Perplexity's $20B Rise: How AI Agents Will Replace Ads, Jobs, and Your Shopping Habits by 2030

Watch the Full Video

Aravind Srinivas, CEO of Perplexity (now valued at $20 billion), reveals how AI agents will fundamentally restructure advertising, eliminate entry-level professional jobs, and create a new economic model where your AI assistant negotiates with brands on your behalf. This isn’t theoretical - Perplexity’s Comet browser already demonstrates this future.

Key Insights

• Perplexity achieved 133x valuation growth (from $150M to $20B) in under 2 years using the 1.01^365 principle - making 1% improvements daily compounds to 37.78x annual growth
• Comet browser represents the first AI agent architecture where AI works for users against advertisers, not the other way around
• Traditional advertising margins will collapse as users gain AI agents that can be instructed to ignore ads entirely
• Entry-level financial advisory and real estate jobs face immediate obsolescence as AI agents can analyze portfolios and scan listings more effectively
• Value in professional services now comes exclusively from exclusive access: off-market real estate deals, private equity funds, hedge fund access
• PhD education losing relevance for AI careers - “learning to learn” now happens faster through AI tools than through 4-6 year academic programs
• Immigrant survival instinct (“we are all survivors”) often more valuable than credentials in competitive AI markets
• Future revenue model: advertisers pay your AI agent for attention, agent shares revenue with you, creating user-aligned economics
• Building successful AI products requires obsession over market strategy - “bet on yourself, not the market”

From $150M to $20B in Two Years

“If you do a 1% improvement every day, how much do you improve at the end of the year? You improve 3,700%.”

Perplexity’s growth story demolishes conventional startup wisdom. When Srinivas last appeared on Silicon Valley Girl in November 2023, Perplexity was closing a Series A at $150 million valuation. Two years later, the company is valued at $20 billion.

The mechanism behind this exponential growth is deceptively simple. Srinivas wakes up every morning and reads everything users are saying on social platforms. Most bugs get fixed immediately. The team maintains an obsessive focus on the compound effect: 1.01 to the power of 365 equals 37.78.

“It may look like small steps every day but as the changes propagate to a larger scale of users they bring in more users through word of mouth,” Srinivas explains. Users experience tangible improvements week over week. The product feels faster, cleaner, more accurate. This builds trust, which compounds into growth.

The company secured brand partnerships with Formula 1 and Lewis Hamilton, associating Perplexity with excellence. As Srinivas notes, Steve Jobs ran the “Think Different” campaign celebrating Einstein, Gandhi, and the Beatles. Perplexity follows the same playbook: associate with the greatest of all time.

Comet Browser: Your AI Agent That Protects You

Perplexity released Comet in July 2024 - the first agentic browser that can think, act, and execute multi-step tasks on your behalf.

In the interview, Srinivas demonstrates Comet finding a specific Jensen Huang video clip. He asks Comet: “There’s a video where Jensen Huang says ‘I would rather torture people to greatness than fire them.’ Can you find it and play it for me from that exact moment?”

Comet performs multi-step reasoning: searches for the video, pulls the transcript, finds the exact quote, opens a tab, and starts playback at the precise timestamp.

But Comet goes further. It can analyze hour-long videos and extract only non-trivial insights. When Srinivas asks “What are some non-trivial things Jensen said in this interview?” Comet reads the entire transcript and identifies unique statements like “GPUs are time machines” - something Jensen hasn’t repeated in other interviews.

Then Comet can email those insights directly. Or schedule meetings. Or find and purchase the leather jacket Jensen is wearing. The browser becomes an extension of your cognition and action.

The Advertising Model Flip

“Amazon has AI do the recommendation ranking. Google has AI to do the search ranking. So they can influence things for and manipulate you to buying stuff. But for the first time you have an AI in your hands. It actually protects you. It actually gives you power against the big tech.”

Srinivas articulates a profound shift in the advertising economy. Until now, AI existed exclusively in the hands of advertising companies. Amazon’s AI optimizes recommendations to maximize purchases. Google’s AI ranks search results based on ad revenue.

Users had no AI protection. You were defenseless against algorithmic manipulation.

Comet changes this power dynamic. For the first time, you have an AI agent working for you, not for advertisers. You can instruct Comet: “Skip all the ads.” When Comet watches YouTube videos for you, it can extract core content while ignoring sponsorships.

The future advertising model Srinivas envisions has two layers of protection. First, you never see ads directly. Second, even if your agent sees ads, the platform shares revenue with you.

“The company that runs the agent gets some of it. But the user also gets a little bit,” Srinivas explains. “That’s the best way to continue to win the user’s trust by saying okay I’m actually open to being advertised to through my agent and in return for that I actually want some money back.”

Advertising margins will compress dramatically. But user trust and lifetime value increase because users maintain control.

Entry-Level Jobs Are Disappearing

“I certainly think so like people need to push further.”

When asked directly whether entry-level jobs are disappearing, Srinivas doesn’t hedge. “I certainly think so.”

The mechanism is straightforward. Entry-level financial advisors primarily manage stock portfolios. But Perplexity (or any AI agent) can read all analyst reports, morning news, and financial filings across every stock in seconds.

