OpenAI’s dominance is unlike anything Silicon Valley has ever seen

Virtually every successful tech startup throughout history has faced the reality that it could get swallowed up or run over by a large incumbent at any moment. It’s part of daily life for an entrepreneur.

But the company at the center of the current boom is a different kind of beast.

Unlike industry giants of past eras, OpenAI is privately held. Its financials are mostly secretive and its ability and willingness to spend other people’s money is unrivaled.

And as OpenAI has proven recently, through a dizzying array of mammoth deals and product rollouts, the artificial intelligence lab is investing up and down the stack — from the picks and shovels of data center development to consumer applications, coding tools and even devices. Its flagship ChatGPT chatbot has reached 800 million weekly users.

“If you’re an entrepreneur, you have to ask yourself, ‘Where is the white space?'” said Nina Achadjian, a partner at Index Ventures who focuses on AI.

Not that Achadjian is staying out of the market.

In a transaction announced on Wednesday, Achadjian and Index led a $25 million investment in Quilter, which uses AI to develop software for printed circuit boards (PCBs). The company was founded in 2019 by former SpaceX engineer Sergiy Nesterenko.

Achadjian described Quilter as “pretty niche” and “not built on top of any model.” She said it’s unlikely that OpenAI will compete in a space where companies like Cadence Design and Synopsys have long developed technology for chip design, adding that a PCB is in every consumer device, light bulb, car tire and virtually all electronics.

Still, “there is no predictability,” she said. Relative to past cycles, “it’s more opaque and hard to predict which direction those guys are going to go.”

In less than three years, OpenAI has ballooned from an AI startup led by the guy who used to run Y Combinator to a $500 billion goliath spearheading a data center buildout plan endorsed by the White House and in partnership with the world’s most valuable company, Nvidia.

The past few months have only gotten crazier.

CEO Sam Altman is everywhere, forging massive infrastructure-related agreements with Nvidia, Broadcom Oracle and AMD. Last week, his company rolled out its Sora AI video app, which hit 1 million downloads in less than five days, and this week he keynoted OpenAI’s DevDay in San Francisco, attended by roughly 1,500 developers.

At DevDay, Altman announced the general availability of Codex, OpenAI’s software engineering agent, and said Sora 2 is now in he application programming interface (API), and can be tested by coders. He also took the stage with iPhone designer Jony Ive, who joined OpenAI in May as part of a $6.4 billion talent hire, with the mission to build AI hardware.

Ive has remained vague about what exactly he’s building, and he told Altman on stage that he hopes to develop tools that “make us happy and fulfilled and more peaceful and less anxious and less disconnected.”

OpenAI has clearly positioned itself as the defining company of the generative AI era. It follows other category-defining consumer internet brands of the last few decades, namely Amazon in e-commerce and cloud infrastructure, Google in web search and digital ads, Facebook in social media and Apple in mobile apps.

In those market booms, successful startups were born, and many more failed for any number of reasons, including an inability to find a big enough lane to build a sustainable business or get direct distribution.

Mobile app developers have been forced to reach users through extremely crowded app stores owned by Apple and Google. Facebook and Google became the requisite avenues for finding customers on the internet, and Amazon Web Services emerged as the go-to platform for startups to launch their companies as a more efficient alternative to buying their own servers.

In each case, the big platform companies released tools and features that competed directly with their customers, sometimes wiping them out.

Ethan Kurzweil, managing partner of Chemistry Ventures, said the biggest difference today is speed.

AI startups are formed and quickly propelled to historic valuations, and OpenAI is moving even more rapidly, launching services that compete with AI coding tools, agent kits and other apps running inside ChatGPT.

“It’s the fastest-moving time in startup creation and disruption in my 17 years of investing,” said Kurzweil, who spent the first 16 of those years at Bessemer Venture Partners before starting Chemistry in 2024.

Kurzweil said that OpenAI is doing a lot of things that are “theoretically scary for a lot of people,” but that there’s a “gold rush mentality where a lot of companies will do well.”

One prevailing view is that AI startups should target industries that are heavily regulated or have other dynamics that make them less likely to choose general purpose AI services.

In health care, Heidi Health and DUOS announced big rounds this week, while EvenUp and Spellbook pulled in hefty amounts of capital to go after lawyers.

“There are a lot of areas, like in finance and health care, where buyers want somebody that will speak their language,” Kurzweil said.

In late September, Chemistry hosted an event featuring OpenAI operating chief Brad Lightcap. Kurzweil said a big topic of conversation among attendees was that there are no “technical moats.”

That’s evident at the foundational model level, where OpenAI is facing off against Anthropic, Google, Meta and others.

Rather, the advantage companies have is their momentum, which helps explain why OpenAI has lined up hundreds of billions of dollars worth of high-profile deals with major tech players of late while simultaneously giving its rapidly expanding user base more apps and features.

In the process, OpenAI is burning mounds of cash without having to worry about Wall Street’s reaction.

“There’s no reckoning, because none of the companies are public,” said Achadjian, referring primarily to OpenAI and Anthropic, which last month said it raised $13 billion at a $183 billion valuation. “That further fosters the exuberance of capital raising, capital spending and vertical integration.”

Representatives from OpenAI and Anthropic didn’t respond to requests for comment.

In the first half of the year, venture growth-stage investments reached $83.9 billion, driven by five billion-dollar-plus AI deals, according to the second-quarter report from the National Venture Capital Association and Pitchbook.

On an annualized basis, that would far surpass the peak in 2021, when $96.1 billion was deployed at the growth stage.

“AI continues to dominate the upper end of the deal spectrum,” the report said.

Exa Labs, which describes its product as “search built for AI,” raised an $85 million Series B round in September, from investors including Nvidia, at a $700 million valuation. Founded in 2021, Exa introduced its first search engine in November 2022, two weeks before the release of ChatGPT.

“It would be really surprising to see a company that doesn’t compete with OpenAI,” Exa co-founder Jeff Wang said in an interview this week. “We’re in the same boat as everybody.”

But Wang said that OpenAI is an asset to his startup and the broader ecosystem in that it’s building tools that make other companies’ products better.

Wang said that while OpenAI may be in the search market — many people are using ChatGPT now instead of Google — the new world that we’re entering won’t be dominated by a single search engine.

Wang said that hobbyists and people building AI products are paying for Exa’s service, and that it’s being used within companies that have specific and “gigantic needs.”

“The pie is really big and OpenAI is just one company,” Wang said.

— CNBC’s MacKenzie Sigalos and Ashley Capoot contributed to this report.

WATCH: OpenAI’s deal spree