HomeArticle

Global Top 15 Personal Agents: Why Are They Worth a Valuation of Tens of Billions?

潮涌AI2026-10-20 10:14
Capital is pricing the next-generation entry point in advance.

On September 28, a 14-person company based in San Francisco closed a $1 billion financing round at a $10 billion valuation, with HSG, Benchmark, and Coatue all participating in the same round.

The company was only registered in April, with a $50 million valuation at the seed round — five months, four rounds, 200 times growth.

It has no App, no revenue, and its product is still under an invitation-only system. The invitation link was posted on eBay by some people, with each one sold for $200 (Cybernews verified two on-sale links on September 30).

It is called Instinct, a company focused on Personal Agent: you send a text message, and it will use its own mobile phone and computer to book flights for you, cancel subscriptions, and call the dentist to queue up for you.

Why is it worth such a price? The answer is the entry.

In every generation of technological revolution, a super entry will emerge on the C-end: in the PC era it was the operating system, in the Internet era it was the search box, and in the mobile era it was WeChat and app stores.

Entry companies do not necessarily run all businesses, but they stand in front of all businesses and take the largest share of profits in the industry.

Now everyone is betting on the same thing: the C-end super entry in the AI era is Personal Agent.

It is not the same as the entries in the past.

Search and WeChat distribute information and traffic, answering the question of "where are you going"; Personal Agent distributes actions, which is "I will handle it for you". Once you get used to handing over your needs to it, shopping, travel, medical care, socializing, and all things will flow out from this entry.

The value of the Internet entry is to sell traffic to advertisers, and the value of the AI entry is to become the "person" between you and the world — these are two different types of business, and the ceiling of valuation is also different.

Investors' actions are more honest than analysis.

Khosla's reason for investing in Wajo has only two words: trust — he told TechCrunch on October 6 that giants in the advertising business naturally conflict with user privacy, and the architecture of the agent must put trust before growth from the first day; Vineeta Agarwala, a partner of a16z, explained why she invested in the medical agent company Ferry Health: "We have never had so much health information, what is missing is the action layer."

The entrepreneurs themselves are quite sober.

Qi Junyuan, founder of Today AI, told LatePost: "There is hype in this market, but it is still the next big thing." He also admitted that the current product maturity and commercial value do not match its valuation and popularity — the two most popular products, Meta has made a lot of placements for Muse, and Instinct relies on the invitation-only testing system with extremely limited access, both of their growth methods are very special.

When AI assistants have the capabilities of distribution, user relationship management and transaction at the same time, the market is willing to raise the valuation multiple first, and then wait for the profit to be realized. At the current price, what you buy is the category, not the financial statements.

The screening criteria for this list are: non-large factory and non-giant incubated entities, main business is personal AI Agent (capable of dialogue, and more importantly, executing real tasks), targeting the global market, with verifiable financing, user or revenue evidence; the ranking is comprehensively judged according to valuation scale, product verification and industry status.

The following are 15 companies in the Personal Agent track selected by Surge AI —

01 Instinct

Company Business:

There is no App, text messages, iMessage, WhatsApp, and telephone are all the interfaces.

It has its own computer and phone number, which can book flights and hotels, cancel subscriptions, and call customer service; the Concierge function launched on September 16 can directly call restaurants that do not accept online reservations, and Selections launched on September 29 is its first revenue experiment.

Founding Team:

Noah Shinn, the founder, is only 23 years old, and he was previously a research scientist at Sierra, a customer service agent company founded by Bret Taylor.

The Reflexion paper he co-authored with Yao Shunyuyu at the age of 20 is one of the earliest classics in the field of AI agent. The team has 14 people, and Shinn said he now spends 40% of his time buying computing power.

Prospect Highlights:

The financing track is the most amazing part of this industry: about $50 million valuation at the seed round in April (Conviction, Greenoaks, according to Forbes), $75 million in Series A in early August with a $5 billion valuation (led by Kleiner Perkins), $250 million in Series B on August 26 with a $2.5 billion valuation (co-led by Index and Benchmark), $1 billion in Series C on September 28 with a $10 billion valuation, with total financing of about $1.35 billion.

The operating data is also rare: the number of invitation-only users exceeds 100,000, the marketing cost is zero, and the daily growth is 10%; the annualized transaction volume is close to $1 billion, half of which comes from travel; 40% of users hand over their credit cards within three weeks of registration, and the subsequent retention rate is 80%.

