Nvidia vice president says GPUs are the 'currency' of AI researchers
In the world of AI research, the speed of development is limited in large part by available computing resources, according to Jonathan Cohen, Vice President of Applied Research at Nvidia.
"These days, the currency in any AI researcher is how many GPUs they get access to, and that's no less true at Nvidia than at any other company," Cohen said in an interview on Nvidia Developer.
Cohen led the team responsible for developing Nvidia's Llama Nemotron family of models. Released in March of this year, they represent the company's entry into the world of " reasoning" AI systems.
The speed at which the models came together was remarkable, Cohen said, taking "no more than one to two months." He partially credits the efficiency of their development to other workers being willing to sacrifice their processing power.
"So, there were a lot of researchers who very selflessly agreed to give up their compute so that we could get these Llama Nemotron models trained as quickly as we did," he said.
Cohen also attributed the speed of development to Nvidia's company-wide culture of prioritizing major projects, regardless of current team goals.
"How do you have a team to do a thing you've never done before? Part of the corporate culture is — we call them a 'swarm' — where you identify, 'This is something that's important,'" he said. "And everyone, every manager who has people who might be able to contribute, thinks about, 'Is this new thing more important than the current thing everyone on my team is doing?'"
If the manager can spare anybody, they'll "contribute" their direct reports to the new priority.
"Llama Nemotron ended up being a very cross-discipline, cross-team effort," Cohen added. "We had people from across the whole company working together without any formal organizational structure."
Llama Nemotron required a series of sacrifices, Cohen said, both in terms of computing power and personnel — but people were able to set aside self-interests for the benefit of the whole.
"It was really great to see, great leadership," he said. "There were a lot of sacrifices that people made, a lot of very egoless decisions that brought it together, which is just awesome."
Nvidia did not respond to a request for comment by Business Insider immediately prior to publication.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
an hour ago
- Yahoo
Why Wall Street's Dr. Doom now wants to be Dr. Boom
Nouriel Roubini, who's been known as "Dr. Doom" for 17 years, is feeling more upbeat. The economist has scaled back his recession call and thinks the US is headed for an investment boom. He told BI there are three things that have driven his newfound optimism. Wall Street has been calling him "Dr. Doom" for 17 years, but Nouriel Roubini — the economist famous for his persistently bearish and frequently dystopian takes on the world economy — is sounding surprisingly positive lately. He's rescinded his earlier call for a recession, and now sees a US tech and artificial intelligence investment boom unfolding that will uplift the economy through the rest of this decade. By 2030, Roubini thinks economic growth in the US will double from around 2% to 4%, while productivity growth surges from around 1.9% to 3%. The stock market is also likely to climb higher, he told Business Insider in an interview, predicting the S&P 500 would see high single-digit percentage growth in 2025, on par with its historical return. It's a sharp turnaround from the gloomy forecasts he' is known for. Roubini told BI the nickname started to stick in 2008, when the New York Times referred to him as "Dr. Doom" after he correctly called the Great Financial Crisis, he told BI. "Even before, I always said I'm not Dr. Doom and I'm Dr. Realist, first of all," Roubini said. He said that he's made numerous forecasts that were more bullish than the consensus throughout the years when the evidence lines up. "So I don't know why people think that I'm always Dr. Doom. It's not the case." His outlook, though, has brightened considerably since 2022. Back then, he appeared on TV and penned op-eds warning of a coming stagflationary debt crisis. At the time, he described the turmoil he saw looming as an all-in-one financial crisis involving spiraling debt levels, soaring inflation, and a severe recession. Roubini told BI there are a few things that have gotten him to change his tune. Roubini says he began to hear the murmurs of the AI revolution well before ChatGPT went viral at the end of 2022. In his 2022 book, "Megathreats," he acknowledged the potential for artificial intelligence to significantly boost economic growth and serve as a major tailwind for markets. That's become a reality way faster than Roubini expected, and a major reason he's become more bullish, he told BI. He believes the economy could start to reap the growth and productivity benefits of AI in the next several years, particularly as humanoid robots enter the mainstream. A breakthrough in fusion energy would be another bullish force for the economy, Roubini said. Fusion energy hasn't been achieved yet, but tech firms are pouring vast sums of money into making it happen. Chevron and Google contributed to a more than $150 million funding round this week for TAE Technologies, a fusion energy company that plans to have a working prototype power plant by the early 2030s. Type One Energy, another fusion energy firm, also plans to roll out a power plant by the middle of the next