Timeline
- 1994Born on 7 June in Chennai, Tamil Nadu, into a family that valued mathematics and science.Both NDTV and Bharat Fast give the same date and city.
- 2010sEnters IIT Madras for a dual B.Tech and M.Tech in electrical engineering, having missed a place in computer science.While a student he taught courses on reinforcement learning.
- 2020Teaches Deep Unsupervised Learning at Berkeley in the spring semester, and again in 2021.
- 2021Completes a PhD in computer science at the University of California, Berkeley, working on contrastive learning, reinforcement learning and generative vision models.
- 2010s to 2022Works at OpenAI on reinforcement learning, at DeepMind in London on contrastive learning, and at Google on vision models including HaloNet and ResNet-RS.
- 2022Returns to OpenAI as a research scientist and contributes to DALL-E 2, then leaves to co-found Perplexity AI on 22 August with Denis Yarats, Johnny Ho and Andy Konwinski.
- 2022Perplexity launches its search engine on 7 December, answering questions with a short synthesis and a list of sources.
- 2023Begins angel investing in AI startups from January, with early positions in ElevenLabs and Suno.
- 2024Perplexity raises $165 million by April and crosses a billion dollar valuation, then closes a $500 million round in June 2025 at $14 billion.
- 2025The company is valued at $20 billion in September, submits a merger proposal for TikTok's US business in January, and bids $34.5 billion for Google Chrome in August.
- 2025Enters the M3M Hurun India Rich List as the country's youngest billionaire at 31, with an estimated net worth of Rs 21,190 crore.Reported by NDTV on 1 October 2025.
- 2026Perplexity reaches a $21.21 billion valuation after a Series E-6 round, signs a three year $750 million Microsoft Azure agreement in January, and moves to a subscription first model in February.
Who Aravind Srinivas is
Aravind Srinivas is the co-founder and chief executive of Perplexity AI, a company that answers questions on the web rather than returning a list of links to them. He was born in Chennai in 1994, trained as an electrical engineer at IIT Madras, and spent six years in research across Berkeley, OpenAI, DeepMind and Google before starting the company in August 2022.[2][3]
His public reputation rests on a single argument that he has repeated in almost every interview: a search engine built for a language model should return a short answer with the sources attached, not ten blue links that a reader has to open one by one. Perplexity was built to that specification, and the design choice turned out to be the company's identity. Every answer carries numbered citations, and a reader who does not trust the summary can check the page behind it.[3][4]
The other half of his reputation is financial. Perplexity went from a research project to a company valued at more than twenty billion dollars in about three and a half years, and the speed of that rise made Srinivas a subject of coverage that has little to do with computer science. In October 2025 he appeared on the Hurun India rich list as the country's youngest billionaire, an entry built on paper wealth in a private company rather than on anything he has sold.[2]
Chennai, IIT Madras and the electrical engineering detour
Srinivas was born and raised in Chennai. Accounts of his childhood are thin, but the one detail that recurs is an early interest in science that his school and family encouraged, in a city with a deep engineering culture and a large supply of competitive exam coaching.[2]
He went to IIT Madras, and the subject he studied there is the part of his biography that gets left out of most profiles. He wanted computer science and did not get a seat, so he took electrical engineering instead, completing both the B.Tech and the M.Tech in a dual degree. The detour did not keep him out of the field. He taught courses on reinforcement learning and advanced reinforcement learning while still a student, which is unusual for an undergraduate and suggests that his interests had already moved past his enrolment.[1][2]
The episode is worth keeping in the record because he has referred to it himself as the moment that pushed him toward the software side of the discipline. An electrical engineering degree at IIT Madras includes enough mathematics and signal processing to make the transition into machine learning a matter of adding coursework rather than starting again, and Srinivas added it quickly enough to be teaching it.[1]
What is not in the record is which school in Chennai he attended, what his parents did, or how the family financed an American doctorate. Bharat Fast's profile fills those gaps with a father described as a mathematics educator and a mother described as a social worker, but no interview, newspaper profile or institutional biography carries those names, so this entry does not treat them as established.[1][2]
Berkeley and the research years
Srinivas moved to the University of California, Berkeley for a PhD in computer science and completed it in 2021. His dissertation work sat in the middle of the generative AI boom that was then still academic: contrastive learning for computer vision, reinforcement learning, and transformer based models for image generation, image recognition and video generation.[2]
