Who Created Perplexity AI?

The Founders of Perplexity AI

Perplexity AI was founded in August 2022 by four computer scientists and engineers: Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski Perplexity AI Who Owns Perplexity AI? Owners and Investors 2026. The team brought together deep expertise from leading artificial intelligence research labs, high-frequency quantitative engineering, and enterprise open-source software scaling Perplexity AI Aravind Srinivas, Perplexity | AI at Princeton Denis Yarats.

Headquartered in San Francisco, California, the company was established to build an "answer engine"—a conversational search platform that provides concise, synthesized answers to natural language questions supported by cited web sources, rather than the traditional list of ranked links common to conventional search engines Perplexity AI Beyond NYU: From research at Courant to taking on tech ....

                               Perplexity AI Founding Team (Aug 2022)
                                                 │
        ┌────────────────────────┬───────────────┴───────────────┬────────────────────────┐
        ▼                        ▼                               ▼                        ▼
 Aravind Srinivas          Denis Yarats                     Johnny Ho              Andy Konwinski
  Chief Executive         Chief Technology               Chief Strategy             Co-Founder &
      Officer                 Officer                        Officer               Board Director
(OpenAI, Google, UC     (Meta AI, NYU, Quora,         (Tower Research, Jane      (Databricks Co-founder,
Berkeley PhD Scholar)    Courant Institute)            Street, Harvard Math)      UC Berkeley AMPLab)

Founding Team Profiles

Each of the four co-founders brought distinct technical and operational capabilities that allowed the startup to quickly prototype, optimize, and scale an infrastructure-heavy product Perplexity AI Who Owns Perplexity AI? Owners and Investors 2026.

Aravind Srinivas (Chief Executive Officer)

Aravind Srinivas serves as the Chief Executive Officer and public face of Perplexity AI Who Owns Perplexity AI? Owners and Investors 2026 Who Is Perplexity's CEO? Aravind Srinivas Profile - Honeyb. Srinivas earned an undergraduate degree in electrical engineering from the Indian Institute of Technology (IIT) Madras before completing a Ph.D. in computer science at the University of California, Berkeley Aravind Srinivas, Perplexity | AI at Princeton Who Is Perplexity's CEO? Aravind Srinivas Profile - Honeyb. His doctoral research concentrated on deep learning, computer vision, and reinforcement learning Aravind Srinivas, Perplexity | AI at Princeton.

Prior to co-founding Perplexity, Srinivas worked in frontier artificial intelligence research labs:

• OpenAI: Conducted research on generative modeling, contrastive representation learning, and language models Aravind Srinivas, Perplexity | AI at Princeton.
• DeepMind & Google Brain: Researched reinforcement learning and representation architectures Aravind Srinivas, Perplexity | AI at Princeton.

His academic background gave him early insight into the capabilities and structural limitations of large language models (LLMs)—most notably their propensity to hallucinate facts without external verification mechanisms Aravind Srinivas, Perplexity | AI at Princeton Beyond NYU: From research at Courant to taking on tech ....

Denis Yarats (Chief Technology Officer)

Denis Yarats serves as Chief Technology Officer, overseeing Perplexity's technical infrastructure, algorithm development, and retrieval pipelines Who Owns Perplexity AI? Owners and Investors 2026 Denis Yarats. Yarats earned his master's degree in computer science from Saint Petersburg State University and later completed a Ph.D. in machine learning at New York University’s Courant Institute of Mathematical Sciences, where his research concentrated on sample-efficient reinforcement learning Denis Yarats Beyond NYU: From research at Courant to taking on tech ....

Before co-founding the company, Yarats held several senior machine-learning engineering and research roles:

• Meta (FAIR - Fundamental AI Research): AI research scientist specializing in reinforcement learning and machine learning systems Denis Yarats Perplexity & the Future of AI with Denis Yarats, Co-Founder ....
• Quora: Machine learning engineer focused on content recommendation, search, and feed optimization Denis Yarats.
• Microsoft: Software engineer working on core platform services Denis Yarats.

Yarats's expertise in combining retrieval mechanisms with neural generation models formed the core architecture of Perplexity's search pipeline Denis Yarats Beyond NYU: From research at Courant to taking on tech ....