decade. "We're not in an AI winter anymore. We had the fusion winter for 40 years. We're not anymore," Roubini said, pointing to the stagnation in tech and fusion energy development is the past. "Now it's happening." President Donald Trump's tariffs may not be as harmful to the US economy as some investors think, Roubini says. He thinks it's more likely that markets will throw a tantrum and force Trump to walk back his most aggressive policies. That's already happened a few times this year. Roubini pointed to sharp sell-offs in the bond market that preceded Trump's 90-day pause of his "Liberation Day" tariffs, and the softening of his tone regarding firing Jerome Powell. "That means the bond vigilantes are the most powerful people in the world," Roubini said. "The instincts might be very bad, but then, markets are unforgiving," he added of policymakers. Roubini speculates that tariffs on China, for instance, could wind up somewhere around 39%, well-below the 145% tariff rate Trump proposed earlier in the year. Meanwhile, AI, quantum computing, and other tech advancements in the US can more than offset the impact of the trade war, Roubini said. Tariffs are expected to drag down GDP growth by 0.06% a year through 2035, according to estimates from the Congressional Budget Office. It's a fraction of the 2 percentage point increase in growth Roubini expects to see by the end of the decade. Roubini now pegs the odds of a recession to just around 25%. Even if the US enters a downturn this year, Roubini says he expects it to be shallow and short, as the Fed can cut interest rates to boost the economy, while tech powers growth over the long-run. That's not to say Dr. Doom has shed all of his bearish views. Roubini says many of the things he feared several years ago — stagflation, spiraling government debt levels, and rising geopolitical conflict — still loom. He rattled off a list of potential risks the US could conceivably face in the future: migration controls fueling stagflation in the economy, the US dollar collapsing in value, and China and the US not reaching a trade agreement and seeing an escalating cold war, to name a few scenarios. "So there's plenty of stuff in the world that can go wrong," he said. Read the original article on Business Insider
Yahoo
an hour ago
- Yahoo
Why Costco Stock (COST) May Be a Smarter Investment Than Big Tech
Costco (COST) has long been a market favorite, but its current valuation has even the most bullish investors raising eyebrows as the company trades near all-time highs. Trading at earnings multiples that outpace some of the world's biggest tech giants, it's fair to ask: how can a warehouse selling pallets of toilet paper and rotisserie chickens possibly justify such a steep premium? At first glance, the numbers don't add up. Comparing Costco to tech companies indicates a rather inflated valuation for the wholesaler. Easily unpack a company's performance with TipRanks' new KPI Data for smart investment decisions Receive undervalued, market resilient stocks right to your inbox with TipRanks' Smart Value Newsletter Timid growth projections, razor-thin margins, and a business model that hasn't changed much in decades don't exactly scream innovation. And yet, Costco continues to climb—largely, in my view, due to one key factor: risk perception. Because the company's business model is so heavily anchored in predictable, recurring revenue from membership fees, its earnings stream is exceptionally stable. That stability arguably justifies the stock's lofty valuation—and forms the backbone of my bullish thesis, even if there's little to no margin of safety. I agree with Costco's critics—those who are skeptical about a warehouse full of assorted goods trading at a forward P/E of 58, a multiple literally higher than tech giant and AI trailblazer Nvidia. For a company like Costco, which has grown revenue at a ~10% CAGR over the past five years and operating income by 15%—yet is only expected by analysts to grow revenue at a 5% CAGR over the next five—it really doesn't make much sense, at first glance, for the business to trade at such a steep premium. But here's what a lot of market participants miss when analyzing companies: valuations based on net earnings are heavily anchored to one key factor—risk. And unlike earnings, there's no single, universally accepted measure of risk. Some investors focus on beta, while others consider regulatory threats or macroeconomic exposure. Ultimately, each investor views the situation through their own lens. That's where Costco stands out as a rare case: a recurring revenue model outside of the tech sector. Customers pay an annual membership fee—ranging from $60 to $120 in the U.S.