The teaching record from those years is more informative than the thesis title. He taught Deep Unsupervised Learning in the spring semesters of 2020 and 2021, a course that Berkeley runs at the sharp end of the field and that changes content every year because the field does. Standing in front of that class is a different skill from writing a paper, and it is one of the reasons his later public communication about AI is clearer than that of most founders in the sector.[2]
Berkeley in that period was also a network rather than a building. The people who worked on the same problems in the same labs went on to staff the research teams at OpenAI, Google, Anthropic and a long list of startups, and Perplexity's founding team came out of that network. Denis Yarats, one of the co-founders, was a Berkeley colleague working on reinforcement learning.[3]
Four labs and a credit on DALL-E 2
Before founding Perplexity, Srinivas worked at OpenAI, DeepMind and Google, and the sequence matters more than the list. He started at OpenAI on reinforcement learning, moved to DeepMind in London to work on contrastive learning, and then spent time at Google developing vision models including HaloNet and ResNet-RS. He returned to OpenAI as a research scientist and contributed to DALL-E 2, the text to image model that made generated pictures a mainstream subject in 2022.[2]
That range covers the three technical traditions that the current generation of AI products rests on: reinforcement learning for decision making, contrastive learning for connecting images and text, and large scale vision architectures for seeing. A founder who has worked inside all three is rare, and it explains why Perplexity's early product decisions were made by someone who understood what the underlying models could and could not do.[2][3]
The lab years also gave him a view of how the big companies were likely to move. Perplexity's later bids for TikTok's American business and for Google Chrome were widely read as publicity, but they came from a founder who had watched how platform owners treat the products that sit on top of them.[3]
Since January 2023 he has also been an angel investor, with early money in ElevenLabs, the voice synthesis company, and Suno, which generates songs from a text prompt. Both are consumer products built on models that did not exist when he was in graduate school, and both sit in the same category as Perplexity: an interface that hides a model behind a simple request.[2]
Founding Perplexity in August 2022
Perplexity AI was incorporated on 22 August 2022. The founding team is usually given as four people: Srinivas, Denis Yarats, Johnny Ho and Andy Konwinski, engineers with backgrounds in back end systems, artificial intelligence and machine learning. NDTV's profile names only Yarats and Konwinski alongside him, which is a common simplification in Indian coverage rather than a different founding story.[2][3]
The product launched its search engine on 7 December 2022. The timing was not accidental. ChatGPT had gone public at the end of November, the idea of asking a machine a question in plain language had just become familiar to a mass audience, and Perplexity offered the same conversational experience with one addition that the chatbot did not have: a visible set of sources under every answer.[3]
By February 2023 the site had two million unique visitors. That number is small by the standards of the later valuation stories, but it was reached without a marketing budget, by a team of engineers shipping a product that people told each other about. The growth curve from there is what attracted the investors who would eventually price the company in the tens of billions.[3]
Srinivas has described the founding decision in interviews as a bet on a gap rather than on a technology. The models were available to anyone with a contract; what was missing was an interface that treated the model as a reader of the web rather than a database of things it had memorised. Perplexity's answer was to search first, then summarise, then cite.[3][4]
The answer engine and why the citations matter
Perplexity calls itself an answer engine. A user types a question, the system searches the live web, reads the pages it finds, and writes a response with numbered references. A free tier is available without registration, and a Pro subscription adds stronger models and more searches. The company also sells an API and an enterprise product, and it has released its own models, Sonar, built on Llama, and R1 1776, built on DeepSeek R1.[3]
The citation design is the part of the product with the clearest public argument behind it. A language model that answers from memory cannot tell a reader where a claim came from, and it cannot be corrected when the underlying fact changes. A model that answers from pages can do both. Srinivas has put the point in almost the same words in several interviews: an index built for humans assumes someone will click a link, while an index built for a model has to deliver grounded, real time knowledge that can be checked.[4]
The company has expanded from the answer box into adjacent products, including a Chrome extension, mobile apps and Comet, its own browser. The browser is the most ambitious of these because it competes directly with the surface that Google controls, and it made the later bid for Chrome legible as strategy rather than as a stunt.[3][4]
