For years, Amazon has set the bar for package delivery. When Prime launched in 2005, two-day shipping was unheard of. By 2019, one-day shipping was standard for millions of items. Now, the retail giant is turning to generative AI to drive more same-day shipping.
Amazon is using the technology to optimize delivery routes, make more intelligent warehouse robots, create more-ergonomic environments for employees and better predict where to stock new items, said Steve Armato, Amazon’s vice president of transportation technology and services.
During an exclusive tour of Amazon’s largest California sort center, located in Tracy, Armato told CNBC that 60% of Prime orders in March were delivered the same day or next day in the top 60 metropolitan areas in the U.S. Amazon is betting on generative AI to increase that figure.
“It seems subtle, but at this scale, getting just one more product in the right spot means that it’s shipping less distance when you order it,” Armato said in an interview at the warehouse.
In 2020, Amazon began developing models for demand forecasting and supply chain optimization using transformer architecture, the backbones of what we know today as generative AI.
“Generative AI is the next big evolution in technology,” Armato said. “It’s remarkable, and we’re already applying it in very practical ways across our operations.”
But not all the changes that generative AI may bring to the e-commerce giant are positive. There are concerns about the high-energy needs of generative AI and about its ability to enable robots to replace Amazon’s human workforce, analysts told CNBC.
Robots and new roles
The number of Amazon warehouse robots grew from 350,000 in 2021 to more than 750,000 in 2023, according to the company.
Amazon began adding AI transformer models to its warehouse delivery robots in 2022 so the machines can dash around each other more intelligently. CNBC watched hundreds of them move in a coordinated grid in the warehouse. Armato calls this “the dance floor.”
“Some of the two-day deliveries might stand aside, let the robot with a next-day delivery go on its mission first and take a straight line to its destination,” Armato said.
Hundreds of robots dash around each other with the help of generative AI at Amazon’s largest California sort center in Tracy, California, July 31, 2024.
Lisa Setyon
While these robots navigate using a series of QR codes, Amazon’s next generation of drive units, called Proteus, are fully autonomous, the company said.
“They’re using generative AI and computer vision to avoid obstacles and find the right place to stop,” Armato said.
As part of the company’s AI strategy, Amazon in August struck a deal with AI startup Covariant. Amazon hired the startup’s founders and licensed its models that help robots handle a wider range of physical objects. Amazon is also developing a bipedal robot called Digit that can grasp and handle items in a humanoid way.
CNBC saw a row of 20 robotic “Robin” arms that use computer vision to determine how much pressure to use when picking up various package shapes and sizes. Amazon said generative AI teaches the arms how to handle products they’ve never seen before based on data from similar products in Amazon’s vast catalog.
A similar model is used to better assess damaged items and keep them from shipping out. Amazon’s AI is three times better at identifying damaged products than humans are, the company said.
Introducing more robotics with generative AI without replacing human workers is a balancing act for Amazon, said Tom Forte, senior equity analyst at the Maxim Group.
“How can they implement automation to improve efficiency and manage labor expenses, but how can they do it in a way that complements their use of humans and doesn’t replace them?” Forte said.
Rather than replacing workers, the robots are reducing the burden on employees and creating new roles, Armato said. Amazon said it plans to spend $1.2 billion to upskill more than 300,000 employees by the end of 2025 as generative AI and robotics change the company’s processes.
“Someone needs to maintain [the robot] if it breaks down,” Armato said. “Or if something does get dropped on the dance floor, we have a process and special training to go clean that up. And so each of those creates new categories of jobs, some of which have higher earnings potential.”
Many of Amazon’s robots move tall bins of items to workstations where employees pick and pack them, which reduces how much humans have to walk, Armato said. AI is also reducing the need for workers to reach and bend, he said.
“One algorithmic improvement is to take our faster-selling products and place those on the shelves at waist height,” Armato said. “That’s your ergonomic power zone.”
Robotic drive units bring tall stacks of items to workstations for picking and packing at an Amazon same-day center in Richmond, California, Aug. 31, 2024.
Katie Tarasov
Predicting orders and routes
With all those robots and workers, Amazon delivered more than 2 billion items the same day or next day in the first quarter of 2024, according to the company.
