The AI revolution is bringing back what made Silicon Valley great in the first place

The opportunity to “put a dent in the universe” has sparked a return of the builder culture

A year ago, the residents of the spectacular $68 million mansion at 1868 Floribunda Ave. in Hillsborough, Calif., began holding regular get-togethers for Silicon Valley’s elites. With so much bad mojo in the air — crypto scandals, destructive social apps, egotistical tech icons, emigration to Texas and Miami — the residents reckoned the palatial digs and poolside parties with other talented people would remind locals what the area’s tech culture still had to offer. “It gave the impression that it was just a party house,” says Rocky Yu, who began living there in late 2022.

Now there’s a new kind of party going on — one that doesn’t require any promotion. With Silicon Valley in the throes of what seems to be an earthshaking technology revolution, Yu and his colleagues decided in early 2023 to focus house activities on generative AI. Now, hundreds of people come to a stream of presentations, dinners, and mixers each week. “A year ago, there were maybe 100 people in the world who were working closely enough with foundation models to really understand the potential,” says Yu, founder of what is now called AGI House. “When we held a hackathon in March, we had a waiting list of 300.”

Thanks to the generative AI boom, Silicon Valley seems to have gotten its mojo back. But this isn’t a typical Gold Rush. We believe it’s also a cultural transformation that is upending the blueprint for company-building. For years, much of the focus was on following a proven path that included finding the right team and finding the product-market fit and business model to win over investors. Now, it’s once again all about using powerful technology to create a great product, and finding a way to turn it into a great business.

“It feels like Silicon Valley went from being almost like a casino, where people could come and try to make a quick buck, to what made Silicon Valley so special in the first place: an ecosystem of people who create something that matters, who want to change the world,” says NEA partner Ann Bordetsky. “If you’re a technology-first founder, this is your time.”

Hackathons and coding in cafes are back

This new mood is an ecosystem-wide phenomenon. It’s infecting everyone from 20-something engineers to mid-career project managers to 50-something AI luminaries. “You’re seeing new companies developing everything from foundation models to infrastructure to applications, and for every industry from entertainment to healthcare,” says Sally Shin, co-founder of Raive, a stealth company developing a foundation model for images and video, focused on IP attribution. Shin, who was CNBC’s San Francisco bureau chief before catching the startup bug, says she gets invited two or three times a week and hosts a monthly poker game for female founders.

“During the pandemic, everyone said the Bay Area was dead — and it felt dead,” says NEA partner Vanessa Larco. “You’d go to cafes and people would be writing emails and working on pitch decks. Now they’re coding.”

The new cultural vibe was evident at a recent hackathon hosted by NEA at AGI House. The topic was image generation, and while many of the attendees were there for purely capitalist reasons — particularly non-tech executives on hand to figure out the implications of generative AI — “I’d say 50% of the attendees were just there because they love building things. It’s their hobby, and their craft,” says Bordetsky. As judges, she and two other NEA partners singled out for recognition a group of strangers who’d created a song-to-music video service that pulled together images based on the lyrics and mood of a song. “Is there a business there? Probably not, but that’s not the point,” she says.

Migration back to the Bay…and to the office

Rather, the point is that entrepreneurs who are serious about generative AI once again want to be in the Bay Area. “Two of our co-founders are European, so we did think about other locations, but in the end the choice was really easy,” says Shin. Not only are noted entrepreneurs like Brex co-founder Henrique Dubugras returning after brief sojourns to trendy locales, but a high percentage of the most prominent technologists and thinkers never left. And the more people come back, the more essential it is for Silicon Valley artificial intelligence startups to be situated locally for recruiting purposes.

The transformation is in many ways a restoration of Silicon Valley past — and not just because of the flowering of new silicon companies intent on creating chips optimized for AI. The energy and openness reminds Silicon Valley veterans of epic periods when brilliant people shared their knowledge and provided assistance out of sheer passion for the potential of technology. As with the Homebrew Computer Club where Steve Jobs and Steve Wozniak honed the ideas that launched Apple, “people feel like they’re involved in something that can fundamentally change the world,” says Elaine Zelby, co-founder Tofu, which is developing a generative AI–based marketing automation system. “There’s an energy and speed at which people operate that just doesn’t exist elsewhere,” she says. “You can see it. You can feel it.”