—just to shop in its warehouses. Costco likely converts nearly all of that directly into operating profit (though the company doesn't explicitly disclose this), which allows the business to be highly profitable even while selling goods at razor-thin margins. In the most recent quarter, membership income rose 10% year-over-year, with a staggering 93% renewal rate in North America. That kind of brand loyalty and perceived value creates a business model with very little risk of earnings volatility. And because of that, Costco functions like a defensive stock, offering low risk and a high degree of predictability in returns. So, while the valuation may look stretched on the surface, this level of stability and cash flow certainty arguably justifies a premium multiple. While only about 2% of Costco's total revenue comes from membership fees, those fees account for more than half of the company's total operating income. At this kind of scale, there's simply no other retailer that can pull off a high-volume, low-margin model quite like the Issaquah, Washington-based warehouse giant. Main Street Data indicates that Costco's revenues are inching higher while membership fees remain a staple. That's why, even during periods of weak consumer spending, Costco tends to remain highly profitable—not just because of its membership model, but because it consistently gains market share from competitors. By keeping margins razor-thin on discretionary items, and with shoppers naturally shifting toward essentials during times of financial stress, Costco puts pressure on rival retailers that simply can't match its pricing without cutting into their own profitability. Take Q3, for example: Costco's comparable sales rose 8% year-over-year to $62 billion, while peers like Target (TGT) saw comparable sales shrink and were forced to revise earnings guidance downward. This highlights how Costco isn't just weathering macroeconomic headwinds—it's actually growing through them. In other words, macro risks seem to have a much lighter impact on a company like Costco. While Costco's relatively low-risk business model may warrant a valuation premium, it's worth exploring what an appropriate stock price might look like under simplified assumptions. For this analysis, I'll estimate an equity value for Costco using a basic perpetuity-based discounted cash flow (DCF) model. Let's begin with the fact that, in fiscal year 2024, Costco generated approximately $4.8 billion in membership fee revenue. While this isn't equivalent to free cash flow, it serves as a reasonable proxy for recurring operating cash flow due to its stable and high-margin nature. With Costco's current market capitalization at approximately $470 billion, we can reverse-engineer the valuation. If we assume a 6% discount rate—reflecting the company's relatively low risk profile despite the 10-year Treasury yield sitting around 4.4%—and a perpetual growth rate of 5%, the implied equity value is approximately $480 billion, or about $1,083 per share. This closely aligns with Costco's current share price. These assumptions appear fairly reasonable and suggest that Costco's valuation is justified under optimistic, yet not unrealistic, conditions. However, using a more conservative 7% discount rate, while keeping other assumptions constant, would cut the implied equity value significantly, roughly in half. In summary, while Costco does not appear overvalued under favorable scenarios, the current valuation leaves little room for error. A meaningful margin of safety may only emerge if macroeconomic conditions shift, such as a recession prompting lower risk-free rates and, in turn, greater tolerance for elevated valuation multiples. Out of the 25 analysts who have covered COST in the past three months, 17 rate the stock as a Buy, while the remaining eight call it a Hold. So, while the overall sentiment is moderately bullish, the average price target of $1,096.36 implies ~8.5% upside from the current share price over the next twelve months. Given the strength and stability of Costco's underlying business model, its premium valuation multiples appear justifiable. At present, there are no significant concerns that would challenge a constructive outlook on the stock. However, it's important to note that the current valuation offers limited—if any—margin of safety. As such, I maintain a cautious Buy rating on Costco. While near-term upside may be constrained, the company remains well-positioned to deliver steady, long-term compound growth. Disclaimer & DisclosureReport an Issue
Yahoo
2 hours ago
- Yahoo
Prediction: This Artificial Intelligence (AI) Stock Could Hit a $6 Trillion Valuation by 2030
Nvidia is currently the world's largest company by market cap. The semiconductor giant still has a lot of room for growth, considering the potential investment in artificial intelligence (AI) infrastructure over the next five years. Nvidia has additional catalysts coming into play that could allow it to sustain terrific growth and hit a $6 trillion valuation in the future. 10 stocks we like better than Nvidia › Nvidia (NASDAQ: NVDA) is the most valuable company in the world as of this writing, with a market cap of $3.4 trillion, and it has reached this position thanks to a tremendous rally of more than 1,500% in its stock price in the past five years. Investors will now be wondering if Nvidia has the potential to deliver more upside over the next five years following this phenomenal run. However, don't be surprised to see this semiconductor giant's shares jump higher and attain a $6 trillion valuation by the end of the decade. Let's take a closer look at the factors that could help Nvidia hit that milestone by 2030. Robust demand for Nvidia's AI chips has been the biggest reason behind the stock's terrific surge in recent years. The good part is that Nvidia continues to generate a massive amount of revenue from its AI chip business despite tackling headwinds such as export restrictions to key markets like China. This was evident from Nvidia's latest results for the first quarter of fiscal 2026 (which ended on April 27). The company's revenue shot up 69% year over year to $44.1 billion during the quarter, even though it lost $2.5 billion in revenue in fiscal Q1, owing to the restrictions on sales of its chips to China. The chipmaker also