Amazon on Thursday revealed its first chip for quantum computing, and said its design will help the company build highly efficient hardware systems.
The processor is called Ocelot, and the announcement comes as more tech companies tout their advancements in quantum. Last week, Amazon cloud rival Microsoft showed off its inaugural quantum chip. Microsoft had a paper in the journal Nature documenting its quantum work, and this week Amazon followed suit.
Some technologists hope quantum computers will be capable of solving problems that stump classical computers. PCs and phones run calculations and store data with bits that are either on or off, while quantum computers work with quantum bits, or qubits, that can operate in both states simultaneously.
“We believe that scaling Ocelot to a full-fledged quantum computer capable of transformative societal impact would requireas little as one-tenth as many resources as common approaches, helping bring closer the age of practical quantum computing,” Fernando Brandão, Amazon Web Services’ director of applied science, and Oskar Painter, the cloud group’s quantum hardware chief, wrote in a blog post.
The U.S. Defense Advanced Research Projects Agency (DARPA) has funded quantum computing research for two decades, but the technology has been slow to make its way to consumers and businesses.
“That’s because they’re not big enough yet,” said Peter Barrett, founder and general partner at Playground Global, which has backed quantum startups Phasecraft and PsiQuantum.
At a million qubits, there are enough bits that the technology will work even if there are some problems, Barrett said. Google’s Willow, the world’s top quantum chip, features just 105, while Amazon’s Ocelot has only nine, Painter told CNBC.
Amazon CEO Andy Jassy said in 2020, when he was head of AWS, that the company was “optimistic in the future that quantum computing will play a role” as cloud gets bigger in large companies and the public sector.
Six months after Jassy made those comments, AWS released the Amazon Braket service, allowing developers to experiment with quantum computers from other companies, including IonQ and Rigetti Computing. Microsoft’s Azure cloud has a similar offering. Amazon is planning for its in-house quantum chip to become available through Braket, Painter said.
In 2023, AWS senior vice president Peter DeSantis talked about building a quantum processor at the cloud group’s Reinvent conference in Las Vegas, promising more details in the future.
Like Microsoft, Amazon fabricated its chip internally. Building a system boasting a million qubits will take collaborations with world-leading semiconductor manufacturers, according to Barrett. Outsourcing to a partner is an option as Amazon progresses with quantum hardware, Painter said.
Public interest in the space has risen lately, Painter said, as companies have discussed new ways of assembling qubits that are resistant to errors. Amazon designed Ocelot to tackle the problem of error correction, and Google’s Willow also demonstrated improvements in that area, Painter said.
Painter estimated that commercial workloads won’t be running on quantum computers for 10 years or more.
At a meeting with analysts in January, Nvidia CEO Jensen Huang said useful quantum computers could be 15 to 30 years away. Days later, Meta chief Mark Zuckerberg told Joe Rogan that he isn’t a quantum computing expert but said most people believe it’s at least a decade from being viable.
Pat Gelsinger, Intel’s former CEO, is more optimistic.
“I stand by my prediction years ago — by 2030, useful quantum computing,” Gelsinger wrote in a LinkedIn comment on Wednesday.
Standard Chartered’s bullish crypto analyst still sees bitcoin’s price hitting $500,000 during Donald Trump’s presidency — even after a selloff that sank the world’s largest digital currency to a three-month low.
Geoffrey Kendrick, who heads up digital assets research at Standard Chartered, told CNBC he believes bitcoin will hit the $200,000 mark this year before climbing even further in the coming years.
“Within the crypto ecosystem, what we need are traditional financial players, like Standard Chartered, like BlackRock and others that have the ETFs now to really step in,” Kendrick said in an interview with CNBC’s “Squawk Box Europe” Thursday.
“As the industry becomes more institutionalized, it should be safer,” Kendrick said, adding that this should result in fewer negative headlines — such as the recent $1.5 billion hack on cryptocurrency exchange Bybit last week.
This increase in crypto adoption by institutions, coupled with some “regulatory clarity” in the U.S., should lead to less volatility over time, he added.
“That should add to that medium term, top-side potential, which for me is bitcoin up to $200,000 this year, and $500,000 before Trump leaves office,” Kendrick told CNBC.
Kendrick said the catalyst necessary for large financial institutions to gain confidence to invest in bitcoin and other crypto assets is a stabilization in prices and increased regulatory clarity.
Bitcoin earlier this week sank to a three-month low below $90,000 amid declines in global equity markets. As of Thursday, the token was trading at $86,418. That means it’s down about 20% from an all-time high of $108,786, which the coin peaked at in January, according to CoinGecko data.
