Yasmin Razavi holds an unusual position: She is the only outside investor on Anthropic's board of directors.
The general partner at Spark Capital invested in Anthropic relatively late by traditional venture capital standards, leading the company's Series C financing in May 2023, its third round and two years after siblings Dario and Daniela Amodei left OpenAI to found it amid concerns over AI safety.
Razavi is part of a new generation of investors transforming venture capital. As companies stay private longer, the venture-capital world is experiencing a historic reordering -- and a new way of investing.
The recent initial public offering of SpaceX, along with the expected debuts of Anthropic, OpenAI and others, place a group of believers atop the investing world. Some made career-defining bets from inside established firms. Others are newer venture funds that built their brands around backing the AI boom.
Some of the new powers, including Spark, Gigafund and Greenoaks, are virtually unknown outside of Silicon Valley and Wall Street. Even there, they were often overshadowed until scoring massive windfalls from recent AI and other investments.
What many of the new winners have in common: they've embraced a new model of investing. Today, high-growth companies remain private for much longer -- often well over a decade -- locking the public out of key, wealth-creation phases of their growth. Many of these investors are investing patiently in private companies and writing check after check, helping the companies scale and establish an edge over rivals.
Paying up for stakes in the hottest startups is a riskier strategy than venture capital's traditional approach, one that can lead to disappointment if companies fail to live up to lofty expectations. Today's venture capitalists often don't have the influence to guide or shape companies the way they once did because of heightened competition. Instead, many are willing to invest in companies later on, sometimes years after they were started, betting that there's more growth ahead.
It's a radical shift from the old days of venture capital, when firms like Sequoia Capital and Kleiner Perkins wrote smaller checks, worked side by side with startups and helped take them public, typically within five years. It was a model designed to produce strong returns because investors acquired large stakes at low prices.
Today's venture-capital investments often resemble IPOs, says Michael Moritz, who co-ran Sequoia Capital for 16 years, with investors buying stakes without seeking influence.
"The amount of money that's available is astronomical compared to yesteryear," says Moritz, who led Sequoia's investments in Google, PayPal and Yahoo. "The business has changed entirely."
The size of the venture capital industry has expanded significantly. U.S. startups raised $321 billion last year, compared to $84 billion in 2016, per PitchBook. Average late-stage venture rounds have swelled from $65 million for a Series D investment, to over $1 billion so far this year.
For Razavi, a single decision changed everything. In the fall of 2022, she learned that a research lab called Anthropic was planning to commercialize its AI efforts. The then-33-year-old investor walked over to Anthropic's headquarters in the San Francisco neighborhood of Jackson Square to hear Anthropic co-founder Dario Amodei and a new hire, Neerav Kingsland describe Anthropic's plans and how AI could help people flourish.
At the time, OpenAI was far ahead. But Razavi, who grew up in Tehran, and moved to Canada when she was a teenager, believed that more than one AI model developer could succeed -- and Anthropic would be one of them.
Investors across Silicon Valley passed on Anthropic, uninterested in paying a $4 billion valuation for a company with no product or revenue. Razavi and her colleagues at Spark, a venture firm founded over two decades ago, debated the investment.
Razavi, who had previously worked as an analyst at McKinsey & Co., wrote an investment memo that examined each layer of AI, from chips to models and applications.
She wrote a second memo, working through potential objections, point by point. With her team on board, in February she told Amodei that Spark wanted to lead the round.
In early 2023, Spark made its $75 million investment, the largest check it had ever written to a company at the time. Today, Spark's stake is worth about $7 billion at Anthropic's latest $965 billion valuation, people familiar with the matter said.
Razavi's home run has elevated her and Spark's reputation. In June, it announced it had raised $3.3 billion for two new funds, the most money it has ever raised.
The risk of paying such high prices is smaller returns. Howard Morgan, a venture capitalist for nearly 50 years, sees more capital chasing fewer deals today. Morgan began investing in the late 1970s before co-founding First Round Capital in 2004, which backed companies including Uber and Roblox. At the time, early-stage startups were typically valued at around $1.5 million, he said.
"There was much less money in venture capital in those days," Morgan said. "It's harder now. You have to be more selective and more selective about what you're going to pay for something."