incurred a $4.5 billion inventory charge to write down the value of chips that were intended for the Chinese market. Moreover, Nvidia's fiscal Q2 revenue would take an $8 billion hit on account of the China-related restrictions. But the good part is that the company's guidance for the current quarter still calls for a 50% year-over-year increase in revenue, while its earnings are expected to increase by 44% despite anticipated loss in Chinese revenue. CEO Jensen Huang admitted on Nvidia's latest earnings conference call that the $50 billion Chinese market is now effectively closed to U.S. players such as Nvidia. Even then, analysts have increased their revenue estimates. That's not surprising, as Nvidia still has a massive sales opportunity in AI chips beyond the Chinese market. It is now entering new markets such as Saudi Arabia to build AI factories "powered by several hundred thousand of Nvidia's most advanced GPUs over the next five years," according to the company. Additionally, massive AI infrastructure projects such as Stargate, from which Nvidia has started benefiting already, could help it mitigate the lost opportunity in China. Management consulting firm McKinsey & Company predicts that AI-capable data centers could require investments worth a whopping $5.2 trillion by 2030 to build enough computing power to handle training and inference workloads. So investors would do well to look past the China-related problems that Nvidia is currently facing, as the broader opportunity in the AI data center market should be lucrative enough to help the chipmaker keep growing at a healthy pace for the next five years. Moreover, Nvidia is showing no signs of losing its grip over the AI chip market. Its data center revenue shot up an impressive 73% year over year to $39 billion in the previous quarter. That was miles ahead of Broadcom's $4 billion AI revenue and AMD's $3.7 billion data center sales in the previous quarter, the two chipmakers that are considered to be the closest to Nvidia in the AI chip race. Nvidia's data center growth was higher than the 57% growth recorded by AMD in this segment and close to the 77% growth in Broadcom's AI revenue, even though it has a much larger revenue base. This is a testament to just how popular Nvidia's AI chips are, with the company's latest generation of Blackwell processors already a major hit among cloud computing giants within two quarters of hitting the market. Even better, Nvidia has moved past just selling AI hardware. It also offers access to models that help customers train and deploy AI agents, along with other enterprise AI applications that allow customers to improve the efficiency of their large language models (LLMs). Its enterprise platforms are gaining traction in diverse industries such as cybersecurity and restaurants where companies are deploying Nvidia's solutions to streamline their operations or to build agentic AI applications. All this indicates that investors shouldn't miss the forest for the trees, as Nvidia's long-term prospects aren't dependent on just China. There is still a lot of room for growth in the AI chip market, and the company's diversification into other areas such as enterprise AI applications and automotive should be enough to power remarkable growth over the next five years. Nvidia is currently trading at 23 times sales. While that's three times the U.S. technology sector's average price-to-sales ratio, the company's dominant position in AI chips, the prospects of this market, and the other catalysts that are coming into play help justify that valuation. We have already seen in the chart that Nvidia's top line is expected to jump to $292 billion in three years. If it maintains its sales multiple at that time, it will be able to easily surpass a $6 trillion valuation in just three years, representing a big jump from current levels. However, if we assume Nvidia's top-line growth slows after fiscal 2028 to an annual rate of 15% in fiscal 2029 and 2030 from the 31% compound annual growth rate that it is forecast to clock between fiscal 2026 and 2028 (using fiscal 2025 revenue of $130.5 billion as the base), its annual revenue could jump to $386 billion after five years. If Nvidia trades at a discounted 15 times sales at that time, it could still hit a $6 trillion valuation by 2030. So, investors can still consider buying this AI stock in anticipation of more upside in the long run, as it seems capable of maintaining its healthy growth rate over the next five years. Before you buy stock in Nvidia, consider this: The Motley Fool Stock Advisor analyst team just identified what they believe are the for investors to buy now… and Nvidia wasn't one of them. The 10 stocks that made the cut could produce monster returns in the coming years. Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $669,517!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $868,615!* Now, it's worth noting Stock Advisor's total average return is 792% — a market-crushing outperformance compared to 171% for the S&P 500. Don't miss out on the latest top 10 list, available when you join . See the 10 stocks » *Stock Advisor returns as of June 2, 2025 Harsh Chauhan has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Advanced Micro Devices and Nvidia. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy. Prediction: This Artificial Intelligence (AI) Stock Could Hit a $6 Trillion Valuation by 2030 was originally published by The Motley Fool Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