Standard Chartered’s Kendrick said digital currencies have dropped more broadly due to uncertainty around tariffs and resolutions to major wars such as Russia-Ukraine and Israel-Gaza.
“Risk assets don’t like uncertainty, and so that’s what we’ve seen. We’ve seen tech stocks in the U.S. coming lower,” Kendrick said, adding that the breach of Bybit has also contributed to negative sentiment surrounding crypto more broadly.
He expects the outlook for crypto will improve later in the year as traders await key regulatory developments in the industry, such as new rules around stablecoins and anti-money laundering.
“That should further legitimize, so you’ll see more U.S. banks involved. You’ll see larger institutions in the U.S. continue to push through,” Kendrick said.
Kendrick was one of the numerous market analysts who predicted a doubling in bitcoin’s price this year to $200,000. Bitcoin broke the highly anticipated $100,000 mark in December following Trump’s election to the U.S. presidency.
Crypto bulls view Trump positively given his support for digital currencies. In January, Trump signed an executive order promoting the advancement of cryptocurrencies in the U.S. and developing a national digital asset stockpile.
The company reported adjusted earnings of 30 cents per share on $987 million in revenue, surpassing the 17 cents per share and $956 million in sales expected by analysts polled by LSEG. That reflected 27% year-over-year revenue growth.
“We see tremendous opportunities ahead to support our customers throughout their end-to-end data lifecycle, and we are laser-focused on delivering on this vision,” said CEO Sridhar Ramaswamy in a press release. He called Snowflake the “most consequential data and AI company in the world.”
Like its peers, Snowflake has pushed to offer new artificial intelligence tools to its customers as the race for advanced large language models and AI capabilities accelerates. It announced an expanded partnership with Microsoft Azure to offer access to OpenAI models on Wednesday.
Product revenue also topped analyst estimates, growing 28% to $943 million. That came in ahead of the roughly $914 million LSEG estimate. The company also said it anticipates $4.28 billion in product revenue for the year, ahead of a $4.21 billion estimate.
Read more CNBC tech news
Guidance for the current quarter, however, came up short of estimates. Snowflake said it expects product revenues to range between $955 million to $961 million, versus a StreetAccount estimate of $961 million.
Goldman Sachs analyst Kash Rangan said the results further boosted the firm’s confidence in the revenue add from new products in the second half of the fiscal year for Snowflake and he views the company as set to become a long-term generative AI winner.
“By expanding the reach and accessibility of its core data platform to more avenues such as [large language models], Hyperscalers, etc., Snowflake can become core to the development of AI applications, evidenced by 4,000+ accounts using Snowflake AI/ML and Cortex AI’s early momentum,” he wrote.
Snowflake said it had 11,159 customers during the period, up from 10,618. Analysts polled by FactSet had expected 10,987. The company also said that Chief Financial Officer Michael Scarpelli will retire, but remain in the role until a successor is found.
Excluding Thursday’s premarket moves, shares are up about 8% year to date.
Stripe announced a tender offer for employees and shareholders on Thursday that values the payments startup at $91.5 billion, the closest the company has been to its peak valuation of $95 billion in 2021.
“We very much care about providing good liquidity for employees and existing shareholders,” Stripe co-founder and President John Collison told CNBC’s Andrew Ross Sorkin in an interview on “Squawk Box.”
As for the company’s long-awaited public market debut, Collison said, “We are not dogmatic on the public vs. private question,” and “have no near-term IPO plans.”
Stripe also revealed in its annual letter on Thursday that it generated $1.4 trillion in total payment volume in 2024, up 38% from the year prior. The company said it was profitable in 2024, and expects to remain so this year.
Collison said the business can’t be managed on a “super tight quarterly EPS basis because this growth tends to come in waves.”
Collison said the artificial intelligence boom has been key to the company’s recent growth. High-profile AI startups OpenAI, Anthropic, Perplexity and Mistral are all Stripe clients.
“Unlike maybe previous booms that were more speculative in nature where you had asset price speculation, here we are seeing an AI boom that is very real,” Collison said. “There’s a bunch of companies that have grown and grown and grown over the past few years, but they’ve grown because they have real revenue and they have real revenue because they have customers that find their products really useful.”
More than 700 AI agent startups launched on Stripe last year, according to the company’s annual letter. Collison said a future where agents will make purchases for human customers is inevitable.
Founded in 2010, Stripe has regularly conducted tender offers to allow early investors and employees to sell a portion of their equity in order to reduce the pressure to go public. A year ago the company announced a tender offer at a $65 billion valuation.