Johnny Ho (Chief Strategy Officer)

Johnny Ho serves as Perplexity's Chief Strategy Officer, initially guiding engineering execution before directing strategic technical initiatives Who Owns Perplexity AI? Owners and Investors 2026. A former competitive programmer and International Olympiad in Informatics (IOI) medalist, Ho studied mathematics and computer science at Harvard University.

Prior to Perplexity, Ho worked in the quantitative trading sector as an algorithmic quantitative engineer at firms such as Tower Research Capital and Jane Street. His background in low-latency systems and mathematical optimization proved critical in engineering Perplexity's real-time retrieval-augmented infrastructure, where sub-second latency is required to make conversational search practical.

Andy Konwinski (Co-Founder & Board Director)

Andy Konwinski is a veteran systems researcher and technology entrepreneur Perplexity AI Who Owns Perplexity AI? Owners and Investors 2026. He earned his Ph.D. in computer science from UC Berkeley, where he was part of the celebrated AMPLab (Algorithms, Machines, and People Lab). Konwinski co-created Apache Mesos and was an early contributor to Apache Spark.

Konwinski previously co-founded Databricks, the data analytics and AI infrastructure company, as well as the open-source software accelerator Perk.com. At Perplexity, Konwinski provided early executive guidance, enterprise strategy, and distributed computing architecture support Perplexity AI Who Owns Perplexity AI? Owners and Investors 2026.

Origin and the Shift to Search

The concept of Perplexity emerged from the founders' friction with existing information-retrieval systems Beyond NYU: From research at Courant to taking on tech .... In early 2022, language models demonstrated advanced reasoning and text synthesis, but commercial search engines remained centered around returning ad-supported links requiring users to click through multiple tabs Perplexity AI Beyond NYU: From research at Courant to taking on tech .... Conversely, raw generative language models suffered from "hallucination"—generating factually incorrect assertions with high confidence due to the lack of grounding in real-time information Beyond NYU: From research at Courant to taking on tech ....

The founders initially experimented with natural-language text-to-SQL interfaces and database tooling under the Perplexity banner in August 2022 Perplexity AI. However, they pivoted quickly toward building a web-grounded conversational search engine, releasing their first prototype, Bird SQL (a tool for querying Twitter data using natural language), followed by the first iteration of the Perplexity Ask search product in December 2022 Perplexity AI.

The company chose the name Perplexity from the fundamental metric in information theory and natural language processing:

In computational linguistics, perplexity measures how well a probability model predicts a sample; a lower perplexity score indicates a model that is less "surprised" by the text and more accurate in its predictions. The founders adopted the name as an ironic nod to clarifying confusion and reducing user uncertainty when seeking knowledge.

Technical Architecture: How Perplexity Was Built

Perplexity differs structurally from both pure search indexers (such as Google or Bing) and pure LLM chat interfaces (such as baseline ChatGPT). The founders designed the system using an orchestration architecture known as Retrieval-Augmented Generation (RAG) Denis Yarats Beyond NYU: From research at Courant to taking on tech ....

By combining real-time search indexing with dynamic inference synthesis, Perplexity ensures that the output is informed by up-to-date web data while maintaining traceability to the underlying publisher or domain Perplexity AI Beyond NYU: From research at Courant to taking on tech ....

Corporate Structure, Funding, and Growth

Perplexity AI operates as an independent, venture-backed enterprise registered as Perplexity AI, Inc. in Delaware Perplexity AI. While the four co-founders guide operational and technical leadership, the company has raised capital across multiple financing rounds from notable venture capital firms, technology leaders, and angel investors Perplexity AI Who Owns Perplexity AI? Owners and Investors 2026:

• Venture Capital Firms: New Enterprise Associates (NEA), Bessemer Venture Partners, IVP, and Institutional Venture Partners Perplexity AI.
• Strategic Corporate Investors: Nvidia, Databricks Ventures, and SoftBank Vision Fund 2 Perplexity AI.
• Prominent Angel Investors: Jeff Bezos (via Bezos Expeditions), Nat Friedman (former CEO of GitHub), Elad Gil, Naval Ravikant, and Yann LeCun Perplexity AI Denis Yarats.

