Ex-Twitter CEO Parag Agrawal Launches AI Startup Parallel Web Systems with $30M Funding

Parallel Web Systems Inc., a cloud platform designed to assist artificial intelligence systems in carrying out extensive web research, was introduced by former Twitter CEO Parag Agrawal. After taking over Twitter (now known as X) in 2022, billionaire Elon Musk fired Agrawal.

Parallel Web Systems: The AI Startup Redefining Research

He launched Parallel in 2023 and has since gathered a staff of 25 people in Palo Alto. The startup has already raised $30 million from well-known investors, such as Index Ventures, Khosla Ventures, and First Round Capital. We already power millions of research jobs every day across ambitious startups and public companies, according to Agarwal’s LinkedIn post.

He went on to say that some of the AI startups with the quickest rates of growth use Parallel to integrate web intelligence into their agents and platforms. With Parallel, a publicly traded firm automates processes that are typically performed by humans with precision that surpasses that of humans. Our search is used by coding agents to locate documents and troubleshoot problems.

Deep Research API: Beating GPT-5 and Humans

Additionally, Parallel announced the release of its Deep Research API, which, according to Agrawal, is the first to beat top models, such as GPT-5, and humans on two of the most difficult benchmarks. The web is humanity’s memories, the business wrote in a blog post.

Large-scale publication, education, and cooperation have been made possible by the open internet. It served as the basis for training contemporary AI. AI is becoming the main user of the web. AIs can search through whole databases, digest information for hours, or extract facts instantaneously on a large scale, in contrast to humans who just look at a few pages or perform quick searches. Antiquated company models are ineffective.

Clicks, advertisements, paywalls, and locked APIs are examples of how the modern web depends on human attention. However, these models run the risk of enclosing important knowledge in silos and are not machine-use-friendly. An AI “Programmatic Web” Parallel contends that the internet needs to develop into a machine-designed system that facilitates computing, reasoning, and reputable sources.

How Parallel Works: AI-Specific Internet for the Future

By integrating data, computation, and reasoning on a single infrastructure, Parallel generates insights and actions rather than only static texts. Additionally, it functions on declarative interfaces, in which AIs specify their needs and the system determines how to provide them.

Furthermore, it offers clear attribution, allowing contributions to be quantified and all sources to be acknowledged. Likewise, Parallel strives to increase transparency in free marketplaces where value is monetarily rewarded, guaranteeing that openness is supported by both financial incentives and goodwill.

Parallel is developing an AI-specific version of the internet. AIs will be able to directly request information, and Parallel’s system will collect, process, and arrange it in place of people clicking and searching. Additionally, it will reward contributors and acknowledge sources. To put it briefly, Parallel makes it simpler and more equitable for AIs to access the internet to generate insights and find answers.