Tech
Nvidia campus in Silicon Valley
(Getty Images)
MORE WITH LESS

Nvidia added only 6,400 new employees to its workforce last year

The company grew its revenues by 114% last year, while adding only 20% to its headcount.

Whether AI brings about the golden era of productivity that many tech leaders endlessly talk about — enabling all of us to do more with less — is still worth arguing about. But the main character of the AI boom itself, Nvidia, is certainly finding ways to squeeze a little more juice out of its employees, with the company’s Q4 earnings revealing not only blockbuster revenues and profits but a remarkable fact: the world’s hottest company employs only 36,000 people, about 20% more than it did last year.

Considering that Nvidia more than doubled its revenue in 2024, from ~$61 billion to an eye-watering ~$130 billion, the fact that the responsibilities of its HR department have grown only 20% is nothing short of remarkable. But, when you look back at the last few years, the divergence between employees and revenue is even more stark.

Nvidia headcount
Sherwood News

Clearly, the kind of highly specialized work that Nvidia chip designers spend their days on isn’t one of those problems that’s aided by simply throwing more cooks into the AI mixer. Indeed, the company reports that 75% of its employees work on “Research & Development,” while less than a quarter are in more generic corporate functions like sales, marketing, or operations.

Last year, Bloomberg reported some employees claimed to be working seven days week in a pressure cooker environment, and CEO Jensen Huang has said previously that he doesn’t fire people — he would rather “torture employees to greatness” than fire them. (He did also add that he was being “tongue in cheek” about the torturing.) But, when looking solely at financial metrics, it’s hard to argue that Huang’s approach isn’t working.

But just how remarkable is Nvidia’s employee efficiency compared to its Big Tech peers?

tech

Nvidia makes $3.6 million of revenue per employee — more than any of its Big Tech rivals

Among the $17 trillion BATMMAAN tech giants (Big Tech’s version of the Avengers), Nvidia is squeezing the most out of every employee.

After reporting blowout earnings, which the market found reasons to dislike anyway as Nvidia tumbled 8% in trading on Thursday, we can now calculate that last year Nvidia pulled in $3.6 million in revenue for each one of its employees.

That’s more than 1.5x that of Apple and Meta, and nearly double that of Alphabet. The gap is even wider when it comes to profit: Nvidia’s net income per employee sits at $2 million, surpassing Apple, Meta, Alphabet, and Amazon.

Of course, this isn’t entirely surprising given that Nvidia had the lowest number of employees among the BATMMAAN group, with its workforce heavy on research and development.

Contrast that with Amazon and Tesla — the least tech-y of the eight behemoths — which rank at the bottom in revenue and profit per employee, and it makes sense. After all, Amazon is the world’s second-largest retailer behind Walmart, relying on a massive workforce to keep its logistics engine running — hence its 1.6 million employees, nearly twice as many as the rest combined. Meanwhile, Tesla, despite its software-heavy ambitions, is still a capital-intensive car company.

Golden handcuffs

Keeping those employees motivated to work long hours, and be tortured into greatness, could be a problem for Nvidia, but it’s one that’s been solved predominantly by the company’s astonishing stock surge. Nvidia leaping more than 1,800% over the last five years has been great news to say the least for its leather-jacket-loving CEO Jensen Huang, its suppliers, institutional shareholders, independent individual investors like Nancy Pelosi, and, of course, Nvidia employees.

With equity grants and employee stock purchase programs par for the course in Silicon Valley, it’s not hard to conclude that many of the workers that joined the high-flying chipmaker five or more years ago are probably millionaires today — assuming that they held on long enough to see the share price soar.

Last year, unsubstantiated reports that 76% of Nvidia employees were millionaires flooded social media. The true number is hard to guess, but the impact of Nvidia’s golden handcuffs are clear to see in its official turnover statistics. And while some employees may not enjoy the intense environment, the company’s low turnover suggests a workforce that’s more than happy to show up every day. The company’s 2024 Sustainability Report reveals that the company’s overall turnover rate was just 2.7% last year — a fraction of the semiconductor industry average of 17.7%.

Put another way:

A 2.7% turnover rate, if maintained, would equate to an average tenure of 37 years at the company.

Chances are that the low turnover rate won’t hold up forever. Some employees will presumably eventually decide that they have enough money to retire, and of course the stock could drop if the insatiable Big Tech capex orgy ever slows down or rivals build GPUs able to compete with Nvidia. But, for now, if you’re an Nvidia employee working seven days a week with a bunch of stock waiting to vest over the next few years, those golden handcuffs are still looking pretty tight.

Go Deeper: Gaming was once Nvidia’s golden goose. Now it’s the most low-key $11 billion business you can imagine.

