May 14, 2024

Davids & Goliath

Hi there, 

The battle for search is on. 

Reports swirled last week that OpenAI is working on an AI-powered search product. 

While the company didn’t announce a search engine on Monday, the speculation underscores how genAI is transforming the search landscape — and potentially Google's place in it.  

Below we cover Google’s AI search plans, the startups entering the scene, and what might be next.

The king of search

Google remains by far the dominant search engine, with an estimated ~90% global market share. 

But it’s clear Google is feeling the heat.

AI was the dominant topic on its Q1’24 earnings call late last month and at its developer conference today.

Google CEO Sundar Pichai emphasized the importance of the genAI moment for search on the call: 

“AI innovations in search are…perhaps the most important point I want to make. We have been through technology shifts before to the web, to mobile, and even to voice technology. Each shift expanded what people can do with search and led to new growth. We are seeing a similar shift happening now with generative AI.”

The company has been experimenting with a “search generative experience,” which gives AI overviews to beta users in their main search results page. It's rolling out this feature to users in the US this week. 

“We are being measured in how we do this, focusing on areas where genAI can improve the search experience while also prioritizing traffic to websites and merchants. We have already served billions of queries with our generative AI features.” 

Search is currently responsible for more than half of Alphabet’s revenue, so  Wall Street analysts are focused on Google's ability to maintain its formidable ad machine, as highlighted by earnings insights on CBI. 

Startups targeting search

A number of startups are trying to rethink search by using generative models to parse and summarize results upfront in response to users’ queries.

Perplexity, one of the biggest names in the space, hit unicorn status in April after raising $62.7M. 

Others include You.com — which has brought together a team of research scientists with experience at Salesforce and Stanford — and Andi, which is backed by Y Combinator.

Some are targeting more specific use cases:

• Twelve Labs ($27M in funding) — building search for scenes within videos
• Elicit ($9M in funding) — offers researchers search and summarization for academic papers
• Phind (backed by YC)  — answer engine targeted at software developers

The business model 

AI-powered search engines still need to figure out a search business model to compete with Google. 

Importantly, these startups are primarily building “answer” engines that try to resolve queries within the search platform itself, whereas Google has built its business on navigational queries — getting users to other websites via result links.

Perplexity, which operates under a freemium model, offers a $20-a-month subscription for premium features including allowing users to select their preferred model and upload files. Neeva, another search engine that tried its hand at a subscription search model, struggled and was ultimately acquired by Snowflake. Of course, Neeva pre-dated a lot of the advancements in generative AI, so Perplexity is a different animal.

In April, Perplexity said it had plans to start selling ads this year related to users' queries. 

That same month it launched an enterprise offering with additional data privacy and security features. Early customers include Stripe, Zoom, and HP. 

But it’ll be a tricky road to compete with Google’s free search. 

Microsoft’s new AI-powered Bing, launched last year, has barely dented Google’s market share. 

Beyond search: agents

Search is already moving to become more personalized, multimodal, and intuitive. 

Google highlighted its plans today to leverage Gemini as an agent "under the hood" of search to answer more complex queries, like planning a trip or creating a meal plan.

This enterprise customer of Perplexity we recently spoke with envisions a future where AI chatbots not only find information but also understand context, interact with websites, and take action on the user's behalf.

At OpenAI’s Spring Update on Monday, the company debuted a fresh model, updating the ChatGPT experience with real-time conversation, vision, and translation capabilities.

It’s not a hard leap to imagine this “assistant” becoming adept at interacting with the web. 

All this competition should mean good things for customers who will benefit from this innovation.

I love you. 

Anand 

@asanwal 

P.S. New buyer interviews on CB Insights this week include: 

• AI development: Hugging Face (CSAT: 9/10)
• Banking: Personetics (CSAT: 8/10)
• ML security: Protect AI (CSAT: 9/10)
• Medical devices: Theranica (CSAT: 9/10)

Get started with CB Insights

CB Insights' emerging technology insights platform provides all the

analysis and data from this newsletter. Our data is the easiest way to discover and respond to emerging tech. 

Was this email forwarded to you? Sign up here