The model choice inside Perplexity has always been plural. The Pro tier lets a subscriber pick between back end models from several vendors rather than committing to one, which is a quiet statement about where the company thinks the value sits: in the retrieval and the interface, not in owning the largest model.[3]
Funding, valuation and the shape of the rise
Perplexity raised $165 million by April 2024, at a valuation above a billion dollars. In June 2025 it closed a $500 million round that put the valuation at $14 billion. By September 2025 the figure was $20 billion, and a Series E-6 round in early 2026 took it to $21.21 billion. Each of those steps happened within months of the previous one, which is why the valuation is often quoted with a date attached.[3][4]
The investor list is unusual for a company of its age. Jeff Bezos, Tobias Lutke, Nat Friedman, Nvidia and Databricks have all put money in, and 1789 Capital, a fund associated with Donald Trump Jr, is also an investor. Nvidia's participation is the one worth noting, because a chip company investing in an application layer is a statement about where it expects demand to come from.[3]
In January 2026 Perplexity signed a three year agreement worth $750 million to use Microsoft's Azure cloud and its Foundry service, which gives it access to models from OpenAI, Anthropic and xAI. Amazon Web Services remained the main cloud provider, so the deal added a second supplier rather than replacing the first. A company that depends on other people's models needs more than one supplier, and the Microsoft agreement bought that redundancy at a fixed price.[3]
In February 2026 the company moved to a subscription first model and discontinued the advertising strategy it had been testing. Management framed the change as a defence of user trust: an answer engine that also sells placement has a reason to prefer the advertiser's page, and the product's whole claim is that it has no such reason.[3]
The Hurun listing and what it does and does not mean
On 1 October 2025 NDTV reported that Srinivas had entered the M3M Hurun India Rich List as the country's youngest billionaire, at 31, with an estimated net worth of Rs 21,190 crore. The figure was calculated from his stake in Perplexity against the company's then valuation, not from cash, salary or realised gains.[2]
That distinction is the whole story. A founder's paper wealth moves with the last round, and Perplexity's last round moved several times in eighteen months. The Hurun entry is a snapshot of a private valuation applied to an equity stake whose size has never been disclosed, which is why the number should be read as an estimate produced by a list rather than as a disclosed fact.[2][3]
What the listing did do is change how he is covered in India. Before October 2025, the Indian coverage was about a Chennai born founder running an American AI company. After it, he was a billionaire, and the interviews shifted toward Indian policy, Indian hiring and what the country should do to produce more companies of the same kind.[2][4]
The bids that did not happen
Perplexity has made two public bids that came to nothing, and both are part of how the company is understood. On 18 January 2025, the day before the American ban on TikTok took effect, it submitted a proposal to merge with TikTok's US business. In August 2025 it offered $34.5 billion for Google Chrome, arguing that a sale would remedy the antitrust case the search giant was facing.[3]
Neither bid succeeded, and it is reasonable to treat both as partly strategic communication. A company with two hundred million dollars of revenue does not buy Chrome, and it does not merge with TikTok. What the bids did was place Perplexity in the same sentence as the platforms it wants to displace, in front of an audience that included regulators, investors and the press.[3]
The Chrome bid is the one that aged best. Perplexity had already built Comet, its own browser, and Google was in the middle of adding AI features to Chrome. The offer put a price on the asset everyone in the industry was arguing about, and it made the argument that the browser, not the search box, is the real point of control.[3][4]
In December 2025 Cristiano Ronaldo took an undisclosed stake in the company and signed a global brand partnership with it. The deal is the clearest sign that Perplexity had moved from a tool for technology workers into consumer marketing, where the currency is a face rather than a benchmark.[3]
India, the largest market
By late 2025 India had become Perplexity's largest market by user base, a fact Srinivas gave to The Economic Times in an interview at the company's San Francisco office. The growth followed a partnership with Airtel that bundled Perplexity Pro with the telecom operator's plans, a carrier strategy the company had already run in Germany, Japan, South Korea and Spain. After the Airtel launch the app reached number one on the Indian iOS App Store.[4]
The reason India works, in his account, is language rather than price. A large English reading population gives the retrieval system a deep set of pages to work with, and Perplexity's product is only as good as the material it can find. Most companies assume English usage means the United States, Britain or Australia; Srinivas has argued that Indian English usage is large enough to be central to the product's development.[4]