Amazon has always used algorithms to predict how much of what inventory is needed, when and where. The company said it’s using generative AI to predict where best to place items it hasn’t previously sold.
“When we place a product in the right place ahead of time, before you click buy, it’s traveling less distance, which is a win for speed and sustainability,” Armato said.
Amazon Web Services has data centers filled with servers running AI workloads that give the company an edge over its retail rivals because it can train its AI in-house. As an early online-only retailer, Amazon got a head start on collecting mass aggregate data on shopping behavior and delivery logistics. Amazon is now using that trove of data to create AI models for use in everything from supply chain optimization to warehouse robotics, according to the company.
“It’s not that Walmart and Target and Costco and others don’t have their own reams of data, but they’re looking at things a little bit differently, and they have much older systems,” said Sucharita Kodali, retail analyst at Forrester Research.
How eco-friendly generative AI will be in the long run is unclear. That’s because training and running generative AI is a carbon-intensive process, and by 2027, AI servers worldwide are projected to use as much power every year as Sweden or the Netherlands.
That’s in conflict with Amazon’s 2019 commitment to reach net-zero carbon by 2040.
The company claims that the use of AI is helping cut down the carbon footprint of package delivery. Amazon is reducing carbon by using more than 20 machine learning models to improve mapping for its vast network of 390,000 delivery drivers, predicting road closures and choosing more efficient routes, the company said.
Beyond its warehouses, Amazon has also introduced generative AI to help its sellers and shoppers.
Additionally, Amazon said, it has invested $4 billion in AI startup Anthropic, which makes chatbot Claude, a competitor to OpenAI’s ChatGPT. Amazon also makes its own AI-focused microchips and its own generative AI tools for developers, which it also uses in operations, the company said.
Whether Amazon’s huge investment in generative AI will translate to profits remains an open question.
“I have yet to see huge lift in anybody’s retail business due to generative AI, including Amazon,” Kodali said. “I think a lot of their biggest impact has happened because of the earlier investments, not necessarily some of these more recent investments.”
From left, Veza founders Rob Whitcher, Tarun Thakur and Maohua Lu.
Veza
Tech giants like Google, Amazon, Microsoft and Nvidia have captured headlines in recent years for their massive investments in artificial intelligence startups like OpenAI and Anthropic.
But when it comes to corporate investing by tech companies, cloud software vendors are getting aggressive as well. And in some cases they’re banding together.
Veza, whose software helps companies manage the various internal technologies that employees can access, has just raised $108 million in a financing round that included participation from software vendors Atlassian, Snowflake and Workday.
New Enterprise Associates led the round, which values Veza at just over $800 million, including the fresh capital.
For two years, Snowflake’s managers have used Veza to check who has read and write access, Harsha Kapre, director of the data analytics software company’s venture group told CNBC. It sits alongside a host of other cloud solutions the company uses.
“We have Workday, we have Salesforce — we have all these things,” Kapre said. “What Veza really unlocks for us is understanding who has access and determining who should have access.”
Kapre said that “over-provisioning,” or allowing too many people access to too much stuff, “raises the odds of an attack, because there’s just a lot of stuff that no one is even paying attention to.”
With Veza, administrators can check which employees and automated accounts have authorization to see corporate data, while managing policies for new hires and departures. Managers can approve or reject existing permissions in the software.
Veza says it has built hooks into more than 250 technologies, including Snowflake.
The funding lands at a challenging time for traditional venture firms. Since inflation started soaring in late 2021 and was followed by rising interest rates, startup exits have cooled dramatically, meaning venture firms are struggling to generate returns.
Wall Street was banking on a revival in the initial public offering market with President Donald Trump’s return to the White House, but the president’s sweeping tariff proposals led several companies to delay their offerings.
That all means startup investors have to preserve their cash as well.
In the first quarter, venture firms made 7,551 deals, down from more than 11,000 in the same quarter a year ago, according to a report from researcher PitchBook.
Corporate venture operates differently as the capital comes from the parent company and many investments are strategic, not just about generating financial returns.
Atlassian’s standard agreement asks that portfolio companies disclose each quarter the percentage of a startup’s customers that integrate with Atlassian. Snowflake looks at how much extra product consumption of its own technology occurs as a result of its startup investments, Kapre said, adding that the company has increased its pace of deal-making in the past year.