The company generates revenue through consumer subscription tiers (Perplexity Pro), API credits for developers, enterprise search solutions (Perplexity Enterprise Pro), and sponsored follow-up query models Perplexity AI Who Owns Perplexity AI? Owners and Investors 2026. Through this hybrid technical foundation and capital backing, Perplexity developed from a four-person research startup into one of the most prominent independent platforms challenging traditional search paradigms Perplexity AI Beyond NYU: From research at Courant to taking on tech ....

I'll research the founding and current status of Perplexity AI before writing.Now let me check the company's current status, funding, and products.## Who Created Perplexity AI

Perplexity AI was created by four engineers and researchers who incorporated the company in San Francisco in August 2022: Aravind Srinivas, Denis Yarats, Johnny Ho, and Andy Konwinski. Srinivas is the chief executive and the public face of the company, which is why he is often named as the founder, but the product was a four-person effort from the start. The company is based in San Francisco, California, and was founded in 2022 by Aravind Srinivas, Andy Konwinski, Denis Yarats and Johnny Ho. One account dates the founding specifically to 22 August 2022, with the founding group comprising Srinivas, Yarats, Ho, and Konwinski. The public launch of the answer engine followed on 7 December 2022.

If you are asking who makes or who develops Perplexity today, the short answer is Perplexity AI, Inc. — a private, venture-backed company that is independent of Google, OpenAI, Microsoft, and Nvidia, even though several of those firms appear in its investor list or its supply chain. That distinction matters more than it first appears, and it is explained further below.

The four co-founders and what each brought

The founding team is unusual in that none of its members came from a traditional search background; they came out of frontier AI research labs, infrastructure engineering, and competitive programming.

Public summaries of the team describe Srinivas as a former researcher at OpenAI, Google, and DeepMind; Yarats as a former Meta AI researcher; Ho as a former Quora engineer; and Konwinski as a Databricks co-founder. Reporting on the team adds that Ho moved from Quora into quantitative trading on Wall Street before joining the company. Ho's trading stint is elsewhere associated with Tower Research Capital, and Konwinski is listed as president and a board member.

Srinivas's path is the most widely documented. He completed his bachelor's and master's degrees at IIT Madras before a PhD in computer science at UC Berkeley, and co-founded Perplexity afterward. Before founding the company he did machine learning research during internships at Google and DeepMind — contributing to vision architecture work including HaloNet and ResNet-RS — followed by a research scientist appointment at OpenAI. His research positions at OpenAI and DeepMind directly preceded the 2022 founding.

The way the team came together is a useful corrective to the "lone founder" framing. By Srinivas's own telling, he recruited Johnny Ho — a standout competitive programmer — and Denis Yarats, a fellow Berkeley researcher who brought deep AI engineering experience. Konwinski's prior experience building Databricks, a company that commercialised research-born infrastructure, gave the group someone who had already navigated the path from lab to enterprise product.

Why they built it: the "answer engine" premise

Perplexity was not conceived as a chatbot or as a general-purpose assistant. It was built around a narrower thesis: that a large language model should answer a question directly and show you where each claim came from, rather than returning a ranked list of links for you to read yourself.

The founding idea was a system that could understand the context of a question, give a direct answer, and cite its sources — integrating large language models into the search process itself to produce conversational, context-aware results. Srinivas has described this as a "citation-first" approach, aimed at verifiable answers rather than traditional search links. The company describes its core product as an answer engine that returns direct, cited answers instead of a list of links.

This is the key architectural decision behind nearly everything else the company has done. Because the product's value proposition is verifiability, Perplexity leans on retrieval — fetching live web content at query time — rather than relying only on what a model memorised during training. That design choice also created the company's most persistent business tension: it depends on publisher content it does not own.

Who develops and operates it now

Perplexity AI, Inc. remains a private company run by its founding CEO, and the engineering organisation it has built since 2022 develops the products. But "who develops Perplexity" has a second, more technical answer that confuses many users: Perplexity does not train most of the underlying foundation models it serves.