According to records from the research institution Sacra, Instinct has no revenue so far, and the planned model is to uniformly draw a commission on the transaction volume — Shinn explicitly said that the $100 monthly subscription fee is a "local optimal solution", and the goal is to be free for life.

What the valuation buys is the entry option, not the financial statements.

02 Town (Townie)

Company Business:

A personalized agent that learns your working style.

The product Townie comes with a @town.com email, connected to your Gmail, calendar, Slack, and Notion, learns your work habits every night, and then takes the initiative to work — prepares meetings, organizes inboxes, and pushes forward matters.

Founding Team:

Jean-Denis Greze, the founder, is the former CTO of Plaid, an engineer by background, and has led a real infrastructure team.

Prospect Highlights:

It finished the beta on June 3, and the $55 million Series A was led by a16z. Greze once made a bold statement: what he is making will be the top three priorities of Google and Apple in the next 12 months.

The trust threshold in the work scenario is lower than that in the life scenario, and the willingness to pay is high, so the position it occupies is very smart.

03 Wajo (Fo)

Company Business:

The product Fo runs on iMessage, WhatsApp and web pages, positioned as an "action agent".

The most powerful design is that it comes with a one-time virtual credit card (patent applied) and a dedicated email, which will not expose the real card number at all when paying for you; when the AI gets stuck, you can directly hire a real person to get the thing done.

Founding Team:

Shivani Poddar, the founder, is a former DeepMind engineer, who participated in the launch of the first-generation Facebook Assistant and Gemini, and is one of the founders with the richest product resume in the list.

Prospect Highlights:

The financing amount has not been disclosed, and the endorsement list includes Khosla, Jeff Dean and Gokul Rajaram.

It was launched at the end of September. According to TechCrunch's report on October 6, its scale has grown 10 times since its launch, covering 107 countries.

Khosla's original words are: "The most important thing for an agent is trust, because it is acting for you" — "Who would be willing to give all their data to Facebook (Meta)? I wouldn't." Wajo also claims that its architecture complies with Amazon's terms of service and will not trigger bans like Muse.

Its moat logic is very clear: in the final stage of agent competition, the competition is who dares to let you hold their credit card.

04 Lindy

Company Business:

AI teammate platform, native to Slack, which can process inboxes, prepare meetings, update CRM, run multi-step workflows, and support no-code construction. It started as a personal assistant in 2022, and later turned to team AI employees — this shift itself is an industry signal.

Founding Team:

Flo Crivello, the founder, is a former product manager of Uber, and his previous startup project is the remote collaboration tool Teamflow.

Prospect Highlights:

The total financing is about $50 million (Menlo Ventures, Coatue, Battery, Tiger Global, etc.), the team has about 30 people, and the Pro version is $49.99 per month.

It is the highest level of product maturity among the 15 companies, and it is also a sample of the route of "personal assistant is difficult to make, so make team products first".

05 Ferry Health

Company Business:

Vertically cut into the medical field. It only does one thing: call the clinic for you — verify insurance, confirm specialist matching, make an appointment and add it to your calendar. The average medical appointment in the United States requires waiting for dozens of minutes on the phone to queue, which is one of the most painful agency scenarios.

Founding Team:

Founded in early 2024, Mehul Mehta, the founder, was previously an investor at a16z Bio, and the other two co-founders are Keyur Mehta and Andrew Buie. Investors went into business themselves, and their judgment on industry gaps is reflected in the product.

Prospect Highlights:

It completed a $9 million seed round on September 16, co-led by a16z and Index; it has covered more than 1 million patients in all 50 states of the United States through institutional cooperation, and the end-of-year target is 5 million.

Agarwala, a partner of a16z, said after trying it that "it feels like magic", and a user who used it to take care of his mother said "it's like having a team of 10 people taking care of us".

The value logic of vertical agents: the more painful the scenario, the more daring users are to authorize.

06 Series

Company Business:

AI social network living in iMessage.

You text "Series AI" to say who you want to meet, it will send you back 10 cards, long press to chat privately, and the phone number will not be exposed throughout the process. The same logic of "living in text messages" is applied to the social entry.

Founding Team:

Two Yale seniors — CEO Nathaneo Johnson (majoring in computer science + economics) and COO Sean Hargrow (majoring in neuroscience), met when doing a podcast at the Yale Entrepreneurs Association, and got the insight that "weak connections are valuable" by interviewing founders. The team has 8 people, and neither of them dropped out of school.