‘Sleeping industry’
Within the tech startup world, Veza is also in a relatively advantageous spot, because the proliferation of cyberattacks has lifted the importance of next-generation security software.
Veza’s technology runs across a variety of security areas tied to identity and access. In access management, Microsoft is the leader, and Okta is the challenger. Veza isn’t directly competing there, and is instead focused on visibility, an area where other players in and around the space lack technology, said Brian Guthrie, an analyst at Gartner.
Tarun Thakur, Veza’s co-founder and CEO, said his company’s software has become a key part of the ecosystem as other security vendors have started seeing permissions and entitlements as a place to gain broad access to corporate networks.
“We have woken up a sleeping industry,” Thakur, who helped start the company in 2020, said in an interview.
Thakur’s home in Los Gatos, California, doubles as headquarters for the startup, which employs 200 people. It isn’t disclosing revenue figures but says sales more than doubled in the fiscal year that ended in January. Customers include AMD, CrowdStrike and Intuit.
Guthrie said enterprises started recognizing that they needed stronger visibility about two years ago.
“I think it’s because of the number of identities,” he said. Companies realized they had an audit problem or “an account that got compromised,” Guthrie said.
AI agents create a new challenge. Last week Microsoft published a report that advised organizations to figure out the proper ratio of agents to humans.
Veza is building enhancements to enable richer support for agent identities, Thakur said. The new funding will also help Veza expand in the U.S. government and internationally and build more integrations, he said.
Peter Lenke, head of Atlassian’s venture arm, said his company isn’t yet a paying Veza client.
“There’s always potential down the road,” he said. Lenke said he heard about Veza from another investor well before the new round and decided to pursue a stake when the opportunity arose.
Lenke said that startups benefit from Atlassian investments because the company “has a large footprint” inside of enterprises.
“I think there’s a great symbiotic match there,” he said.
Arvind Krishna, chief executive officer of International Business Machines Corp. (IBM), during a Bloomberg Television interview at the World Governments Summit in Dubai, United Arab Emirates, on Tuesday, Feb. 11, 2025.
“We have been focused on American jobs and manufacturing since our founding 114 years ago, and with this investment and manufacturing commitment we are ensuring that IBM remains the epicenter of the world’s most advanced computing and AI capabilities,” IBM CEO Arvind Krishna said in a release.
The company’s announcement comes weeks after President Donald Trump unveiled a far-reaching and aggressive “reciprocal” tariff policy to boost manufacturing in the U.S. As of late April, Trump has exempted chips, as well as smartphones, computers, and other tech devices and components, from the tariffs.
IBM said its investment will help accelerate America’s role as a global leader in computing and fuel the economy. The company said it operates the “world’s largest fleet of quantum computer systems,” and will continue to build and assemble them in the U.S., according to the release.
Read more CNBC tech news
IBM competitor Nvidia, the chipmaker that has been the primary benefactor of the artificial intelligence boom, announced a similar push earlier this month to produce its NVIDIA AI supercomputers entirely in the U.S.
Nvidia plans to produce up to $500 billion of AI infrastructure in the U.S. via its manufacturing partnerships over the next four years.
Last week, IBM reported better-than-expected first-quarter results. The company said it generated $14.54 billion in revenue for the period, above the $14.4 billion expected by analysts. IBM’s net income narrowed to $1.06 billion, or $1.12 per share, from $1.61 billion, or $1.72 per share, in the same quarter a year ago.
IBM’s infrastructure division, which includes mainframe computers, posted $2.89 billion in revenue for the quarter, beating expectations of $2.76 billion.
The company announced a new z17 AI mainframe earlier this month.
Meta CEO Mark Zuckerberg looks on before the luncheon on the inauguration day of U.S. President Donald Trump’s second Presidential term in Washington, U.S., Jan. 20, 2025.
Evelyn Hockstein | Reuters
Mark Zuckerberg’s plan is to make Meta the market leader in artificial intelligence. Investors will want to know how President Donald Trump’s tariffs-heavy trade policies will impact that strategy.
Those answers could start to come as soon as this week as Meta’s AI strategy takes center stage when the company hosts its first Llama-branded conference for AI developers on Tuesday then reports its latest quarterly earnings the next day.
