Colossal SoftBank fund could shake up tech world


Japan-based SoftBank is sending tremors through the tech world with a massive new venture capital fund for investing in startups that's expected to dominate the industry so thoroughly it's playfully referred to as a "gorilla." The Vision Fund's $100-billion coffers nearly equals the total amount pumped into venture capital-backed companies last year, according to market intelligence firm CB Insights, and some say it may be a game-changer for Silicon Valley.

"SoftBank shows a remarkable amount of bravery, confidence and optimism to look to apply this much money in technology," said Bill Maris, who started Google Ventures nearly a decade ago and runs his own California-based investment firm Section 32.

"I can't say it's a wrong bet, if you think the trends in tech will continue in the future. I would be much more worried if SoftBank was saying tech is dead."

Last year, VC-backed firms received $100.8 billion across 8,372 deals around the world, according to CB Insights data. The huge amount of cash could accelerate the trend where fast-growing startups remain private - without the scrutiny and transparency of a stock market debut. Some investors worry that the Vision Fund will buy into startups at high prices, overinflating the market, while crowding out other investors and prolonging the time it takes for young companies to go public. SoftBank has outlined plans to focus on late-stage investments when startups are more established, and on investments of at least $100 million. The SoftBank fund is widely expected to pump some $10 billion into ride-sharing giant Uber, which has a whopping valuation near $70 billion. Such a deal would boost the profile of the Japanese group in Silicon Valley. Maris predicted the venture capital market would adapt to the Vision Fund, and in the end there would be more money available for entrepreneurs.

"It is looking for mega-investments," he said. Targets for the Vision Fund were expected to include e-commerce, ride-sharing, robotics and machine learning.