This time, they are focused on the massive amounts of information generated in part by the Internet and known collectively as big data.
The creation of Data Elite, a hybrid between a venture fund and an incubator, underscores the obsession among the valley's investors of finding promising companies at the very earliest stages.
Data Elite is backed by some big Silicon Valley names. They include former Facebook executive Chamath Palihapitya's fund, Social+Capital; Palantir founder Joe Lonsdale's fund, Formation8; angel investor Ron Conway; former Amazon executive Anand Rajaraman; and venture firm Andreessen Horowitz.
Last month, regulations changed so private companies can advertise for investors, which is invigorating relatively new vehicles such as angel-investment syndicates.
Declining costs are leading to growing numbers of start-ups; the Bay Area is on track for a record year in terms of the number of companies receiving the early-stage financing known as seed money, according to consultancy CB Insights.
Meanwhile, incubators and accelerator-programs that provide guidance and facilities to start-ups-are multiplying. Almost 1400 are listed at AngelList, a high-profile site for start-up companies and their backers.
It is unclear if Data, or any other group, will achieve better results than most in venture capital, a business where most funded companies fail or stumble along-and just one or two out of ten enjoy meaningful success.
Partly to blame: the one-size-fits-all approach of many funds and incubators, said Data Elite managing director Stamos Venios in an interview.
"Generic advice-it's not enough anymore," he said, emphasizing that big data companies need particular expertise in areas such as storage, transfer, and analytics.
Data Elite aims at a slightly later stage than most incubators and startups. The companies' founders will need to have at least five years' experience in their field, or proven success, such as an existing start-up that sold to a bigger company, Venios said.
"We're more interested in the technical expertise of founders than the ideas," he said.
It plans to accept up to ten companies in its first three-month program, starting in January. That compares to around 50 companies in recent sessions of 500 Startups, and around 30 at Y Combinator, both well-known accelerator programs in Silicon Valley.
It also plans to invest at least $150,000 in each company, including $50,000 for a 6 percent stake and a $100,000 convertible note with terms that vary per company.
By comparison, Y Combinator used to invest $150,000 per company, via a convertible note. Last year it cut the amount to $80,000, citing falling costs to start a company. The funds come from Russian investor Yuri Milner, plus venture firms Andreessen Horowitz, General Catalyst, and Maverick Capital.
Data Elite won't be the first entrant specialized in big data. Accel Partners, the firm noted for its investment in Facebook as well as big data companies such as software-services provider Cloudera, has started two big-data funds of $100 million each.
Another group, Data Collective, has made some early-stage investments in companies such as predictive-analytics service Kaggle and big-data services company Continuuity.
As is the case at some other similar programs, advisors to Data Elite's start-ups will have a financial interest in its companies. Advisors include Ken Rudin, Facebook's head of analytics, and Jeff Mangusson, manager of data science platform architecture at Netflix.
Big data has led to some top-performing initial public offerings lately, including data-analytics businesses Splunk and Tableau.
© Thomson Reuters 2013