai or Adept. Don’t be surprised if you find coding agents and customer service bots sitting next to Tickle Me Elmo in the attic come next Christmas.
New From Our Reporters