It’s not just the social scene and sense of community that’s drawing people back to the Bay. Just when many techies had decided the right to work remotely was universal, the generative AI crowd appears to have taken a hard turn back toward the office. Rental space is already getting tight in San Francisco neighborhoods such as the Mission, Jackson Square, and Cerebral Valley (formerly known as Hayes Valley). “People are itching to get back to the office,” says Shin. “The energy you get when people are working in the same place and feeding off each other is so important for companies like us.” There’s even a half-serious new mantra among some investors, says Yu: “If a company is fully remote, proceed with extreme caution before investing.”

You can also move faster with fewer employees, especially with the right AI talent on the team. And this desire to run lean — not to mention the unavoidable risks involved with a market that’s evolving as fast as generative AI — is impacting many founders’ views on fundraising. With an all-star team hailing from OpenAI and elsewhere and high-profile backers such as Google chief scientist Jeff Dean and YouTube CEO Susan Wojcicki, Perplexity AI could potentially have raised more than the $28.6 million A round it secured in March. But the team, which created a novel gen-AI search engine with just four people in four months, wanted to stay focused on speed rather than on laying the groundwork for a bigger company. "It's not always about raising a gazillion dollars,” says Larco. “In some cases it's 'let's raise as little as we can and figure out if the thing's going to work.’”

As a participant in the great technological shifts that have shaped the firm’s nearly 50-year history, today NEA is doubling down on Silicon Valley by hosting events and building a more formalized founder network. “We’ve seen this story before,” says NEA CEO Scott Sandell. "Netscape wasn't a big company, but everybody could see it was a game changer. The frenzy we're seeing in AI is similar to that. Yes, people are paying irrational prices for some companies, and there will no doubt be a bust cycle. But there's no doubt that some incredibly transformative companies will come out on the other side."

New Norms

So how is this cultural shift changing the art of company building? Here are a few clear trends we see emerging from Silicon Valley’s AI boom:

Build first, then fundraise

When a market is developing this fast, VCs are not as interested in who you are, where you went to school, or how much you’ve raised in the past. That’s because top teams are coming to seed investors with a team already in place and a beta product to show, not just a slide deck. Perplexity, for example, had launched its AI-based search engine and was getting 300,000 queries a day when they first pitched NEA in February 2023, says the company.

Don’t be a tourist

Given the huge potential for companies that take the early lead in generative AI markets, VCs are unlikely to get excited about the fifth entry into a particular segment. And investors are hyperaware that ChatGPT can be used to create a cute front-end for a weak product idea. “It’s not hard for almost anyone to spin something up in a few weeks that integrates GenAI,” says Anand Kulkarni, CEO of Crowdbotics. “The barriers to entry are low and getting lower.”

Think lean

Generative AI clearly has the potential to increase productivity by taking on some or even all of the responsibilities of certain jobs. This is a sensitive subject, given concerns about job loss. But startups that understand generative AI are already thinking of how to build companies with far fewer people than are typically employed today. They’re also looking to realign incentives around headcount, so that people are not rewarded for the size of the groups they run. “We're building an amazing company with 15 people,” says Dan Siroker, founder of Rewind AI. “In the past we might have needed hundreds of people. It’s a different way of thinking.”

Don’t over-raise

While marquee AI companies are securing massive valuations, there’s a sense among generative AI founders that less can be more for many companies. “There’s a huge group of founders who have seen their companies killed because they raised too much,” says Zelby. As a result, she notes, we’re seeing a growing number of founders looking to raise seed rounds in the $4 million to $6 million range. “That’s larger than historical seed rounds, but not by a ridiculous amount given the market opportunity of generative AI.”

Be open

Silicon Valley has always been a melting pot of ideas, Larco notes, but this is a time to be particularly generous in sharing what you know to help other companies and the overall ecosystem. Especially given the many ethical and other concerns about AI, efforts to develop effective, fair norms and standards will help build an industry that benefits everyone.

Aim high

Given how fast generative AI is developing, nobody can claim to know how industry dynamics and market structures will settle out. So be bold, and pursue ideas that do not fit into existing business models and categories.

Speed is key

“The biggest thing I perceive as being different is the pace of change,” says Luke Hoban, CTO of Pulumi, a multicloud infrastructure startup. “It feels like there’s a fundamental new capability every few weeks.”

Of course, Silicon Valley has never had a monopoly on technology entrepreneurship or startup culture. In fact, AGI House’s Yu is in talks with VCs in New York about opening an AGI House in Manhattan, where NEA also has an office. No doubt, great talent exists everywhere, especially in great cities like the Big Apple. But for now, at least, Silicon Valley has staked a compelling claim as the place to be to participate in the AI technology revolution. “Those VCs in NY are blown away by the energy out here. They want to know how we can bring some of that out here!"