2025创业报告：人工智能时代的创业之道

What venture capitalists are looking forfrom cutting-edge AI founders The Future Starts Right Here3AI Investment in 2025: Bubble, Rising Tide or Both?5ARR vs. ERR: Weighing Experimental Income for AI Startups8Pricing Software in the Age of AI10Product Roadmap: Separating Signal from Noise12The Pull of Data Gravity14The KPIs That Matter15GTM and Partnerships for AI-Native Startups17How AI Founders Nail the VC Meeting19Early Days, Distant Horizons21Next Steps22 TABLE OFCONTENTS THE FUTURE STARTS RIGHT HERE We’re told, repeatedly and fairly convincingly, that the vanguard of advancedAI technology being led by large language models and generative AI is going tofundamentally change the world. The work being done in labs and office parks,the funding decisions being made by venture capital firms, will cumulativelycatapult the world into a new era. EIGHT PERSPECTIVES ON AI STARTUPS But if AI is going to change everything, surely it will start with that very process,the way the businesses bringing AI to market are born and nurtured. It seemedlike a smart theory, and it matched what we were seeing from the new businessesworking with the Powered by Snowflake Startup Program. So we decided to doa little digging. MATS OLOV MAGDESJÖ AKASH BAJWA Operating Partner for Analytics, EQT Ventures Principal, Earlybird Venture Capital ROHINI CHAKRAVARTHY LIAM MULCAHY Managing Partner, NewBuild Venture Capital Go-to-Market, Kleiner Perkins To understand what the age of AI has done to the perpetual pursuit of funding,we interviewed a group of venture capitalists who partner with Snowflake’sstartup program. Eight experienced funders in the United States and Europeexplained how their approach has changed when it comes to considering andfunding an AI-based startup — and how much it hasn’t. SHRAVAN NARAYENPartner, IVP PATRICK CHASE Partner, Redpoint Ventures The first thing we learned is that it’s kinda crazy out there. Every time you blink,the performance and cost of artificial intelligence has taken a significant leapforward. Patrick Chase, a partner at Redpoint Ventures, says he saw that whenhe and his wife had their first child in mid-2023. SAM TEDENInvestor, Anthos Capital SAKIB DADIPartner, Stage 2 Capital “One of the big differences with AI is that moreexecutives have ownership over it and benefit fromit,” says IVP partner Shravan Narayen. “In the past,a business could theoretically isolate its data andanalytics strategy under a few people. But your AIstrategy encompasses finance, product, engineering,IT, sales — every single person affects it, is affectedby it, and has a stake in it.” “I was on paternity leave for about eight weeks, andwhen I came back, things were completely different.Llama had shipped 7B, 13B and 70B models, OpenAIhad shipped new models and so did Anthropic,”he says. “It’s just crazy how fast things are moving.” POWERED-UP STARTUPS Not only do the specifics of the technology changeon an all-but-daily basis, but the playing field haseffectively been leveled by a sheer tidal wave of new. ThePowered by Snowflake Startup Programprovides startups the resources they need tobuild and grow data-intensive applicationsand products. “AI is an incredible trend to invest behind,” Chasesays. “The world will probably be really differentin one, three, five, 10 years from now, and I thinkthat creates a lot of opportunity for startups andventure capital.” “There are all these people who have been doing AIfor a long time, but in terms of the state of the artright now, the transformer models, it’s only existedsince 2017,” Chase says. “So in some ways, the mostexperience someone can have is a couple of years.” Hundreds of startups are already buildingapplications with Snowflake for a wide varietyof use cases, including marketing automation,monitoring and observability, IoT, machinelearning, embedded analytics and many more. The insights and opinions shared across eightin-depth conversations largely break down intoeight categories, ranging from whether experimentalrecurring revenue is actually a thing to what,specifically, excites these investors when firstmeeting a founder. It’s not just the technology that’s in flux. How anAI startup sells its product, and who it sells to,is changing as well. AI INVESTMENT IN 2025:BUBBLE, RISING TIDE OR BOTH? AI is turning up everywhere that something can be sold to an enterprise client. Half the sandwich shops in SiliconValley are probably spelling the name of their fancy mayonnaise as “AI-oli.” There is broad public excitement andanticipation, of course, around new generative AI products that provide direct interaction with large languagemodels. But the investors we talked to are also looking at, and for, AI capabilities being applied to existingproducts or business workflows — less “AI” and more “AI under the hood.” But in addition to the questionof whether Startup X is a good investment, they also have to weigh a lot of talk about bubbles and crashes. 2024 w