Srinivas describes asking Perplexity: “What would Warren Buffett do if he had exactly your portfolio?” No human financial advisor can answer questions like that.

The interviewer shares her personal experience. She asked Perplexity to analyze her portfolio and asked whether she should fire her financial advisor. Perplexity’s response was definitive: “Yes you should do this immediately.” The AI identified expensive mutual funds and created a complete rebalancing strategy.

Entry-level real estate agents face the same fate. Srinivas envisions AI agents running recurring tasks: “Every morning just look at all the new listings on Redfin. These are the preferences I have. Every time there’s a house below a certain budget and that’s in within this neighborhood, make sure to send me a push notification and submit an application on my behalf and schedule a meeting or a visit.”

If your only value is showing listings that already exist on Zillow or Redfin, you’re obsolete.

What Financial Advisors Must Do to Survive

Srinivas uses his own private wealth managers as an example of what still provides value. They help him move money around - mundane operational tasks he doesn’t have time for. They provide tax optimization advice. They manage administrative complexity.

More importantly, they provide access. Access to private equity funds. Access to hedge funds like Citadel. Access to venture capital opportunities. These assets don’t trade on public markets. They require relationships, credentials, and minimum investment thresholds.

“If your financial advisor does more stuff for you like getting you some access to certain funds like private equity or hedge funds or access to like funds of Citadel or something like that that you cannot get through just the public markets then maybe they’re useful.”

The pattern holds across professions. Your value cannot come from analyzing publicly available information. AI does that better. Your value must come from exclusive access, proprietary networks, or unique expertise that doesn’t exist in training data.

Real Estate Agents: Access Over Listings

“If your agent only shows you listings on Redfin and Zillow, I think they’re useless. But if your agent is doing off-market deals for you, they’re still pretty useful because those things do not exist on the web.”

This quote crystallizes the survival criterion for real estate agents. Listings on Zillow and Redfin are structured data. AI agents can scan, filter, and analyze them instantly.

Off-market deals don’t exist on the web. They require relationships with sellers who haven’t listed publicly. They require trust, reputation, and negotiation skill. These remain human advantages.

Srinivas emphasizes that having AI intelligence at your disposal means you can “push the existing industry to work harder for you.” Don’t accept the status quo. If your real estate agent only emails you Zillow links, demand more value or find someone who provides off-market access.

The same logic applies everywhere. Having an AI agent means professional service providers must justify their fees with value AI cannot replicate.

PhD Education in the AI Era

“The fundamental thing you learn in a PhD is learning to learn. But 18 is too early to know what you’re interested in.”

Srinivas completed a PhD before founding Perplexity. But when asked if he would do it again if he were 18 today in the US, his answer is revealing.

“Realistically, it’s gone.”

The core value of a PhD is “learning to learn.” You acquire the ability to identify a new topic, dig deep, gather information, consult experts, ask the right questions, develop subject expertise, and make decisions.

But Perplexity now provides instant access to expert knowledge. The consultation phase that once required years of academic networking now happens in seconds.

Srinivas did his PhD partly because it was his only path to America. Indian students who can’t afford master’s tuition can get funded PhDs. But beyond immigration logistics, he genuinely loved AI and enjoyed deep learning.

For someone starting today, the calculus is different. “Unless you’ve done some things like that in your life already like try to actually go and do something like commit yourself to one thing and do it for a sustained period of time because it’s very hard to be good at something if you just do it for like a couple of months.”

The recommendation: pick something and do it deeply for 1-2 years. Build the confidence that you can bet on yourself to learn anything and do it well. That confidence - not the credentials - is what matters.

Building in AI: Bet on Yourself, Not the Market

“The only thing you can bet on is whether you are so obsessed about a topic that you will do it anyway regardless all the odds stacked against you.”

When asked about opportunities for new AI founders, Srinivas is brutally honest. Building foundation models is “almost impossible” for new entrants. Not because training one model is expensive, but because it requires relentless iteration.

An investor once told Srinivas it costs $200,000 to train a GPT-3 model. “I’ll give you a million and you spend 200k on the model.” But that’s not how it works. You need continuous model iteration, clusters, talent, infrastructure, and a path to revenue.

Even product companies face intense competition. “Everyone’s going to do anything that works,” Srinivas warns. ChatGPT will do shopping. Perplexity will do commerce. There are no protected niches.

So what should founders do? “Do what you truly are obsessed about because fundamentally it’s a bet on yourself. It’s not a bet on the market.”

Don’t try to be a “whiteboard strategy master” analyzing competitive dynamics. When your idea works and generates significant revenue, existing players will enter your market. That’s guaranteed.

The only sustainable advantage is obsession. “Build for yourself hopefully that’s a thing that a lot of people in the world want and therefore you can turn it into a scalable product.”

Srinivas describes Perplexity as built for the founders themselves. “We built it for ourselves.” The bet was that everyone in the world loves having an AI they can summon to do tasks. Everyone wants more time. If you’re solving your own problem intensely, there’s a chance millions of others share it.

Key Quotes