Perplexity AI

Magic AI

Perplexity AI

Perplexity uses a freemium model. The free tier gives unlimited basic searches powered by a standard model.

Perplexity Pro costs $20/month and gives access to more powerful frontier models from OpenAI, Anthropic and Google, alongside Perplexity's own, unlimited Pro searches, file uploads, and image generation. Perplexity Enterprise Pro is designed for businesses with team management and API access.

The company also recently introduced advertising through "sponsored follow-up questions" — a controversial move that some see as inevitable and others see as the beginning of the end for ad-free search.

Magic AI

Magic is building an AI that works like a software engineer, not a chatbot. The goal is a model you can hand a whole code base and a task, and it writes, reviews, and debugs like a teammate.

Their edge is memory. In 2024 they showed off LTM-2-mini, a model with a 100 million token context window.

To put that in perspective, that is about 10 million lines of code the model can look at in one shot, by far the biggest of any model at the time. The plan is to sell this as an AI coworker to companies that write a lot of software.

As of the 2024 rounds, the product was still not for sale.

Perplexity AI

Aravind Srinivas grew up in Chennai, India, got a PhD in AI from UC Berkeley, and worked as a research intern at OpenAI and a researcher at DeepMind. He watched the ChatGPT explosion in late 2022 and saw something everyone else missed: AI chatbots were interesting, but AI-powered search was potentially more valuable.

Google search hadn't fundamentally changed since 1998. It still showed you a list of links and made you do the work of reading and synthesizing information.

Srinivas co-founded Perplexity AI in August 2022 with Denis Yarats (ex-Meta AI), Johnny Ho (ex-Quora), and Andy Konwinski (co-creator of Apache Spark). Their idea was an "answer engine" — you ask a question, Perplexity searches the internet in real time, reads the sources, and gives you a synthesized answer with inline citations.

No ads. No SEO spam.

Just answers.

The product launched quietly and grew through word of mouth among researchers, developers, and knowledge workers who were frustrated with Google's increasingly ad-cluttered, SEO-gamed results. Within a year, Perplexity was handling millions of queries daily.

Magic AI

Eric Steinberger and Sebastian De Ro started Magic in 2022. Steinberger is an AI researcher who became convinced that the missing piece for a true AI software engineer was memory.

Chatbots forget what you told them a few thousand words ago. He wanted a model that could read your entire code base and remember all of it.

They raised a small seed in 2022, then a $23 million Series A in early 2023 led by Alphabet's growth fund CapitalG. From day one the pitch was not a chatbot.

It was an AI colleague that could handle real engineering work.

Perplexity AI

Perplexity grew almost entirely through product quality and word of mouth. The first users were AI researchers and tech workers who shared it on Twitter.

Every time someone used Perplexity and got a better answer than Google, they posted about it. The product was its own marketing.

The mobile app was a growth accelerator. By launching on iOS and Android early, Perplexity captured users who wanted quick answers on their phones.

The app was faster than opening a browser and Googling — one tap, ask your question, get an answer. Mobile usage now exceeds desktop.

Strategic distribution deals expanded reach. Perplexity partnered with SoftBank (default AI assistant on SoftBank phones in Japan), Deutsche Telekom, and various hardware manufacturers to be pre-installed as the default search assistant.

These deals put Perplexity in front of millions of users who had never heard of it.

Magic AI

Magic's strategy is research first, product later, which is both unusual and expensive. Instead of shipping an app and iterating in public, they raised huge sums to chase a technical breakthrough in model memory.

The headline move was the 100 million token context claim, which got the whole AI world talking. They also cut a deal with Google Cloud in 2024 to build two supercomputers, called Magic-G4 and Magic-G5, using NVIDIA chips, to train bigger models.

Star backers like Eric Schmidt served as validation. The bet is that if the tech is far enough ahead, the product and the customers follow.

Perplexity AI

Google is the elephant in the room. Google controls 90%+ of the search market and generates $175 billion annually from search advertising.

Google has responded to Perplexity by adding AI Overviews to search results — essentially copying the answer engine concept. Google has unlimited data, unlimited compute, and the default position on every browser and phone.

Competing with Google for search is historically a death sentence.

The publisher problem is real and litigious. Perplexity synthesizes content from news articles and websites without driving traffic back to those publishers.

Multiple media companies have accused Perplexity of scraping their content and delivering it as answers, depriving them of page views and ad revenue. The New York Times, Forbes, and Condé Nast have all raised concerns.

If publishers successfully block Perplexity or win legal challenges, the product's utility decreases.

Monetization without ads is nearly impossible at scale. Perplexity initially positioned itself as the ad-free alternative to Google.

But $20/month subscriptions can't fund the compute costs of an answer engine serving hundreds of millions of queries. The introduction of sponsored questions signals that pure subscription revenue isn't enough.

Whether Perplexity can add ads without becoming what it set out to replace is the defining question.

Magic AI

Magic has raised close to half a billion dollars and, years in, still has no product the public can buy. That is the whole problem in one sentence.

A giant context window is a great demo, but rivals like Cursor, GitHub Copilot, and Cognition's Devin already have real users and revenue. Magic has to prove that more memory actually makes a smarter engineer, not just a bigger number on a slide.

And it has to do that while burning enormous sums on compute against competitors who are already shipping. The clock and the cash both keep running.

Perplexity AI

Perplexity Search is the core — ask any question and get an AI-synthesized answer with cited sources. Pro Search performs multi-step research, asking clarifying questions and digging deeper for complex queries.

Perplexity Pages lets users create shareable research articles from their queries. Collections organize saved searches and threads.

The Perplexity API lets developers integrate the answer engine into their own products. Perplexity recently added Spaces for team collaboration on research projects.

Magic AI

The centerpiece is Magic's LTM, or Long-Term Memory, model family. In 2024 they unveiled LTM-2-mini, a model with a 100 million token context window, pitched as the key to an AI that can understand an entire code base at once.

The broader product is an AI software engineer, or coworker, that can write and change code across a whole project. The company also announced Magic-G4 and Magic-G5, two supercomputers built with Google Cloud and NVIDIA to train these models.

Much of it is still research rather than a shipping product.

Perplexity AI

Jeff Bezos, IVP, NEA, Databricks Ventures, Nvidia, SoftBank

Magic AI

Magic's backers are a murderers' row of AI insiders. It started with a $5 million seed round in 2022 and a $23 million Series A in early 2023 led by Alphabet's CapitalG.

Then in February 2024 came a $117 million round led by Nat Friedman and Daniel Gross, two of the most connected investors in AI. The blowout was August 2024.

Magic raised $320 million with checks from former Google CEO Eric Schmidt, CapitalG again, Atlassian, Jane Street, Sequoia, and Elad Gil. That pushed total funding to around $465 million, all for a company still building toward its first sellable product.