Instead it combines its own search-tuned models with third-party ones. The company pairs its in-house, search-tuned Sonar models with a multi-model picker offering models from OpenAI, Anthropic, Google, and xAI, and monetises through subscriptions and enterprise rather than advertising. Its surface area has grown from a single website into iOS, Android, Windows and macOS apps, the Comet browser, an enterprise product, and the Sonar developer API.

So a single Perplexity answer may involve work from several organisations at once: Perplexity's retrieval stack and ranking, a frontier model licensed from another lab, and content from the publishers it cites. When people ask who "made" the answer they received, this layered reality is usually the honest response. Model availability on any given plan also changes frequently, so specific model names should be checked against the current product rather than taken from any written source.

The product line has expanded well beyond search. Perplexity's Comet browser became free worldwide in October 2025, and a Computer agent product launched in February 2026 added usage-based revenue on top of subscriptions. Comet is positioned as an agentic browsing surface, with a Comet Plus tier intended as a response to the publisher-compensation problem that agentic browsing creates. On the publisher side, the company introduced a Publishers Program to share revenue with outlets whose work the engine cites.

Ownership, funding, and why valuation figures disagree

Perplexity is venture-funded and not publicly traded, so no single shareholder "owns" it in the way a parent company would. Reported investors have included Nvidia, Jeff Bezos, SoftBank, IVP, Accel, and Databricks, among others. Total funding has been reported as exceeding $1.5 billion from that investor group. Early backers are commonly listed as including Elad Gil and Nat Friedman.

Published valuation figures vary noticeably, and it is worth understanding why rather than trusting any one number. Private-company valuations come from different rounds, from primary versus secondary transactions, and sometimes from unconfirmed reporting. One timeline puts the company at roughly $121 million in April 2023, about $9 billion by late 2024, $18 billion in July 2025, and about $20 billion in September 2025. Another account describes a Series E-6 closing on 9 January 2026 at roughly $23 billion, and notes that as of late August 2026 Nvidia was reportedly in talks to invest at a valuation above $30 billion — a round that had not closed. A third summary gives the January 2026 Series E-6 valuation as $22.6 billion, with total funding near $1.72 billion after a $200 million raise reported on 5 June 2026.

These accounts are not all reconcilable, and some secondary trackers restate each other rather than reporting independently. The defensible summary is that Perplexity has been valued in the low tens of billions of dollars since late 2025, with higher figures reported but not confirmed. As of mid-2026 no IPO had been announced, with no known S-1 filing and no stated listing timeline.

Common confusions worth clearing up

A few recurring misunderstandings cluster around this question:

• "Aravind Srinivas founded it alone." He is the CEO and the most visible founder, and much coverage reduces the story to him. Contemporary accounts consistently list him alongside Denis Yarats as CTO, plus Andy Konwinski and Johnny Ho.
• "It's an OpenAI or Google product." It is not. Srinivas previously worked at OpenAI, Google, and DeepMind, and Perplexity licenses third-party models — but the company is independently incorporated and independently run.
• "Perplexity AI" versus "perplexity" the metric. In natural language processing, perplexity is a long-established measure of how well a probability model predicts a sample — lower is better. For a model assigning probability $p(x_i)$ to each token in a sequence of length $N$, perplexity is:

The company borrowed this term; nobody "created" the metric as a product, and searches mixing the two concepts are genuinely ambiguous.

• Founding date precision. Sources agree on August 2022 for incorporation but differ on whether to date the company's "creation" to that month, to a specific day in it, or to the December 2022 public launch. All three dates appear in circulation, and none is wrong so much as answering a slightly different question.

How to verify founder claims yourself

Because founder attributions get garbled as they are recycled across aggregator sites, it helps to know which evidence actually settles the question. Incorporation records and the company's own about/team pages are primary. Reputable business press coverage of specific funding rounds typically names the executive team and is dated, which makes it checkable. Profiles that list founders without naming a round, a filing, or an interview should be treated as derivative. For any claim about current leadership, roles, product capabilities, model lineups, or valuation, check the company's current materials — executive titles in particular shift at fast-growing private companies, and several of the figures above carry explicit uncertainty from the sources that reported them. Even Srinivas's own title is reported inconsistently, with some alumni citations calling him President and CEO while major outlets simply use CEO.