210,000 annual fee Bloomberg terminal recreated by AI! Perplexity unveils a new "PC" in two months: integrates all AI functions, with Opus at the core to coordinate 19 models

Perplexity is also joining the fun, announcing its participation in the intense lobster battle.

Perplexity tweeted that it has launched a new product, Perplexity Computer, a browser-based, all-in-one digital assistant.

According to official statements, it can reason, delegate, search, build, remember, code, deliver, deploy projects, design documents, research topics, and manage workflows…

Almost covering all the processes needed in daily scenarios.

Perplexity CEO Aravind Srinivas posted a ??? indicating that Perplexity Computer is the result of the team’s secret two-month development.

Within just five hours of release, a user used Perplexity Computer to create a civilian version of a Bloomberg Terminal.

Although its features are far from the full capabilities of the original, the all-in-one workspace feel is already taking shape.

Friends!!

A Bloomberg Terminal is a professional financial information and trading system designed for finance professionals, used for market data, research, order placement, and communication.

It integrates real-time data, historical data, news, research reports, and analysis tools across global stock, bond, forex, and commodities markets, and is almost a standard for investment banks, brokerages, funds, and central banks.

Here’s the key—

In 2025-2026, the standard annual fee for a Bloomberg Terminal single user is about $25,000–$30,000, roughly $2,000–$2,665 per month, or about 14,500–19,321 yuan.

Perplexity Computer, on the other hand, only requires subscribing to Perplexity Max to use.

$2,000 per year is just 1/12 of the standard Bloomberg Terminal annual fee.

What is Perplexity Computer?

Currently, the AI agent market is mainly dominated by two players: OpenClaw, which controls the desktop, and Claude Cowork, focused on enterprise collaboration.

Perplexity Computer’s positioning is precisely between the two.

Officially, it is defined as an “agentic AI system”.

The core idea is “ability to act.”

Users only need to describe the desired outcome, and the system will automatically break down the task, generate execution steps, and then multiple sub-agents handle different parts, enabling cross-platform task execution.

In other words, once a task is issued, sub-agents quickly divide the work and start working.

One model might handle reasoning, another coding, another research, and yet another image generation.

The entire process can run continuously for hours, theoretically up to months.

Only when “really needed” will human intervention occur.

Currently, Perplexity Computer supports calls to 19 different models.

The core is Claude Opus 4.6, with access also to Gemini, Grok, ChatGPT 5.2, Nano Banana, and Veo 3.1, among others.

Reasoning, coding, web search, lightweight tasks, as well as image and video work.

See? This is what specialization looks like.

Additionally, Perplexity states that the system can interact with software just like a human using a computer:

• Access real browsers

• Access real file systems

• Access command-line tools

• External integrations

These interactions can be asynchronous, so if you’re busy with other things, you can let it run in the background.

The official claim is that you can run dozens of Perplexity Computers simultaneously.

What’s the difference between Perplexity Computer and OpenClaw?

Perplexity states about Perplexity Computer:

This is what a personal computer should look like in 2026.

(Eh, why isn’t Perplexity Computer considered a PC?)

But many friends and netizens are still a bit confused.

What exactly is the difference between this and OpenClaw? How does it differ from Claude Cowork? What about OpenAI and Gemini’s agents?

According to Perplexity, Perplexity Computer is safer and more controllable.

Take the recently deleted emails of Meta’s super-intelligence alignment director as a comparison with OpenClaw—

OpenClaw emphasizes direct control over the operating system, with access to local files, application manipulation, and even browser behavior.

Perplexity repeatedly emphasizes that Perplexity Computer runs in a browser “safe and reliable sandbox,” not touching the main system, so even if vulnerabilities occur, they won’t spread to the main network.

This is to prevent giving it too much freedom.

Clearly, Perplexity hopes to reassure users through browser-level isolation mechanisms.

But there are always those who love to stir up trouble.

Given Google’s restriction of OpenClaw’s high-frequency automated calls to Gemini, which was seen as malicious use and impacted platform stability, some netizens asked:

What if tomorrow the big tech giants ban Perplexity access too??

Besides architecture and security positioning, Perplexity also revealed some internal usage data.

The company says they have executed thousands of tasks with this system, from web content publishing to app development, with output quality exceeding expectations.

Of course, these statements are official, and the actual stability remains to be verified over time.

One More Thing

Perplexity Computer is currently only available to Max subscribers, who get 10,000 points per month.

It is understood that Perplexity Pro and Enterprise will soon launch this product as well.

Source: Quantum Bit

Risk Disclaimer and Terms of Use

Market risks are present; invest cautiously. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should consider whether any opinions, viewpoints, or conclusions herein are suitable for their particular circumstances. Invest at your own risk.