Jensen Huang, co-founder and CEO of Nvidia, displays the new Blackwell GPU chip during the Nvidia GPU Technology Conference in San Jose, California, on March 18, 2024.
David Paul Morris/Bloomberg via Getty Images
Nvidia CEO Jensen Huang is expected to reveal details about Rubin, the chipmaker’s next AI graphics processor, on Tuesday at the company’s annual GTC conference.
While other tech companies usually name their products using combinations of inscrutable letters and numbers, most of Nvidia’s most recent GPU architectures have been named after famous women scientists.
Nvidia is naming its next critical AI chip platform after Vera Rubin, an American astronomer.
The company has never explained its naming convention, and hasn’t emphasized the diversity aspect of its choices, but Nvidia’s chip names that highlight women and minority scientists are one of the most visible efforts to honor diversity in the tech industry during a period where diversity, equity and inclusion, or DEI, initiatives are being slashed by the Trump administration.
Rubin discovered a lot of what is known about “dark matter,” a form of matter that could make up a quarter of the matter of the universe and which doesn’t emit light or radiation, and she advocated for women in science throughout her career.
Nvidia has been naming its architectures after scientists since 1998, when its first chips were based on the company’s “Fahrenheit” microarchitecture. It’s part of the company’s culture – Nvidia used to sell an employee-only t-shirt with cartoons of several famous scientists on it.
It’s one of Nvidia’s quirks that has received more attention as it’s risen to become one of the three most-valuable tech companies and one of the most important suppliers to Google, Microsoft, Amazon, Openai, Tesla and Meta.
Investors want to hear on Tuesday how fast the Rubin chips will be, what configurations it will come in and when it might start shipping.
Before revealing a new architecture, Nvidia CEO Jensen Huang usually gives a one-sentence biography of the scientist it’s named after.
“I’d like to introduce you to a very, very big GPU named after David Blackwell, mathematician, game theorist, probability,” Huang said at last year’s GTC conference. “We thought it was a perfect name.”
Rubin is a fitting name for Nvidia’s next chip, which comes as the company tries to solidify the gains it has made in recent years as the leader in AI hardware. “Vera” will refer to Nvidia’s next-generation central processor, and “Rubin” will refer to Nvidia’s new GPU.
World famous astronomer Vera Rubin, 82, in her office at Carnegie Institution of Washington in Washington, DC on January 14, 2010.
Linda Davidson | The Washington Post | Getty Images
Born in Philadelphia in 1928, Rubin studied deep space and worked with other scientists to develop better telescopes and instruments that could collect more detailed data about the universe. In 1968, according to a Nova documentaryshe started observing the Andromeda galaxy and collecting the data that would upend science’s understanding of our universe.
Her primary claim to fame came after she observed how quickly galaxies rotate.
“The presumption was that the stars near the center of a galaxy would be orbiting very rapidly, and stars at the outside would be going very slowly,” Rubin said in 1987.
But Rubin realized that she was observing that outer stars were moving quickly, contrary to expectations. They weren’t flying out of orbit, which meant that there had to be more mass scientists weren’t observing — confirming the concept of dark matter.
She was acclaimed during her lifetime, published over 100 papers and held three advanced degrees, but she still faced discrimination because of her sex. Early in her career, Rubin wasn’t allowed to collect her own data, and some observatories didn’t allow women, according to the documentary.
Rubin died in 2016. In 2019, the Vera C. Rubin Observatory, a state-of-the-art telescope in Chile, was named after her. A biography on the federally-funded observatory’s website was edited to remove details about her advocacy for women in science earlier this year, according to ProPublica.
“I hope you will love your work as I love doing astronomy,” Rubin said at a commencement address in 1996. “I hope that you will fight injustice and discrimination in all its guises.”
Rubin isn’t the first woman to be honored with an Nvidia chip named after her.
Before Blackwell, who was the first Black American inducted into the National Academy of Sciences, Nvidia’s most advanced AI chip family was Hopper, named after American computer scientist Grace Hopper, who coined the term “bug” to refer to computer glitches. In 2022, Nvidia released its “Ada Lovelace” architecture, named after the British mathematician who pioneered computer algorithms in the 19th century.
The scientist names used to be a secondary naming convention, taking a back seat to the actual product name, and primarily appearing in marketing copy. Nvidia users more frequently referred to the “H100” chip or marketing names for consumer graphics cards like GeForce RTX 3090.
But last year, Huang emphasized that Blackwell wasn’t a single chip, it was a technology platform, and Nvidia increasingly started using the term “Blackwell” to refer to all of the company’s latest-generation AI products, such as its GB200 chip and DGX server racks.
Keeping momentum going
Now investors and analysts track how many “Hoppers” are in use, and big companies brag about having early access to “Blackwell.”
It’s critical for Nvidia that Rubin achieve the same last-name familiarity level as Hopper and Blackwell.
The company’s sales more than doubled in its fiscal 2025, ended January, to $124.62 billion, thanks to durable sales for the company’s Hopper chips and early demand for the company’s Blackwell chips.
In order to keep growth rising, Nvidia needs to deliver a next-generation chip that justifies its cost and improves on the previous generation’s speeds, power efficiency and cost of ownership.
The company has targeted 2026 for a rollout of the Vera chips, according to an investor presentation last fall. In addition to Vera Rubin, Nvidia is expected to discuss Blackwell Ultra, an updated version of its Blackwell chips that analysts expect the company to start selling later this year.
Huang also teased during an earnings call last month that he’ll show the “next click” after Vera Rubin. That architecture will likely be named after a scientist, too.
“These products should excite partners at the conference ranging from Microsoft to Dell to sovereigns, which normally would please investors,” Melius Research analyst Ben Reitzes wrote in a note on Monday.
Tuesday’s keynote will also be a test of Nvidia’s relatively new release cadence, where it strives to reveal new chips on an annual basis. Investors will also want to see whether Nvidia can continue to impress tech critics and developers while releasing new chip families on a faster schedule than it’s used to. Blackwell was announced last March, and its sales started showing up in Nvidia’s October quarter.
WATCH: Final Trades: Nvidia, Taiwan Semi, Amazon and the IEFA
Be the first to comment