Almost every day, Silicon Valley entrepreneur Grantley hears from investors trying to convince him to make money. Some have sent him a personalized gift basket of him and his co-founder.
Mr. Lee, 41, is usually flattering. In the past, fast-growing startups like Gamma, an artificial intelligence startup that helped set up in 2020, would have always been looking for more money.
But like many young startups in Silicon Valley today, Gamma is pursuing a different strategy. We use artificial intelligence tools to increase employee productivity in everything from customer service and marketing to coding and customer research.
So Gamma, which makes software that allows people to create presentations and websites, doesn’t require more cash, Lee said. His company only hires 28 people to attract “tens of millions” and nearly 50 million users on its recurring annual revenue. Gamma is also profitable.
“If we came from a generation before, we can easily become 200 employees,” Lee said. “It basically gives you the opportunity to rewrite the playbook and rethink it.”
The old Silicon Valley model decided that startups should raise huge sums from venture capital investors and spend time hiring an army of employees to expand quickly. The profit will come much later. Until then, head numbers and fundraisers were badges of honor among founders.
However, Gamma is one of the cohort of startups, most of which work on AI products, using AI to maximize efficiency. They make money and are growing quickly without the funds and employees they previously needed. The biggest bragging rights of these startups are to make the most profitable with the least workers.
The story of the success of “small team” has become a meme. Technicians are excitedly sharing a list showing how Anysphere, a startup that creates coding software cursors, has reached $100 million in annual revenue in under two years with just 20 employees. and how ElevenLabs, an AI voice-activated user, did the same with about 50 workers.
The possibility that AI could allow startups to do lesser amounts has led to wild speculation about the future. Openai CEO Sam Altman predicts that one day there could be a single company worth $1 billion. His company is building a cost-intensive AI called the Basic Model, but employs more than 4,000 people and raises more than $20 billion in funding. They are also meeting to raise more funds.
With AI tools, some startups have declared they will stop adoption at certain sizes. Financial software company Runway Financial says it plans to attract 100 employees, with each worker doing 1.5 jobs. The agency, a startup that uses AI for customer service, also plans to hire fewer than 100 workers.
“It’s about eliminating roles that aren’t needed when you have a small team,” said Elias Torres, the agency’s founder.
The idea of AI-driven efficiency was bolstered by Deepseek last month. This is a Chinese AI startup that has shown that AI tools can be built for a small portion of the general cost. That breakthrough was built on open source tools that are freely available online, causing an explosion of companies using DeepSeek’s inexpensive technology to build new products.
“Deepseek was a basin moment,” said Gaurav Jain, an investor at Afore Capital, a venture company that backed Gamma. “The calculation costs are very fast and very fast.”
Jain compared the company wave that occurred in the late 2000s with new AI startups after Amazon began offering cheap cloud computing services. This reduced the cost of starting a company and caused a surge in new startups that could be built at a cheaper price.
Before this AI boom, startups generally burned $1 million and earned $1 million in revenue, Jain said. According to the analysis of the aforementioned 200 startups, $1 million in revenues could now fall by a fifth, and ultimately drop by a tenth.
“This time, we’re automating people, not just data centers,” Jain said.
However, if startups can be profitable without spending too much, it can be a problem for venture capital investors. Last year, AI companies raised $97 billion in funding, accounting for 46% of national venture investments in the US, according to Pitchbook, which tracks startups.
“Venture capital only works when the winner gets paid,” said Terrence Rohan, an investor in the fund, focusing on very young startups. He added, “If you need less money, because there are fewer future winners, how would that change the VC?”
For now, investors continue to fight to enter the hottest companies, many of which don’t require more money. Scribe, an AI productivity startup, has attracted much more interest from investors last year, and has far more interest than the $25 million it wants to raise.
“It was the smallest amount of negotiation we could probably undertake,” said Jennifer Smith, CEO of Scribe. She said investors were shocked by the size of 100 staff compared to their 3 million users and rapid growth.
Some investors are optimistic that AI-driven efficiency will encourage entrepreneurs, create more companies and increase investment opportunities. They hope that once the startups reach a certain size, companies will adopt older models of large teams and big money.
Some young companies, including Anysphere behind Cursor, have already done that. According to Oskar Schulz, the company’s president, Anysphere has raised $175 million in funding and plans to add staff to conduct the investigation.
Other founders have seen the dangers of old startup playbooks. This has kept businesses on fundraising treadmills that have created more costs beyond their pay by hiring more people.
Large teams needed managers, more robust human resources and support from back offices. These teams needed specialized software along with a larger office with all the perks. , such as the startups have led to burning cash, forcing founders to always raise more money. Many startups from the 2021 fundraising boom eventually reduced, shut down or scrambled to sell themselves.
Early profits can change the outcome. At Gamma, employees use approximately 10 AI tools to use Intercom’s customer service tools to handle issues, Midjourney’s image generator, Anthropic’s Claude Chatbot for Data Analysis, and Google’s NoteBooklm. and analyze customer surveys. Engineers also use AnySphere cursors to write code more efficiently.
Gamma’s products are built on tools such as Openai, but are not as expensive as other AI products. (New York Times sued Openai and its partner Microsoft, alleging that it infringes copyright in news content related to AI systems. Both companies denied the allegations in the lawsuit.)
Other efficient startups have similar strategies. Thinking, the 10 provider of AI mobile phone agents, has made a profit in 11 months thanks to its use of AI.
Payment Processor Stripe has created an AI tool that helps Leonard analyze thoughtful sales. And without AI tools from others, he said it would require 25 people thoughtfully to streamline its operations, far from profitable.
Leonard said he would think and ultimately raise more money, but that’s only when he’s ready. He said not to worry about a lack of cash is a “big relief.”
In Gamma, Lee said he plans to roughly double the workforce this year to 60 and hire them for design, engineering and sales. He has previously planned to hire a different type of worker, and is looking for a generalist who does a variety of tasks, rather than a specialist who only does one thing. He also wants a “player coach” instead of a manager. You can mentor experienced employees, but you can also take part in your daily work.
Lee said that AI-efficient models freed up the time they otherwise spent managing and recruiting people. Now he talks to customers and focuses on improving the product. In 2022 he created a slack room for feedback from top Gamma users. He is often shocked to discover that the CEO is responding to their comments.
“It’s actually a dream for every founder,” Lee said.