Already, tech companies are starting to talk about the potential impact they’re bracing for as a result of the Trump tariffs.
Intel Chief Financial Officer David Zinsner said Thursday during the chip giant’s first-quarter earnings call that U.S. trade policies “have increased the chance of an economic slowdown, with the probability of a recession growing.” Meanwhile, Google CFO Anat Ashkenazi said that day during a first-quarter earnings call that the tech giant remains committed to its $75 billion investment in capital expenditures, or capex, this year, but also acknowledged that the “timing of deliveries and construction schedules” could cause some quarter-to-quarter spending fluctuation.
For now, analysts expect Meta to follow Alphabet’s lead and remain firm in its plan to spend as much as $65 billion in capex for AI infrastructure this year when it reports earnings Wednesday. Some analysts believe Meta could even raise the figure because AI is a core priority for the company.
“We do not expect META to cut its CapX guidance of $60B-$65B in 2025, for its GenAI infrastructure, because they see this as an important 10-year investment, we believe,” Needham analysts wrote in a research note published Wednesday. “However, tariffs add risks of upward cost revisions.”
Investors will also be monitoring Meta’s LlamaCon event at its Menlo Park, California, headquarters for any signs that its AI investments are having an immediate business impact. This will be the first time Meta hosts a developer conference specifically for its Llama family of AI models.
“Investors want to see ROI on all these AI investments, and while Meta has shown clear benefits from leveraging AI to improve its products and drive faster revenue growth, it’s been hard to quantify those benefits,” Truist Securities analyst Youssef Squali told CNBC.
Meta in April released a couple of its new Llama 4 models, which Meta Chief Product Officer Chris Cox previously said can help power so-called AI agents that can perform tasks for users via web browsers and other online interfaces.
It’s critical that Meta keep improving Llama to create a major business involving AI agents that companies can use to interact with their customers within apps like Facebook and WhatsApp, William Blair research analyst Ralph Schackart said.
“Meta has an early mover advantage at scale in a multi-trillion dollar market,” Schackart said in an email. “We believe Meta is very well positioned to leverage its billions of global users across multiple platforms.”
Meta is unlikely to curb its Llama investment any time soon, but should eventually consider doing so if it fails to generates enough money to justify its costs, said Ken Gawrelski, a Wells Fargo managing director of equity research.
“We do believe that over time Meta needs to continue to evaluate whether Llama needs to be competitive with the leading-edge models,” Gawrelski said. “This is a very expensive proposition and thus far, unlike Google, Meta does not directly monetize its model in any material way.”
Chris Cox, Chief Product Officer at Meta Platforms, speaks during The Wall Street Journal’s WSJ Tech Live Conference in Laguna Beach, California on October 17, 2023.
Patrick T. Fallon | AFP | Getty Images
Meta AI and the consumer
Analysts are also following the Meta AI digital assistant. That’s because the ChatGPT rival represents the second pillar of Zuckerberg‘s AI strategy.
Zuckerberg in January said he believes 2025 “is going to be the year when a highly intelligent and personalized AI assistant reaches more than 1 billion people, and I expect Meta AI to be that leading AI assistant.”
In February, CNBC reported that Meta was planning to debut a standalone Meta AI app during the second quarter and test a paid subscription service, in which users could pay monthly fees to access more powerful versions like users can with ChatGPT.
Although Meta’s enormous user base across its family of apps gives Meta AI an advantage over rivals like ChatGPT in terms of reach, they may not interact with Meta AI in the same way they do with rival chat apps, said Cantor Fitzgerald analyst Deepak Mathivanan.
Gawrelski said that people may not want to use Meta AI within Facebook and Instagram if all they want to do is passively watch the short videos that Meta algorithmically recommends to their feeds.
“This is why a separate Meta AI, where Meta could clearly articulate its use case and value proposition, could be helpful,” Gawrelski said.
A standalone Meta AI app could help the company better market the digital assistant and distinguish it from rivals, said Debra Aho Williamson, founder and chief analyst for Sonata Insights.
“ChatGPT has such wide brand awareness, that it’s become a moat that is soon going to be very hard to overcome,” Williamson said.