As venture capital has evolved, one investing truism remains: Home runs usually result from contrarian bets, rather than from diversifying a portfolio or chasing a hot sector. SpaceX, Anthropic and OpenAI all faced serious skepticism, when many of the investors wrote their first big checks.
Luke Nosek made one of the earliest and riskiest such bets. Nosek, 51, studied computer engineering at the University of Illinois at Urbana-Champaign, where he became immersed in technology.
He met Peter Thiel and they helped start a company that merged with an Elon Musk-operated business to become PayPal. Nosek and Thiel went on to launch venture-capital firm Founders Fund in 2005.
Nosek wore flannel pajamas around the office, munched on fresh blueberries from their plastic clamshell containers, and built a soundproof, blacked-out box to improve his sleep. He stayed out of the spotlight as Thiel and others assumed public roles.
Nosek's early investment in DeepMind, an AI company focused on achieving artificial general intelligence, paid off when it eventually was purchased by Google.
He also became convinced lifespans could be extended. One day, he approached a new employee, asking, "How old do you want to live to?"
Ninety-five or so, she said.
"Why not 150?" Nosek asked.
By 2008, Nosek was the firm's biggest believer in SpaceX, say people familiar with the matter. The company had been around for six years but hadn't demonstrated it could launch a rocket inexpensively enough to transform the industry.
Nosek led a $20 million investment -- about 10% of the firm's second fund -- and joined SpaceX's board of directors. At least one Founders Fund limited partner refused to invest in the firm's next fund, unhappy it was wagering on a rocket company. Over time, Nosek and Founders Fund invested several hundred million dollars in SpaceX.
"It was a huge, huge bet for us," says Ken Howery, a Founders Fund co-founder who also worked on the investment.
Nosek pushed others to invest in the company, visiting the home of Cyan and Scott Banister, two venture capital investors and longtime friends.
"Guys, I've never believed in anything more," he told them in 2008, when SpaceX was worth about $400 million, Cyan Banister recalls. "If you have any liquidity to put into this, now is the time."
Nosek wanted to make even larger investments in Musk's companies and in 2017, he left Founders Fund to co-found Gigafund and eventually moved to Austin, Texas. His new firm backed Musk's Neuralink and The Boring Company, and invested more than $1 billion in SpaceX. That stake has ballooned in value to tens of billions of dollars, based on the company's current market value.
Along with bushels of cash, the new investing powers will receive something else from these deals: new scrutiny. Some of the firms scored their windfalls due to their closeness to founders like Musk, raising questions about whether they will have the same access to future entrepreneurs.
There's also concern about how the firms will monetize their holdings, turning their paper profits into actual gains, said David Mann, chief executive of the Mannsion Group, which invests in late-stage private companies.
Some of the big winners cemented their status through bold recent investments. Greenoaks, led by Neil Mehta, first invested in Anthropic earlier this year. The firm co-led the funding round that valued Anthropic at $965 billion just a few months later, and owns a multibillion-dollar stake.
In recent years, the firm has backed some of the industry's most closely watched AI startups, winning competitive deals once reserved for Silicon Valley's establishment. Those include Safe Superintelligence, a startup founded by OpenAI co-founder Ilya Sutskever, and customer-service AI startup Sierra, founded by former Salesforce co-CEO Bret Taylor. The investments have helped elevate Mehta's profile as one of the leading investors of the AI boom.
Mehta, 42, previously worked at the hedge fund D.E. Shaw in New York and Hong Kong before co-founding Greenoaks in 2012. The firm first built its reputation on investments in companies such as Palantir, Stripe and Coupang, the South Korean e-commerce giant, where it invested repeatedly over more than a decade.
Mehta favors making concentrated bets -- and doubling down on his winners.
He grew up in Silicon Valley with two engineer parents. For his 14th birthday, he asked for a trip to Omaha, Neb., to attend Warren Buffett's annual shareholder meeting. He also started a social-networking company while he was in high school during the dot-com boom. By the time he graduated in 2002, the startup had failed and the bubble had burst, a formative experience that fostered his interest in entrepreneurship.
(MORE TO FOLLOW) Dow Jones Newswires
July 17, 2026 21:00 ET (01:00 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments