Jonathan Ross

A Forbes article published in August 2024 gave his age as 42.⁠[1]

Fortune wrote in August 2024 that he dropped out of high school because he was bored and learned computer programming.⁠[2]

In a conversation with Chamath PalihapitiyaChamath PalihapitiyaChamath Palihapitiya is an investor and business operator who founded and leads Social Capital, and is also co-founder and CEO of the AI software company 8090 and a co-host of the All-In podcast. He previously ran user growth at Facebook and took part in taking several companies public through the Social Capital Hedosophia series of special purpose acquisition companies; he is also an early investor in and director of Beast Industries.Open the full entry that Social CapitalSocial CapitalSocial Capital is a US technology investment firm founded in 2011 and led by Chamath Palihapitiya. Its activities evolved from venture funds toward investing and incubation with permanent proprietary capital.Open the full entry published on April 16, 2024, he said that after dropping out of high school he got a job as a programmer and that his boss, noticing that he was clever, urged him to take classes at a university; asked whether he has a high school diploma, he answered no, and added that he does not have an undergraduate degree either. Chamath Palihapitiya’s article of February 2024 calls him a former high-school dropout.⁠[3][3][4]

In a 2026 interview he said that he has not had a driver’s license since he was 18, because he does not feel the need to drive and would rather spend the time thinking and doing something useful; he said this has been the case since he was 19.⁠[5][5]

By his own account, he began going to Hunter College on the side without being formally enrolled, came under the wing of one of the professors there, did well, and then transferred to New York UniversityNew York UniversityNew York University is a private university in New York City, founded in 1831, and one of the largest private universities in the United States.Open the full entry.⁠[3][3]

Fortune wrote that after a stint at Hunter College he got into New York University, took Ph.D. classes as an undergraduate for two years, and then dropped out; he joked, “I didn’t want to cap my earning opportunity by graduating from something.”⁠[2]

The biography attached to an EE Times interview reposted by Groq states that he studied mathematics and computer science at New York University’s Courant Institute and that, in his second year, he became the first computer science undergraduate to complete courses restricted to Ph.D. students.⁠[6][7]

Groq’s leadership page states that he studied under Yann LeCun of the Courant Institute. ForbesForbesForbes is an American business media brand that began with Forbes magazine, founded by B.C. Forbes in 1917. It is known for rankings such as the Forbes 400, the World’s Billionaires list and 30 Under 30, and operates Forbes.com, ForbesLive events and licensed editions in 81 countries. Since 2014 a majority stake has been held by the Hong Kong investment group Integrated Whale Media; a 2021 plan to list on the New York Stock Exchange through a special purpose acquisition company was terminated in 2022. It is headquartered in Jersey City, New Jersey.Open the full entry wrote in August 2024 that Yann LeCun had been his computer science professor at New York University and had by then joined Groq as a technical advisor.⁠[8][1]

He said that had it not been for New York University he did not think he would have ended up at Google: he went to an event at Google, where one of the people with whom he had taken the Ph.D. courses recognized him and referred him.⁠[3][3]

Chamath Palihapitiya wrote that he entered the chip industry while working on ad optimization at Google’s New York office. He himself said that what he first worked on at Google was testing for ads: the team built giant test systems that ran 100 tests live on every ads query but did not have the production system’s budget, so it had to write its own threading library; he considered this harder engineering than the production system itself.⁠[4][3][3]

The biographies published by Groq and by the World Economic Forum both state that he began, as a 20% project, the work that became Google’s Tensor Processing Unit (TPU), and that he designed and implemented the core elements of the first-generation TPU chip.⁠[8][9]

Chamath Palihapitiya’s account is that he overheard Google’s speech recognition team complaining that they could not get enough compute, asked Google for some budget, and began putting together a chip-based machine learning accelerator for them, working in the ads business during the day and with the accelerator team at night; after winning Google’s approval, he and his team built the Tensor Processing Unit and began deploying it across Google’s data centers within a year, and the TPU eventually underpinned more than 50% of Google’s compute power.⁠[4]

He himself said that 20% time, in his case, could be used only outside his working hours: every night he went to work with the speech team on this side project, which was separate from his main project; the speech team bought him some hardware, and the money came from what a vice president called his “slush fund,” or leftover money. Two other projects to build AI accelerators existed in the company at the time and nobody expected his to succeed, which gave the team cover to do counterintuitive things; once it became successful, he said, “they brought in the adult supervision.”⁠[3][3]

He recalled that this began in 2012: no machine learning model had yet outperformed a human being at any task, and the speech team had trained a model that transcribed speech better than people did but could not afford to put it into production. Jeff Dean gave the leadership team a presentation of just two slides, one with the good news that machine learning works and one with the bad news that the company could not afford it; by his account, speech recognition alone would have required doubling or tripling Google’s global data center footprint, which at an average of $1 billion per data center, for 20 to 40 data centers, meant $20 billion to $40 billion.⁠[3][3][3]

He said that Jeff Dean noticed that matrix multiplication was the main algorithm consuming most of Google’s CPU cycles, and that the team decided to build a massive matrix multiplication engine around it, using a systolic array; the two other competing teams took more traditional approaches, and one of them was led by a Turing Award winner. He recounted that this award winner, speaking about the TPU, remarked that whoever came up with it must have been really old, because systolic arrays had fallen out of favor; it was in fact him, he said, and he had not known what a systolic array was until someone explained the term to him.⁠[3][3][3]

The paper “In-Datacenter Performance Analysis of a Tensor Processing Unit,” published by Google researchers at ISCA 2017, lists him among its authors, and its abstract states that the TPU had been deployed in Google’s data centers since 2015. He said that the TPU had been in production at Google for about a year before it was announced.⁠[10][11]

He recalled that he had by then moved to Google X and was no longer working on the TPU; a former member of the TPU team who was also at Google X showed him an email from DeepMindGoogle DeepMindGoogle DeepMind is Google’s artificial intelligence research organization, formed in April 2023 by merging DeepMind, which Google acquired in 2014, with the Brain team from Google Research, and led by CEO Demis Hassabis. It developed AlphaGo, AlphaFold, and the Gemini family of models, and Hassabis and John Jumper received the 2024 Nobel Prize in Chemistry for AlphaFold’s protein structure prediction work.Open the full entry saying that they had a competition with a prize purse, expected to lose it, and wanted to know whether the TPU was as fast as they had heard. They answered yes; the reply was that the competition was 30 days away, against the world champion at Go, and that the test games had been lost, so the program had to be ported to the TPU.⁠[5][5][5]

Rana el KalioubyRana el KalioubyRana el Kaliouby is a computer scientist and entrepreneur in emotion AI who co-founded the MIT Media Lab spin-off Affectiva with Rosalind Picard in 2009 and later became its CEO; after Smart Eye acquired Affectiva in 2021 she turned to investing and launched Blue Tulip Ventures.Open the full entry, host of Pioneers of AI, places the AlphaGo match in 2016; the chapter list of Eric Ries’s show names it as the DeepMind Challenge Match between AlphaGo and Lee Sedol.⁠[11][12]

He said that the last thing he worked on was getting AlphaGo, DeepMind’s Go-playing software, to run on TPUs and that, once the software had been ported, 48 TPUs won 99 games out of 100 against 170 GPUs running exactly the same software. Recalling the ratings in 2026, he said that AlphaGo’s Elo score on GPUs was about 3,200, that Lee Sedol’s was about 3,550, and that on TPUs the score rose to about 3,900, adding that he might have the first digit wrong but that the model itself was exactly the same.⁠[3][3][5][5]

He also said that the famous move 37 of the second game was a one-in-10,000 move in the games on which the system had been trained and that, when the game was replayed on GPUs afterwards, the program never found it because it lay too deep in the search chain. The conclusion he drew was that computing faster allows deeper thinking and so makes a model smarter.⁠[5][5][5]

In 2026 he said that GPUs had since become as good as TPUs and better, which he put down to the whole industry and ecosystem behind them, while holding that the TPU had some novel innovations at the time.⁠[5]

Groq’s biography of him states that after the TPU he joined Google X’s Rapid Eval Team, the initial stage of its moonshot factory, where he devised and incubated new Bets for Alphabet.⁠[8][6]

He said that all big companies end up becoming political and that, when something is that successful, a lot of people want to own it and more senior people start reaching for it; he moved on to Google X’s Rapid Eval Team, the team that comes up with the crazy ideas at Google X. He was having fun there, he said, but nothing was turning into a production system, and he wanted to do something real again from start to finish, taking it from concept to production, so he began looking outside. He also said that he had had no intention at all of building a chip: he had built the highest-performing image classifier as well, but saw that the software and the models were all being given away free, concluded that machine learning would be open source, and could not imagine building a business around that; chips, by contrast, take so long to build that when something innovative is launched it is four years before anyone can copy it, let alone pull ahead.⁠[3][3][3][3]

He said that after the TPU was announced other companies began calling him about building AI chips, and that one company pitched him to join it on the argument that Google’s own chip gave it a large advantage and that AI compute should not be concentrated in one place, while he did not think going from one player to two was much of an improvement. Chamath Palihapitiya wrote that other hyperscalers had tried to hire him to build custom chips for them.⁠[11][4]

Eric Ries’s show notes state that he left Google in 2016 to launch Groq, judging the risk of leaving to be lower than the risk of staying and watching the project die in-house.⁠[12][1]

Forbes wrote that he founded Groq in 2016 with Doug Wightman, a fellow Google engineer, who served as the first chief executive and left a few years later; in the same year Groq raised $10 million in a round led by Social Capital.⁠[1]

Chamath Palihapitiya wrote that he led Groq’s founding investment in 2016 and that it became increasingly clear to Ross that a gap would emerge between companies with next-generation AI compute and those without, so he set out to build a chip that would be available to everyone and founded Groq. In their 2024 conversation Chamath Palihapitiya said that he was shocked when Google began talking publicly about the TPU, concluded that Google must know something that others did not, and met Ross probably a few weeks afterwards; it was probably the fastest investment he had ever made, he said, and because Ross did not yet have a company, the company was incorporated after the check had been written.⁠[4][3][3]

He said that when fundraising began the team was not even 100% sure that it would do something in hardware; when Chamath Palihapitiya asked what he would do differently, his answer was the software: Google could build the chips, but every team at Google had a person dedicated to hand-optimizing the models, and the people Groq had just started hiring from NVIDIA said that NVIDIA hand-optimized its kernels in the same way, which he judged could not scale because every application needs an engineer. The company therefore focused on the compiler for its first six months, and he banned whiteboards at Groq because people kept trying to draw pictures of chips. On Pioneers of AI he likewise said that his first point was to start with the software, because the software on all these AI chips is a mess and even the TPU has to be optimized by hand.⁠[3][3][3][11]

He said that a 15% improvement would not persuade customers to move to a new architecture and that the design had to be five to ten times better; that goal justified using an older, underutilized 14-nanometer process, doing without external memory, and using older interconnect technology.⁠[3][3]

Kisaco Research’s speaker page calls him Groq’s “technical founder and CEO” and gives his title as Co-Founder & CEO; the RAISE Summit page likewise calls him Co-Founder & CEO. Entrepreneur wrote in July 2026 that he helmed the company for seven years.⁠[7][13][14]

In 2026 he said that the first thing a founder has to do is move from the technical work he is good at to learning how to manage people, and that this probably cost Groq three to four years; of himself when he started he said, “I was a terrible leader.” He said that he is by nature a delegator but entrusted autonomy to people who were not suited to working independently, so that things ground to a halt because those people were used to being given instructions; when, in frustration, he went in and gave instructions, it was so unnatural to him that they did not accept it.⁠[5][5][5][14]

In the EE Times interview he said that he was not very confident when he first became a chief executive and that, about two years into Groq, he was struggling as a leader; the chief executive of a well-known company, who managed about 2,000 people, offered to let him shadow him for a day and let him follow from the time he arrived until the time he left, including in highly confidential meetings. He recalled in 2026 that Groq then had about 35 people; in each meeting he silently worked out what he would decide, and every time the other man’s decision was the one he would have made, which made him realize that what he lacked was confidence. From then on, he said, he acted with confidence; his decisions did not change, but people followed his direction much more.⁠[6][5][5][5]

In the EE Times interview he said that three or four times that day the other man stopped to debrief with his chief of staff and they asked for his observations too, and that he had noticed things of value that they had not, which taught him the value of being an outsider to a situation; since then he has made a point of having an appropriate amount of “naivety” in the room when making any important decision. He also said that the day taught him that people are inspired more by a person’s actions than by the person.⁠[6]

Forbes wrote that after the 2016 round Groq found it hard to raise more money; recalling a low point in 2019, he said that the company was then a month from running out of money. He said, “Groq nearly died many times,” and also, “We started Groq maybe a little bit early.”⁠[1]

In 2026 he explained that some of Groq’s early investors later fell out of favor with other venture capital firms, which did not want to co-invest with them, so every fundraising was a struggle. As he put it, venture firms on the U.S. West Coast tend to follow one another—when one invests the others want to, and when one passes the others pass too—whereas investors in New York run their own analysis and do not care what others are doing. He said that in the end Groq had very few of the typical West Coast venture firms among its shareholders and more crossover funds from the East Coast.⁠[5][5][5]

He said that many years ago the company was at one point about three weeks from running out of money. The leadership team of the time was already drawing up a list of layoffs, and when he reviewed it he saw that the layoffs would kill the company: the product did not yet work, the company first had to write a compiler of a kind never written before, one that did not require people to hand-write what are called kernels, and the people on the list were critical to that.⁠[5][5][5][5]

His answer was to cut pay instead of people. The company held an all-hands meeting, put up pictures in the style of Second World War war bonds, and called the plan “Groq Bonds”; he said that it was not technically a bond but an exchange of salary for equity. The company expected high attrition, but 80% of employees took part and about half went down to the statutory minimum salary, which he put at about $50,000 to $60,000, when engineers were being paid hundreds of thousands of dollars.⁠[5][5][5]

He said that this saved more than three weeks of runway—closer to two months—and that the company had three weeks of money left when it next raised, so that without it the company would have gone out of business. Attrition after the plan was announced, he said, was under 10% and perhaps closer to 5%, probably lower than before; he sums the approach up as “put everyone’s hands on the steering wheel,” meaning that passengers are more nervous on a dangerous road than the driver is, and that people who take part in saving the company are more willing to take the risk.⁠[5][5]

Forbes wrote that Groq was valued at $1.1 billion in 2021 and that by August 2024 it had been selling chips for about two years; before ChatGPTChatGPTChatGPT is a generative AI assistant launched by OpenAI on November 30, 2022. Through websites and applications, it provides conversation, writing, coding, search, file analysis, and image and voice interaction. It evolved from a free research preview into personal, enterprise and education plans, gradually adding deep research, Study mode and cross-application task execution. Its use has prompted continuing discussions about factual accuracy, personal privacy, copyright, education and teen safety.Open the full entry was released in late 2022, demand for fast inference was limited. Documents seen by Forbes showed 2023 sales as low as $3.4 million and a net loss of $88.3 million.⁠[1]

He recalled in 2026 that when LLMs were first emerging, the chief executive of GitHub telephoned him to say that they could now do code completion with LLMs but, even as part of Microsoft, could not get GPUs, and asked whether they could use Groq’s chips. He went to his team, which said that such models could not run on Groq’s chips; he thought the opposite but let himself be talked out of it. He was talked out of a second such opportunity as well; the third time, he worked out the performance figures himself, and although everyone disagreed that it could be done, those were exactly the numbers the company ended up hitting.⁠[5][5][5]

He calls these lucky opportunities that he failed to seize and reflects that the outcome would have been very different had Groq been the original platform running LLMs at MicrosoftMicrosoftMicrosoft is a technology company founded by Bill Gates and Paul Allen in 1975 and headquartered in Redmond, Washington, with businesses spanning Windows, Microsoft 365, Azure, enterprise software, artificial intelligence, gaming and devices.Open the full entry for OpenAIOpenAIOpenAI is an artificial intelligence research and product organization founded in 2015 that develops the GPT model series, ChatGPT, Codex and a developer API, and conducts research on image, speech, video and intelligent agents. After its 2025 restructuring, the nonprofit OpenAI Foundation continues to control the for-profit OpenAI Group PBC. Its growth has involved large-scale fundraising and compute partnerships, and has also been accompanied by controversies over governance, copyright, privacy and AI safety.Open the full entry.⁠[5]

He said that he concluded early that fast inference would matter, but that early prospective customers who saw the demonstrations asked why an LLM needed to be faster than a person can read; there was also strong resistance inside Groq, and in the period about three or four years before 2026 the company had heavy turnover, with many people saying as they left that there was no point to fast inference.⁠[5][5][5]

He recalled that about three months before ChatGPT caused its sensation he saw, at an event of something he referred to as “SCSP,” in which Eric SchmidtEric SchmidtEric Schmidt is a technologist, entrepreneur and philanthropist who was Google’s chief executive officer from 2001 to 2011 and later executive chairman of Alphabet; he chaired the U.S. National Security Commission on Artificial Intelligence and has led the rocket company Relativity Space since 2025.Open the full entry was involved, an LLM from AnthropicAnthropicAnthropic is an artificial intelligence research and product company founded in 2021 that develops the Claude models, assistant, and tools such as Claude Code, and operates as a public benefit corporation. Dario Amodei and Daniela Amodei are its co-founders, and its research covers model safety, interpretability, and steerability.Open the full entry answering questions from the audience without the room reacting; the Special Competitive Studies Project (SCSP) is an initiative that Eric Schmidt announced in October 2021 and that was to be headed by him. From this he concluded that people would understand the value of speed only if they could try it with their own questions, and so decided to put the demonstration directly on the internet.⁠[5][5][15]

On February 5, 2024, Chamath Palihapitiya published an article summarizing a Space that he had hosted with Ross the previous Friday, covering Groq’s founding story, how to build great AI hardware, LLMs, and the implications for the key players in AI.⁠[4]

Fortune wrote that in February 2024 Groq launched a chatbot on its website that it said set speed records for LLM output on open-source models, including MetaMetaMeta (Meta Platforms, Inc.) is a publicly listed American technology company, renamed from Facebook in 2021, that operates social communication, digital advertising, virtual reality, smart glasses and artificial intelligence businesses.Open the full entry’s Llama, and that a developer posted a video on X showing Groq running a model from Mistral and answering questions in under a second. Thousands of developers then wanted to build on Groq’s chips, and he called the moment “100% a game changer.”⁠[2]

Forbes wrote that in February of that year he was in Oslo demonstrating the chatbot to members of the Norwegian Parliament and technology executives while pitching a European data center powered by Groq chips; the demonstration lagged slightly, because a developer’s post had gone viral the day before and a surge of new users had overloaded Groq’s servers. He said, “I just kept checking the numbers.”⁠[1]

He recalled in 2026 that he was then on a tour of several countries looking for customers and that, during a presentation in Norway, queries to the open-source models felt a little slower than usual; on checking he found that usage had skyrocketed, because someone had posted on X a video of an LLM running very fast on Groq, after which people began building applications with it and posting them.⁠[5][5]

Fortune wrote that about six months before August 2024 he had sent Elon MuskElon MuskElon Musk is an entrepreneur who founded SpaceX and has long served as chief executive of Tesla; he also took part in creating PayPal, OpenAI, Neuralink, The Boring Company, and xAI. He acquired Twitter in 2022, and his businesses span spaceflight, electric vehicles, energy, artificial intelligence, and social media.Open the full entry a sarcastic cease-and-desist letter over xAI’s chatbot, Grok.⁠[2]

Forbes wrote that GroqCloud launched in March 2024, letting developers rent access to the chips, and drew 70,000 sign-ups in its first month; payments were turned on on June 30. Groq also hired Stuart Pann, a former IntelIntelIntel is a semiconductor company headquartered in Santa Clara, California, founded by Robert Noyce and Gordon Moore in 1968; it introduced the 4004 microprocessor in 1971, its x86 processors have long powered personal computers, it has pursued a foundry strategy since 2021, and Lip-Bu Tan has been its chief executive officer since 2025.Open the full entry executive, as chief operating officer and signed a partnership under which Aramco Digital, the technology arm of Saudi AramcoSaudi AramcoSaudi Aramco (officially Saudi Arabian Oil Company, now branded as Aramco) is Saudi Arabia’s integrated energy and chemicals company. It manages Saudi Arabia’s hydrocarbon reserves base, and its shares trade on the Saudi Exchange (Tadawul) under the ticker ARAMCO.Open the full entry, would use Groq chips.⁠[1]

Groq’s leadership page states that Sunny Madra sold Definitive Intelligence, a company he had founded, to Groq and leads GroqCloud.⁠[8]

In the conversation with Chamath Palihapitiya published in April 2024 he said that Groq had 75,000 developers slightly over 30 days after launching its developer console, and compared this with the seven years it had taken NVIDIA to reach 100,000 developers. Chamath Palihapitiya mentioned that a month earlier he had been on stage in Saudi Arabia with Saudi Aramco announcing a big deal; he answered that the announcement was that Groq would do a deal with Aramco Digital, that its size had not been announced, and that the amount of compute to be deployed would be large. The deployment targets he gave at the time were 100,000 LPUs by the end of that year and 1.5 million by the end of the next, which he said would probably be about 50% of the world’s inference compute; the deals already done, he said, came to more than 10% of the 1.5 million goal.⁠[3][3][3][3][3]

In August 2024 Groq announced a $640 million funding round valuing it at $2.8 billion. The round was led by BlackRockBlackRockBlackRock is a publicly traded investment management company headquartered in New York, founded in 1988 by Larry Fink and seven others and listed on the New York Stock Exchange (ticker BLK); its businesses include the iShares family of exchange-traded funds and the Aladdin investment and risk management platform. As of June 30, 2026, its assets under management totaled $15.3 trillion.Open the full entry Private Equity Partners, with CiscoCiscoCisco Systems, Inc. is an American networking technology company, incorporated in 1984, headquartered in San Jose, California, and listed on Nasdaq.Open the full entry Investments and the Samsung Catalyst Fund taking part.⁠[2][1]

Fortune wrote that Groq then had 200 employees and that 300,000 developers were accessing its solutions and hardware through its AI cloud service, the vast majority of them free of charge; he said that because of limited capacity paid access had been made available to only “little over 30 customers.” He said that revenue that year would be pretty good, without giving figures, and that the company intended to recoup the money it had raised, getting back every dollar on the hardware it deployed, and had no intention of being a startup that lives off venture funding. Forbes put GroqCloud sign-ups at 350,000 and wrote that, by the account of its sources, Groq had forecast $100 million in sales for the year, a target that those sources doubted it would hit.⁠[2][1]

Forbes wrote that Mark ZuckerbergMark ZuckerbergMark Zuckerberg is the founder of Facebook, Chairman and CEO of Meta, and a co-founder of the Chan Zuckerberg Initiative. His work spans social networking, virtual reality, and artificial intelligence.Open the full entry announced that Groq would be one of the companies providing chips to run inference for Meta’s Llama 3.1 model. The article has him speaking from inside Groq’s semiconductor lab in San Jose, CaliforniaCaliforniaCalifornia is a U.S. state on the Pacific coast; its capital is Sacramento and its largest city is Los Angeles. California was admitted to the Union on September 9, 1850, as the 31st state. Its population at the 2020 Census (April 1, 2020) was 39,538,223, the largest of any state.Open the full entry; in it he said, “Compute is the new oil,” and, of Groq’s position, “It’s sort of like we’re Rookie of the Year,” adding that Groq was nowhere near NVIDIA yet.⁠[1]

On February 10, 2025, Groq announced that it had secured a $1.5 billion commitment from the Kingdom of Saudi Arabia for expanded delivery of its LPU-based AI inference infrastructure, announced at LEAP 2025. The announcement names him as Groq’s chief executive and founder and quotes him as saying that it is an honor for Groq to support the Kingdom’s Vision 2030; it states that at LEAP 2025 he, alongside Ahmad O. Al-Khowaiter, chief technology officer of Saudi Aramco, and others, demonstrated reasoning LLMs, a model created in Saudi Arabia, and text-to-speech models in English and Arabic running live.⁠[16]

The announcement states that from its data center in Dammam, Saudi Arabia, Groq delivers AI inference to customers worldwide through GroqCloud, and that the agreement follows Groq’s building, in December 2024, of what the announcement calls the region’s largest inference cluster, which was brought online in eight days.⁠[16]

On Pioneers of AI he said that Groq had been talking with Saudi Arabia for about a year, that only about four months passed between the decision to proceed and completion, and that 51 days after the contract was signed almost 20,000 chips were up and running; those chips had worked out so well, he said, that a deal had been signed to expand the deployment significantly that year.⁠[11]

Forbes, discussing Groq’s business in Saudi Arabia in an article of March 2026, wrote that a year earlier he had said on the 20VC podcast that the company had not raised $1.5 billion and that the figure was revenue, while a former employee’s account was that the figure referred to the cumulative value of the services that were to be built. The article also states that, by the account of three former employees, the deal with Aramco Digital was structured as a revenue-sharing arrangement in which Groq sold chips to Aramco at a relatively low price, Aramco funded the data center, and the two shared the income from the chips’ computing capacity.⁠[17]

TechCrunchTechCrunchTechCrunch is an American technology news outlet founded in 2005 by Michael Arrington and Keith Teare and headquartered in San Francisco; it mainly covers technology companies, startups, and venture capital, and hosts events such as TechCrunch Disrupt. Its ownership passed successively through AOL, Verizon, and Yahoo, and in 2025 Regent announced its acquisition.Open the full entry wrote on September 17, 2025, that Groq had confirmed raising $750 million at a post-money valuation of $6.9 billion. The round was led by Disruptive Technology AdvisersDisruptive Technology AdvisersDisruptive Technology Advisers (brand name DISRUPTIVE) is a private technology investment firm headquartered in Dallas, Texas. Since 2012 it has invested in privately held technology companies through its private funds, and it is an investment adviser registered with the US Securities and Exchange Commission; in its Form ADV updated in August 2026, it reported regulatory assets under management of about US$7.68 billion.Open the full entry, with BlackRock, Neuberger Berman, Deutsche Telekom Capital Partners, and others taking part, along with the existing investors Samsung, Cisco, D1, and AltimeterAltimeter CapitalAltimeter Capital is a US technology investment firm founded by Brad Gerstner in 2008, investing through public and private market funds. Its legal entity is Altimeter Capital Management, LP, with offices in Menlo Park and Boston. Disclosed investments include Snowflake financing, Grab’s listing transaction and AI company financing.Open the full entry; when word of the raise leaked in July, the figures mentioned had been about $600 million at a valuation near $6 billion.⁠[18]

The article states that, by PitchBookPitchBookPitchBook is a private capital market data, research and software company headquartered in Seattle, Washington, United States. Founded by John Gabbert in 2007, it was fully acquired by Morningstar in 2016 at a valuation of $225 million and operates as a Morningstar subsidiary. The PitchBook segment’s revenue was $671.8 million in 2025.Open the full entry’s estimate, Groq had by then raised more than $3 billion in total, and that Groq said it powered AI applications for more than 2 million developers, up from about 356,000 a year earlier.⁠[18]

On December 24, 2025, Groq announced that it had signed a non-exclusive licensing agreement with NVIDIA for its inference technology. The announcement states that the founder, Jonathan Ross, the president, Sunny Madra, and other members of the Groq team would join NVIDIA to help advance and scale the licensed technology, that Groq would continue to operate as an independent company with Simon Edwards as chief executive, and that GroqCloud would keep running without interruption.⁠[19]

ReutersReutersReuters is an international news agency founded in London in 1851 by Paul Julius Reuter and is now the news business of Thomson Reuters. It provides text, photo, video, and graphics coverage to media and institutional clients and distributes it through channels such as Reuters Connect.Open the full entry wrote that Groq did not disclose financial details of the deal; CNBCCNBCCNBC is an American business and financial news cable channel launched by NBC in 1989; it merged with its rival FNN in 1991 and since January 2026 has been part of Versant Media Group, spun off from Comcast.Open the full entry’s account was that NVIDIA had agreed to acquire Groq for $20 billion in cash, on which neither company commented. TechCrunch wrote that NVIDIA said to it that the transaction was not an acquisition of the company, and that if CNBC’s numbers were accurate the purchase would be NVIDIA’s largest ever.⁠[20][21]

Forbes wrote in its article of March 2026 that the previous winter he walked into a meeting with Jensen HuangJensen HuangJensen Huang is the co-founder, president, and CEO of NVIDIA and has served on its board since the company was founded in 1993. He has a background in electrical engineering education and the semiconductor industry, and under his leadership NVIDIA’s business expanded from graphics chips to GPU general-purpose computing, artificial intelligence, and data center computing platforms.Open the full entry with a proposal that the two companies’ technologies work together, while also seeking approval to buy about 100,000 Blackwell units; three days later Jensen Huang called to say that they should move quickly, and three weeks after that, on Christmas Eve, NVIDIA announced the $20 billion deal to license Groq’s product and hire most of its staff. In his own telling, only about three weeks passed between the call on which the idea was first floated and the money arriving in the bank; Groq had already integrated GPUs and LPUs and meant to deploy them itself, which was why it asked NVIDIA to sell it about 100,000 GPUs, and Jensen Huang, seeing what had been done, thought it better to make the combination available to all of NVIDIA’s customers.⁠[17][5][5][5]

He said that the idea of putting the two kinds of chip together came from Sunny Madra, then chief operating officer, and that he himself did not at first think it would be a big deal and so had not proposed it; the team worked on it for three or four months or longer, establishing that GPUs are better at the attention portion of an LLM’s decoder layer and LPUs at applying the weights, and placed each part on the chip that suited it. Groq was not afraid to show the result to NVIDIA, he said, because it wanted to be NVIDIA’s customer. Forbes wrote that he credited Sunny Madra with unlocking the deal.⁠[5][5][5][17]

He said that this time Groq was nowhere near running out of money, that the amount of the licensing deal was only a little over twice the company’s last valuation, and that the company could also have raised money at that amount. Groq had about 450 employees when the partnership was made, he said, but because the organization was so creative, managing it was in many ways more like managing 5,000.⁠[5][5][5][5]

The Forbes article estimates that he would take about $950 million in cash after taxes, on an estimated stake of 9%; it states that $3 billion is said to have been set aside for the employees who moved to NVIDIA, that the NVIDIA stock compensation he received is probably a significant part of it, and that when that stock is released he will be what the article calls a new “multimillonario.” He declined to comment on his wealth; the article judges that it would have been worth far more had the deal been closed as a stock transaction or an asset sale and not as a licensing agreement paid mostly in cash and taxed twice. The other big winners the article lists are the investor Chamath Palihapitiya and his firm Social Capital, the co-founder Doug Wightman, whose stake was estimated to be similar to his, and Sunny Madra.⁠[17]

The Forbes article also states that, by the account of two sources, Groq’s annual revenue was around $100 million at about the time of the deal, whereas the company’s initial projection for 2025 was said to have been $2 billion, later lowered to $500 million, and that Groq pays EquinixEquinixEquinix, Inc. is a digital infrastructure company headquartered in Redwood City, California, that operates data centers and interconnection services worldwide.Open the full entry up to $300 million to host its LPUs in Australia. The article describes Saudi Arabia’s sovereign fund as having been Groq’s most important contract and main customer, while Ross said that the company was doing well, was meeting some objectives, and was talking to other large customers, and declined to comment on 2025 revenue; it adds that former employees said a sale of the company left after the deal had been expected but had not happened by the time of the article.⁠[17]

The Forbes article describes him as by then NVIDIA’s chief software architect and Sunny Madra as a vice president of NVIDIA; Entrepreneur wrote in July 2026 that he left Groq in December 2025 to join NVIDIA full-time as a chief software architect. The Forbes article records his recollection that he had thought he would die at Groq and had always said to everyone that the company wanted to deliver half of the world’s inference; of arriving at NVIDIA, he said it was like being children in a toy store. He also said that the independent Groq was still growing as an LPU cloud provider.⁠[17][14]

On March 16, 2026, Jensen Huang announced in his GTC keynote that NVIDIA would use Groq’s LPUs to raise the inference performance of its new Vera Rubin rack systems; The Register wrote that each LPX rack holds 256 LPUs and was expected to ship with Vera Rubin later in 2026. NVIDIA’s product page describes NVIDIA Groq 3 LPX as an inference accelerator for the Vera Rubin platform that pairs Rubin GPUs with LPUs.⁠[22][23]

Forbes wrote that the integration uses Groq’s third-generation chip; in Ross’s telling, the second generation arrived late and had not yet reached scale, and Groq moved to a new version because integrating the second generation with NVIDIA would have taken as long as starting from scratch with a third. He also said that production would be at NVIDIA’s scale and not a pilot and that he had taken part in many customer meetings; the article notes that no buyer had yet been announced.⁠[17]

In the interview published in July 2026 he described what he had observed at NVIDIA: it is the least political large organization he has seen, and Jensen Huang does not hold one-on-one conversations in which different people hear different things. He also said that at Groq he had been too fond of playing “3D chess,” whereas Jensen Huang simply asks what the customer needs, sells customers only what they actually need, and always says to customers things that he believes to be true and can support.⁠[5][5][5][5]

On June 22, 2026, Groq announced $650 million in new growth capital, in a round led by Disruptive Technology Advisers and Infinitum; the announcement names as its leadership the chairman, Alex Davis, founder and chief executive of Disruptive, the chief executive, Adam Winter, and the chief financial officer, Matt Eng, and traces the company’s current direction to the non-exclusive licensing agreement signed with NVIDIA in December 2025. Entrepreneur wrote that Adam Winter had previously been a vice president at the company.⁠[24][14]

On August 17, 2026, Groq announced a $350 million Series A led by Disruptive Technology Advisers, with planned participation from NVIDIA, valuing the company at $3.5 billion; the announcement states that, together with the $650 million of June, Groq’s recent funding came to $1 billion and that it operates 13 data centers across North America, Europe, the Middle East, and Asia Pacific.⁠[25]

On July 5, 2026, David SenraDavid SenraDavid Senra is the founder and host of the Founders podcast, on which he reads biographies of entrepreneurs to discuss starting and running businesses and making personal decisions. He began producing the show in 2016, partnered with Colossus in 2022, and in 2025 launched David Senra, a long-form interview show under his own name.Open the full entry published a conversation with him on his YouTubeYouTubeYouTube is Google’s video-sharing and creator platform, founded in 2005 by Steve Chen, Chad Hurley, and Jawed Karim and acquired by Google in 2006. Built around video uploading, viewing, and creator earnings, the platform has developed Shorts, YouTube Music, Premium, YouTube TV, and products for children, and is headquartered in San Bruno, California, United States.Open the full entry channel under the title “Why Asking the Right Questions Is the Most Important Skill in the AI Age.” David Senra said in the episode that the two had had this conversation before, at NVIDIA GTC, and referred to what they had talked about on stage there.⁠[5][5][5]

The conversation opens with the deal between Groq and NVIDIA and moves on to why GPUs and LPUs complement each other, why speed matters more when AI talks to AI, payments made by agents, his side projects and daily brief, and his view of the ability to ask questions.⁠[5][5][5][5][5][5]

He then discusses his understanding of leadership, his delegating style of management, how he learned to manage, what he has learned at NVIDIA, his experience shadowing a chief executive, Groq’s fundraising difficulties, his view of venture capital behavior (which he compares to the Keynesian beauty contest), and how the partnership with NVIDIA came about.⁠[5][5][5][5][5][5][5][5]

The topics of the second half are why speed can raise the quality of a model, the history of the TPU and AlphaGo, the “reality quotient” and choosing the dominant game, change management, “return on luck” and missed opportunities, the bet on fast inference and the viral demonstration, “intentional leadership,” “Groq Bonds,” the “people spec” used in hiring, loss bias, what he calls “manufactured discontent,” his discontent with the shortage of compute, and the effect of AI on software development and education.⁠[5][5][5][5][5][5][5][5][5][5][5][5][5][5]

Entrepreneur published an article on July 7, 2026, quoting his reflections in the episode on his early leadership; the article refers to the show as the Founders podcast.⁠[14]

He said that his way of leading was to set a high-level direction and not to say how to do things: it took him a while to arrive at a goal simple enough to put on a challenge coin, and everyone at Groq had such a coin, which read 25 million tokens per second and showed a rising graph. The goal, he explained, referred to inference capacity in Groq’s data centers, and everyone could connect their own work to it: making the chip faster, making the software faster, deploying more, getting power costs down, finding more data centers, fabricating more chips, and optimizing the supply chain.⁠[5][5][5]

He said that he hired very autonomous people who often go off and execute on their own and who would be terrible in most corporate environments; he also acknowledged that the better the people, the harder they are to manage.⁠[5][5][5]

He said that Groq hired for “reality quotient,” as distinct from intelligence quotient: at its simplest the ability to recognize reality, and in its most developed form the ability to choose the dominant game among those being played. The example he gave was that MySpace focused on the number of accounts signed up while Facebook focused on monthly active users.⁠[5][5][5]

Groq had a “people spec” with version numbers that was revised over time and that set out, in the manner of a product specification, what kind of people the company was looking for. The positive traits he cited were “return on luck” and “poetic design,” meaning the most minimal expression of what one is trying to achieve; their respective negatives were squandering luck and maximalist design that throws in every feature.⁠[5][5][5]

He said that he had been very good at hiring smart and talented people but that many of them caused organizational problems, because he could always come up with a reason to hire someone. Groq later brought in a head of human resources who was very good at noticing problems with people and getting them out; watching her, he realized that he had been hiring all wrong, and he switched from looking for positives when hiring to looking for negatives. The distinction he draws is that when growing people one shows them the positive, and when selecting people one screens for the negative traits.⁠[5][5][5][5]

On building an AI team in Silicon Valley, he said in his 2024 conversation with Chamath Palihapitiya that AI researchers are amazing at AI but that the field is new and they are still green; his recommendation is to hire the best, most seasoned engineers who know how to ship on time and let them learn AI, because they will do that faster than AI researchers can be given 20 years of experience of deploying production code.⁠[3][3]

He said that people have a natural loss bias, to which some attach a number: a loss is six times more painful than a gain; someone who invests and loses 20% finds it very painful, for example, while not having invested in something that grew 100% hurts less, even though the opportunity cost is much higher.⁠[5][5]

He also valued a trait that he calls “booking the win early”: when someone said in an architecture meeting that a certain change would make the chip twice as fast, others heard something to put into the next chip, while he heard that the current chip would be half as fast as it could be without it. As soon as he hears that something is possible, he said, he treats it as already won and as a loss if it is not done, and he looked for people with the same attitude when hiring.⁠[5][5][5]

He said that most people are so afraid of putting themselves out there and suffering a bad outcome that they keep their sights low, whereas entrepreneurs start a company certain that they will succeed, say so to everyone, raise money, put their reputation on the line, and are thereby forced to perform. He said this while discussing an episode of David Senra’s Founders podcast about a basketball star: he suspects that the player, who made bets on everything and constantly taunted other players, did so deliberately so that a loss would be humiliating, to force himself to perform at superhuman levels.⁠[5][5][5]

He said that doing something disruptive meant that his job was full-time change management, and that the difficulty a founder often meets is that he sees the dominant game and explains it, yet people want to stick with the old way of doing things or care more about the process.⁠[5][5]

In the EE Times interview he described himself as still a relatively new chief executive whose leadership style was still developing. The principles he listed were: say things that are difficult to disagree with; don’t whisper, broadcast; hire people who are smarter and better and get out of the way; effectiveness is better than being right; learn, learn, learn; grow people; T-shirt and jeans; let others speak; treat the team’s time as more valuable than his own; maximize talent density; defy gravity; get comfortable being uncomfortable; fear is a sign that you are doing something that matters; recruit, empower, appreciate, and decide; and “Every. Word. Matters.”⁠[6]

In the same interview, asked for advice to early-career engineers and people who want to start a company, he said that he had written and deleted a great deal of advice, because advice is easier to give than to follow and depends on context, and so offered only some signs to watch for: to pay attention when someone says, “That can’t possibly be right, because if it was, everyone would do it,” because it means the idea is either really good or really stupid but certainly not mediocre, and all breakthroughs are at first things that no one does; that not feeling fear means that one is not doing anything important; and that one’s team is the people with whom one is willing to share rough drafts, so that feeling the need to withhold information from someone means that one is not working with one’s team. Asked for the biggest threat that technology poses to society, he answered complacency.⁠[6]

He said that successful entrepreneurs tend to have something he calls “manufactured discontent.” He described a conversation with a group of entrepreneurs and successful people from other fields: the entrepreneurs, though they had more money, were the least happy with their wealth, while the others were unhappy with their own earlier work, and everyone present had some discontent that drove them. Looking at his own life, he said that before Groq found product-market fit his discontent was genuine, and that afterwards he was unhappy with the scale and then with other things, always finding something to be unhappy with.⁠[5][5][5]

In the EE Times interview he said that he lives in the Bay Area and spends his weekends learning: reading, thinking, and meeting other chief executives and founders. He said that he reads about one or two books a week, often as audiobooks, and does not know that he has ever read a book twice, but has books that he recommends over and over again: “If I Understood You, Would I Have This Look On My Face?,” “Turn The Ship Around!,” “The Courage to Be Disliked,” “When Things Fall Apart,” “No Rules Rules,” “Influencer,” “Non-Violent Communication,” “Never Split the Difference,” “Decisive,” and “Superbosses.”⁠[6]

In 2026 he said that after reading “Turn the Ship Around” he adopted its approach heavily in his leadership because it suited his autonomous style, and that very early on as a founder he read a book by Jim Collins in which a chapter on “return on luck” resonated with him; as he put it, the chapter’s thesis is that the most successful companies do not have more lucky events than others but seize on their luck better. Jim Collins’s website states that the concept was developed in the book “Great by Choice.”⁠[5][5][26]

He describes himself as someone who is not a control freak: at a restaurant he can simply ask the waiter to bring whatever the waiter thinks is best.⁠[5]

He said that he likes to do cutting-edge side projects on his personal computer, where he has no access to the work codebase and so runs no risk; he spins up a server on GCP or AWSAmazon Web ServicesAmazon Web Services (AWS) is the cloud computing business of Amazon. Launched in July 2002, it made its public debut with the Amazon S3 storage service on March 14, 2006, and has since added hundreds of services, including Amazon EC2; its segment net sales were about $46.0 billion in the third quarter of 2026.Open the full entry and starts building, and has made applications that work out which flights and routes to take and which seats are best, as well as applications that do certain mathematical things. He said that he uses AI extensively at work but always starts with a hobby project before bringing something to work.⁠[5][5]

One of these projects is a “daily brief”: every morning he receives an email that gathers what is happening in the world on the basis of his interests and how he interacts with it, together with research material, which he compares to a presidential daily brief personalized for him. The brief is text with links; he later changed it to give him only a set of headlines on which he asks follow-up questions, because he considers interaction more useful than static content.⁠[5][5][5]

He said that another hobby project needed several telephone numbers so that a number of agents could be with him on messaging applications, and that getting the numbers required him to prove that he was human, which was a great nuisance; from this he concluded that payments are not yet built for agents and that the number of payments will skyrocket once micropayments are possible.⁠[5][5]

He said that his executive assistant now builds software herself: when he travels, she makes a small application that shows the local weather, updates live, and gathers the telephone numbers and other information he needs.⁠[5]

In the episode David Senra read out two things that Ross had said: “Success in the information age was about being able to answer questions. Success in the AI age will be about being able to ask the right questions,” and “looking forward to a year of massive upleveling for anyone who wants it.” Ross also said that a recent post of his about “code rationing” had drawn a good deal of negative feedback.⁠[5][5][5]

Groq’s website carries a question-and-answer interview that he gave to EE Times Weekend, reposted under the title “Jonathan Ross: Every. Word. Matters.,” which covers what he does at weekends, the books he rereads, who has most inspired him, the biggest threat that technology poses to society, his advice to engineers and founders, and his leadership style.⁠[6]

In the conversation published by Social Capital, Chamath Palihapitiya first asked him about his origins and how he came to Google, after which the two discussed the origin of the TPU, his leaving Google and founding Groq, Groq’s early design choices, how Groq compares with NVIDIA, the difference between training and inference, deployment targets, building an AI team in Silicon Valley, and the partnership with Aramco Digital, ending with his view of the future of AI.⁠[3][3][3][3][3][3][3][3][3][3]

On November 7, 2024, Eric Ries released an episode of The Eric Ries Show with him. The show notes state that Groq had had many near-death experiences in its eight years, which he believes made the company stronger, that he holds that true disruption means doing what others believe cannot be done, and that he puts Groq’s mission as “preserving human agency in the age of AI.” The chapter list shows that the conversation also covered how he came to design chips without the usual background and the jump in Groq’s users from 8,000 to 400,000 in a week.⁠[12]

On Pioneers of AI, hosted by Rana el Kaliouby, he said that Groq’s platform had crossed one million developers the week before and that the company was about to hire its first head of prompt engineering. On DeepSeekDeepSeekDeepSeek is a Chinese AI company developing large language models and providing generative AI services through websites, applications and APIs.Open the full entry, he said that the U.S.United StatesThe United States is a federal country in North America made up of 50 states, the District of Columbia as the seat of government, and territories such as Puerto Rico and Guam; its capital is Washington, D.C. The Continental Congress adopted the Declaration of Independence on July 4, 1776. The 2020 census (April 1, 2020) counted a resident population of 331,449,281, and the 435 seats of the House of Representatives are apportioned among the states on the basis of the census.Open the full entry “AI czar” had advocated using a few U.S. companies that run the DeepSeek model, so that the data would not go to China, and had named Groq as one of them; he said that Groq deletes queries once they have been processed and does not retain user data. The host said that the two had last met in December at the Fortune Brainstorm AI conference; the Masters of Scale profile page also lists another episode in which he took part, “From Fortune Brainstorm AI: What does it take to make AI?,” and describes Groq as an AI inference platform based in Mountain View, California.⁠[11][27]

The RAISE Summit website has a page describing a keynote he gave at that conference on Groq’s expansion into Europe. The points the page lists are that Groq’s first data center in the European Union is in Helsinki and was operational in just weeks, that LPUs use one-third of the energy of GPUs and run on air cooling in place of scarce liquid-cooled facilities, that Groq regards inference as a high-volume, low-margin business, that its customer base was growing by 20% to 30% a month, and that its throughput had risen from 10 million to 20 million tokens per second in six weeks.⁠[13]

The World Economic Forum has a profile page for him; the LEAP website lists him as a past speaker, and the World Governments Summit website also has a speaker page for him, describing him as a technology entrepreneur and the founder and chief executive of Groq.⁠[9][28][29]

He said that what he is most discontented with at present is the lack of compute in the world. As he sees it, AI has pros and cons but the pros are massive: if it takes an extra year to cure cancer because there is not enough compute, “that’s my fault”; he said that he feels all of this is on his shoulders and that he has to make sure the world has more compute.⁠[5][5][5]

He said that AI is very good at using AI: a person can wait a second or two for an answer, but an AI simply sits waiting, so speed matters more when agents call one another. He likens adding LPUs to existing models to moving overnight from dial-up to broadband: an answer that took a minute arrives in ten seconds, and because the model can search deeper in the same time, the quality of the answer rises as well.⁠[5][5][5][5][5]

He holds that because code used to be expensive to write, software engineers were very careful about what they wrote, which amounted to code rationing, and that teams even had people whose job was to say no; now that the marginal cost of code is approaching zero, professional engineers are shifting to implementing something, experiencing it, and then having it re-implemented differently in light of that experience.⁠[5][5][5]

He compares the change to the spread of literacy: once only scribes could read and write, then writing became simpler and education universal, everyone could read and write, and what counted was the quality of the writing. He expects that many people with good taste who know what good looks like, but without technical skills, will be able to build software to solve problems, and that individual founders without large teams will create very valuable companies; he said that if he started another company it would probably consist of AI.⁠[5][5][5]

On Pioneers of AI he said that he is less worried about AI taking over or taking jobs than about how people will find their own purpose and preserve their own agency when they no longer have to work: because decisions are hard, people will hand their decision-making authority to AI, which is wise for simple decisions, but the important decisions should still be made by people themselves. He said that Groq’s mission is to drive the cost of compute to zero so that everyone, and not just the wealthiest in society, has access, and that he worries about a concentration of compute power.⁠[11]

Asked in his 2024 conversation with Chamath Palihapitiya whether people should be afraid of AI, he drew a comparison with Galileo, who popularized the telescope: the better the telescope became, the more obvious it was that humanity was not the center of the universe. LLMs, he said, are “the telescope for the mind,” showing that intelligence is larger than we are, which makes people feel small and afraid; but just as people grew used to a larger universe and came to see its beauty, they will come to understand their place in intelligence and will not be afraid of it.⁠[3][3][3]

He holds that the information age trained people to answer questions, whereas with AI the answers can simply be asked for, and what a person has to do is think of the right question (7:13).⁠[5]

He cautions new founders that there are countless ways to be a leader, that copying other people’s advice does not work, and that one should choose the form that is natural to oneself (11:03).⁠[5]

His advice for a creative team is to set as few constraints as possible while stating clearly the goal that really matters: the fewer the constraints, the more freedom the team has to solve the problem and the more likely it is to surprise you (14:22).⁠[5]

To reduce politics in an organization, he advises against passing messages in separate one-on-one conversations and in favor of bringing everyone concerned together and saying it once; when someone emails a complaint about another person, that person should be copied in (17:38).⁠[5]

He says that the first principle of change management is to make people feel that nothing has changed: give them enough context to see that what they are trying to accomplish is the same as before (37:07).⁠[5]

He recommends saying “I intend to do this” in place of asking whether one should: asking for opinions invites pessimistic objections, stating an intention does not, and people will still speak up when there is a real problem (48:19).⁠[5]

He advises writing down what one is looking for in people, since otherwise hiring will not be consistent and will not find them (55:33).⁠[5]

On children’s education, his advice is to stop teaching them to answer questions and to teach them to ask questions; curricula should be built around problems that actually matter to the community, which students solve with the help of AI (1:10:31).⁠[5]

Google: Official website: https://about.google/ (opens in a new window).⁠[30]

Groq: Official website: https://groq.com/ (opens in a new window).⁠[31]

NVIDIA: Official website: https://www.nvidia.com/ (opens in a new window).⁠[23]

X: Public account: https://x.com/JonathanRoss321 (opens in a new window).⁠[12][32]

LinkedInLinkedInLinkedIn is a social networking platform for professionals. It was founded in 2002 by Reid Hoffman and others, launched in 2003, and listed on the New York Stock Exchange in 2011; after Microsoft acquired it for approximately $26.2 billion in 2016, it became a wholly owned Microsoft subsidiary. Its revenue comes from membership subscriptions, advertising, and recruitment solutions.Open the full entry: Public profile: https://www.linkedin.com/in/ross-jonathan/ (opens in a new window).⁠[12][14][32]