More Tech

See all Tech
tech

OpenAI rolls out age prediction system ahead of allowing adult content

OpenAI is rolling out a new “age prediction” feature for ChatGPT users.

The company will look at various signals from users to predict if a user is underage.

In a blog post, the company said:

“The model looks at a combination of behavioral and account-level signals, including how long an account has existed, typical times of day when someone is active, usage patterns over time, and a user’s stated age.”

If the system suspects the user is a minor, it will reduce content with graphic violence, harmful viral challenges, sexual or romantic role play, depictions of self-harm, and material promoting “extreme beauty standards, unhealthy dieting, or body shaming.”

If a user is incorrectly flagged as under 18, they will have to submit a selfie to an identity verification service to have the restrictions removed.

An age verification system is part of OpenAI’s plan to reduce harmful mental health encounters with the chatbot, while also allowing ChatGPT to generate “erotica” in the near future.

“The model looks at a combination of behavioral and account-level signals, including how long an account has existed, typical times of day when someone is active, usage patterns over time, and a user’s stated age.”

If the system suspects the user is a minor, it will reduce content with graphic violence, harmful viral challenges, sexual or romantic role play, depictions of self-harm, and material promoting “extreme beauty standards, unhealthy dieting, or body shaming.”

If a user is incorrectly flagged as under 18, they will have to submit a selfie to an identity verification service to have the restrictions removed.

An age verification system is part of OpenAI’s plan to reduce harmful mental health encounters with the chatbot, while also allowing ChatGPT to generate “erotica” in the near future.

tech

Google’s YouTube maintains its top spot as streaming accounts for nearly half of all TV-watching time

People spent a record 47.5% of their TV-watching time on streaming platforms in December, according to new data from Nielsen, up from the previous record of 47.3% in July. Google’s YouTube once again was the most popular streaming service by time spent, but Netflix’s share inched slightly upward to 9% from 8.8% in July, while YouTube’s fell to 12.7% from 13.4%. The jump was largely thanks to Stranger Things, which was the most watched streaming title last month.

tech

Amazon CEO says tariffs are inflating prices and buyers are looking for bargains

While the legality of President Trump’s tariffs winds its way through the courts, their effects are beginning to show up in prices.

During an interview at the World Economic Forum in Davos, Switzerland, Amazon CEO Andy Jassy said he is starting to see tariffs “creep into” pricing, as some sellers are “passing on those higher costs to consumers in the form of higher prices.”

Jassy said that while consumers are still spending, they are becoming more price conscious.

“I think that wherever they can, they are trying to trade down in price — they are looking for bargains wherever they can find bargains,” he said. “I see people a little more hesitant on higher-priced discretionary items.”

Trump has maintained that other countries are footing the bill for his tariffs. But new research suggests Americans will ultimately be the ones paying those higher prices.

Jassy said that while consumers are still spending, they are becoming more price conscious.

“I think that wherever they can, they are trying to trade down in price — they are looking for bargains wherever they can find bargains,” he said. “I see people a little more hesitant on higher-priced discretionary items.”

Trump has maintained that other countries are footing the bill for his tariffs. But new research suggests Americans will ultimately be the ones paying those higher prices.

tech

Musk: Tesla restarting Dojo supercomputer effort as “AI5 chip design is in good shape”

Tesla CEO Elon Musk said in a post on X over the weekend that the company plans to restart work on its Dojo supercomputer, dubbed Dojo3, saying that the AI5 chip the company had been developing is in “good shape.”

The Dojo supercomputer trains Tesla’s AI models, including the one behind its all-important Full Self-Driving tech. The company stopped work on Dojo in August. “It doesn’t make sense for Tesla to divide its resources and scale two quite different AI chip designs,” Musk said at the time. “The Tesla AI5, AI6 and subsequent chips will be excellent for inference and at least pretty good for training.”

“Pretty good” appears to be good enough.

In the interim, Tesla relied more on companies like Nvidia and Advanced Micro Devices for AI training. Restarting Dojo suggests Tesla plans to bring at least some AI training back in-house.

Musk also runs AI company xAI, which has its own supercomputer and a substantial business relationship with Tesla. A plurality of Tesla shareholders recently voted in favor of investing in Musk’s AI company, but the board declined to approve the measure because of a large number of abstentions.

Latest Stories

Sherwood Media, LLC produces fresh and unique perspectives on topical financial news and is a fully owned subsidiary of Robinhood Markets, Inc., and any views expressed here do not necessarily reflect the views of any other Robinhood affiliate, including Robinhood Markets, Inc., Robinhood Financial LLC, Robinhood Securities, LLC, Robinhood Crypto, LLC, or Robinhood Money, LLC.