San Francisco – Jordan Jacobs, an investor of the Radical Ventures Risk Capital Fund, has spent the past few days serving half a dozen calls from investors in his firm. Everyone wanted to know about the Chinese artificial intelligence app that led application stores over the weekend.
DeepSeek created a powerful AI model with much less money than most experts were found possible, challenging many assumptions underlying the development of rapidly evolving technology. To calm panic, Jacobs said he explained to his investors that the radical ventures has long invested in more efficient Ia models similar to the created by DeepSek.
“We will focus on companies that are actually building real business instead of those who are pursuing science fiction,” Jacobs told them.
Continues after advertising
Read more:
Nvidia, Google, Goal, and other technology giants have faced a flood of questions about DeepSek since last week, when Chinese startup has toppled a longstanding AI notions. But their repercussions are being felt beyond the largest companies, reaching the risk capital industry that has bet on high technology, investing billions of dollars in AI startups.
For two years, risk capital firms were involved in a financing frenzy, dumping more than $ 155 billion in AI startups between 2023 and 2024, according to Pitchbook, which tracks startups. Two of these AI companies – OpenAi and Anthropic – raised $ 24 billion and $ 16 billion with the goal of building was as intelligent as humans. OpenAi’s review reached $ 157 billion – more than Pfizer or Citigroup – while Anthropic’s assessment reached $ 20 billion.
Continues after advertising
What DeepSeek has done now put this funding fever in question. If a Chinese startup can create an application as powerful as OpenAi’s chatgpt or Anthropic’s chatbot Claude with almost no money, why did these companies have so much money?
Read more:
“It’s not a good look now” for some AI companies “given your speech about the need for an increasing scale to create the best model,” Matt Turck, a Firstmark Capital investor. But, he added, AI companies would ultimately need money, computational power and infrastructure to serve their customers.
Continues after advertising
Risk capitalists have debated the best way to invest in AI since OpenAi launched Chatgpt in late 2022. Some investors argued that the technology that supports chatgPT and other products – often referred to as “foundation models” because they can feed Many applications, including chatbots, search engines, and image generators – is not a good investment because the systems are expensive to create and easy for competitors to copy.
Marc Andreessen, a Andreessen Horowitz investor, last year called such systems “running to the background” and speculated that building a business with that type of AI would be like “selling rice” where anyone can compete.
With the uproar caused by DeepSeek in recent days, risk capital investors who have not invested in foundation models such as OpenAi and Anthropic – either because they anticipated the race to the background or because they didn’t have the money or the opportunity – used – used The moment to share your opinions.
Continues after advertising
Eric Vishria, an investor of the Benchmark risk capital firm, said on social networks on Monday that he believed the foundation models were “the most quickly depreciated asset in human history.” Anjney Midha, an investor at Andreessen Horowitz, wrote that DeepSeek has shown that “the current structure of the AI foundation models market is far from stable.”
Investors who supported Foundation Model Companies defended their investments. Gavin Baker, an Atreides Management investor, who invested in Elon Musk’s AI startup, X.Ai, said he felt good about his bet because AI companies are limited by the amount of data they can access. The X.Ai, he said, was in a strong position because he has his own unique source of social platform X data, which Musk also has.
“For me, I feel very, very calm,” said Baker.
Other technology leaders have dissected Deepseek’s claim that it has spent only $ 6 million to create its AI model, which is a fraction of what other companies spend. Some pointed their finger at regulation, including the AI executive order of former President Joe Biden and California’s failed attempt to enact AI state law by trying to contain industry progress.
They also lamented powerful AI chips export restrictions as ineffective to prevent Chinese technological advances. Some attacked the so -called AI security advocates, who tried to slow down the development of AI due to its potential risks to humanity. Others invoked patriotism and said Deepseek was a sign that the United States needed to advance the AI faster. Still others saw the moment as an opportunity.
Turck said DeepSeek’s advance may be bad news for some of the largest AI companies, but it opened possibilities for other companies that were just starting.
“Panic in recent days is a dramatic exaggerated reaction,” he said in a message.
Niko Bonatsos, a risk capital investor at General Catalyst, said in an interview that DeepSeek energized the startups. “If you are building anything that touches it and has not been excited, obsessed, scared and deprived of sleep in the last four days, which planet are you living in?” He said.
c.2025 The New York Times Company