He has also moved the company's attention beyond the product. Perplexity intends to set up an Indian fund and is considering a policy presence in Delhi and an engineering office, with Bengaluru and Hyderabad the two candidates. Partnerships in travel, shopping, education and healthcare are on the table, partly because Indian consumer infrastructure, UPI in particular, allows agentic products that have no direct equivalent in the West.[4]
On policy he has been careful to describe his role as informal. He says he is not an official adviser but shares input when asked, including with the prime minister and some chief ministers, and he has argued for government owned GPU capacity allocated to universities and startups, and for a sovereign fund backing Indian companies that build fundamental technology.[4]
There is a personal argument underneath the commercial one. In the same interview he described growing up with a mother who told him to get a job at Google, and said that Sundar Pichai's rise proved an Indian could run a global company but also reinforced the idea that Indians run other people's companies. His stated ambition for Perplexity is that the next generation sees building the next Google as a realistic option.[4]
Copyright, crawlers and the criticism
Perplexity's model, which reads and summarises pages it does not own, has produced legal and reputational problems that no amount of funding can settle. The company has faced allegations of copyright infringement, unauthorised use of content and trademark issues from several major media organisations, including the BBC, Dow Jones and The New York Times.[3]
Two investigations added a more specific charge. Wired, and later Cloudflare, reported that Perplexity was using undisclosed crawlers with spoofed user agent strings to reach sites that had explicitly blocked automated scraping. The company has disputed the characterisation, and the dispute is unresolved.[3]
For a founder whose product is built on citation, the accusation is uncomfortable in a particular way. A citation is an invitation to check the source, which is a defence against the charge of hiding where information came from; it is not a defence against the charge of taking the information without permission. Perplexity's answer has been a publishers' programme, announced in July 2024, that shares advertising revenue with partners, and the subscription first switch in 2026 removed the advertising that programme depended on. Whether the arrangement satisfies publishers is still an open question.[3]
The company's position in the wider industry is also contested. Its searches are run on models built by other companies and its index competes with Google's, so it sits between two much larger players, dependent on one for capability and in conflict with the other for traffic. That position is profitable while the model vendors need application layer demand, and it is exposed if they decide to keep the consumer relationship for themselves.[3]
What has not been confirmed
There is no Wikipedia biography of Srinivas. The English Wikipedia entry for his name redirects to the article on Perplexity AI, which means the personal detail in circulation comes from interviews and from profile sites rather than from an encyclopaedia that has been argued over by editors. The date of birth, 7 June 1994, and the birth city, Chennai, are consistent across NDTV and Bharat Fast, but they are not independently confirmed by a primary record.[1][2][3]
Bharat Fast names his parents and describes their professions. No interview, newspaper profile or company biography that this entry could find carries those names, and the same page also gives his degree as a dual B.Tech and M.Tech in electrical engineering, which is plausible and consistent with his own account of missing computer science, but is not confirmed by a second source. None of it is presented here as fact.[1][2]
The size of his stake in Perplexity has never been disclosed, so every net worth figure, including the Rs 21,190 crore estimate in the Hurun list, is a calculation by an outside body against a private valuation. His salary, his personal investments beyond the two angel cheques that have been reported, and his property holdings are not public.[2][3]
The founding team is given as four people in most accounts and as three in NDTV's profile, which omits Johnny Ho. This entry follows the four name version because it is the one used by the company and by Wikipedia, and notes the discrepancy rather than resolving it.[2][3]
His height, weight, marital status and family are not published by him, and the biography sites that carry numbers for them do not cite anything. This entry does not repeat them.[1]
Questions about this entry
Who is Aravind Srinivas?
Aravind Srinivas grew up in Chennai, studied electrical engineering at IIT Madras because computer science was full, and finished a PhD at Berkeley in 2021.
What is Aravind Srinivas known for?
Aravind Srinivas grew up in Chennai, studied electrical engineering at IIT Madras because computer science was full, and finished a PhD at Berkeley in 2021.
When was Aravind Srinivas born?
Aravind Srinivas was born in 7 June 1994, Chennai.
How current is this entry?
It was last updated on 8 November 2026, carries 1 revisions and cites 4 sources. The entry has been checked by the editorial team and carries the verified badge.