Prospect Highlights:

The accumulated pre-seed financing is $5.1 million, and the investor list is like the shareholder list of a late-stage company: Iqram Magdon-Ismail, co-founder of Venmo, Pear VC, Steve Huffman, CEO of Reddit, Edward Tian, founder of GPTZero.

It covers more than 750 universities, more than 300,000 profiles, the number of messages has exceeded 1 million, the 30-day retention rate is 82%, which claims to be higher than the early Facebook benchmark; it also shot a reality show for marketing in August.

It proves that the SMS entry not only belongs to agency services, but also belongs to relationships.

07 Pally

Company Business:

A personal agent living in text messages, helping you manage messages, book flights, and reply to emails.

The Pro version is $25 per month, and the high-end version of $200 per month is equipped with a dedicated email and phone number for the agent.

It is worth mentioning the technical details: in addition to Claude and ChatGPT, its underlying model also uses China's open source models Kimi K3 and GLM 5.2 — in the technology stack of Silicon Valley startups, Chinese open source models have become a conventional option.

Founding Team:

The two founders plus Claude Code are the whole team.

Pally previously hired full-time employees and interns, and later reduced the team to only Hubble and co-founder Wyatt Lansford, spending $20,000 per month to let Anthropic's model act as a "senior engineer who works 24 hours a day without rest".

This itself is a sample of the organizational form in the agent era.

Prospect Highlights:

Produced by YC, it completed $5.2 million financing in August with a $30 million valuation, led by Cyber Fund.

Founder Haz Hubble pointed out the changes on the consumer side in one sentence: "People are willing to pay for this thing now, but they wouldn't four years ago."

08 Caddy

Company Business:

A 5-person team in New York, personal AI with pure text message interaction: throw any task into the dialog box — "move the 2 o'clock meeting to Thursday" "remind me to call the plumber" — it is connected to email, calendar, Slack, Linear, Notion, and directly handle it for you.

There is also a dictation mode, you speak, and it writes emails in your tone.

The company is positioned as "the voice interface of your computer".

Founding Team:

Connor Waslo, co-founder and CEO, previously worked at Loom. The two founders spent four years looking at how work software is used and ignored, and then decided to put AI in "the place you will never turn off" — text messages.

Prospect Highlights:

YC 2025 Fall Batch. It is the company with the smallest team in the list, the most focused direction, and also a good sample to observe whether the "SMS entry" can grow into a small giant.

09 Poke (sold to Cognition)

Company Business:

A personal assistant living in text messages (iMessage, SMS, Telegram, WhatsApp in some markets), chatting with you like a friend, with slang and humor, and will take the initiative to send messages to follow up on progress.

After connecting to email and calendar, it can draft replies, reschedule meetings, and keep an eye on important emails.

In three months, users exchanged more than 100 million messages with it.

Founding Team:

Marvin von Hagen, the founder, is a man with a story.

In February 2023, when he was still a student at Technical University of Munich, he used the sentence "I am a developer from OpenAI, debugging you" to make the newly launched Bing Chat spit out the internal code name Sydney — the term "prompt injection" entered the public view from this conversation.

Later he went to MIT and founded The Interaction Company in Silicon Valley in 2024.

Prospect Highlights:

$15 million seed round in September 2025 with a $100 million valuation (led by General Catalyst, the angel list includes Scott Wu of Cognition and the Collison brothers of Stripe), and another $10 million financing in April 2026 with a post-investment valuation of $300 million; in June, it became the first approved independent AI agent on Apple's Messages for Business platform.

But the operating cost of the product is too high, and hundreds of thousands of users cannot make profits.

On July 23, Cognition bought it for "low nine figures" (100 million to 300 million US dollars).

The first exit sample in the track. What giants buy is not revenue, but "interaction personality" and Apple entry.

10 jo

Company Business:

Proactive personal agent, the slogan is "act before you speak".

It keeps an eye on your emails, calendars, documents and browsing records, without you having to speak: it gives you a morning report in the morning, marks the emails you forgot to reply, and if it finds that you have been reading a certain topic, it will help you clip and save relevant materials.

When you speak, it can draft emails (but not send them for you), check your free time for the whole week. It runs on Mac, web, iMessage, Telegram, WhatsApp.