Microsoft is primed to enjoy its next cycle of growth. On Wednesday, the company started selling the Microsoft 365 Copilot artificial intelligence add-on for its Office app subscriptions targeting businesses.
The feature that appears in Word, Excel and other Office programs will cost $30 per person per month. That can add up to over $10 billion in annualized revenue by 2026, Piper Sandler analysts Brent Bracelin and Hannah Rudoff wrote in a note to clients earlier this week.
Microsoft aims to make the most of its commanding lead in the productivity software market, where Google has been working to gain share. Google, meanwhile, is selling the Duet AI enhancement for subscriptions to its Workspace tools.
Piper Sandler’s model assumes that 18% of eligible users will use Copilot. That might be aggressive, but “there’s going to be a FOMO element to this,” Bracelin told CNBC in an interview on Tuesday, using the acronym for fear of missing out. “If you’re in an industry competing against someone that has Copilot and you don’t, you’re at a disadvantage.”
Piper Sandler has the equivalent of a buy rating on Microsoft shares, which are up 41% this year, compared with a gain of 9% for the S&P 500 index, of which it’s included.
“Customers tell us that once they use Copilot, they can’t imagine work without it,” Microsoft CEO Satya Nadella told analysts on a conference call last week.
After revealing plans for Copilot in March, Microsoft announced in September that it would first target the largest companies. On last week’s call, Nadella said 40% of the companies in the Fortune 100, a ranking of U.S. companies by revenue, were using Copilot in an invitation-only paid early-access program announced in May, calling out five clients by name: Bayer, KPMG, Mayo Clinic, Suncorp and Visa.
The preview was announced in May, less than six months ago. As a result, there isn’t a wealth of data on how Copilot affects performance.
“A lot of the conversations we’ve had even with the early-access customers is too short a timeframe to really look at the qualitative aspects of how they’re using the tools,” said Jason Wong, an analyst at technology industry researcher Gartner.
Companies need at least 300 licenses for employees to get access to Copilot. The challenge for Microsoft is to go beyond a small core of end users and land a wide deployment. That could take time.
Wong said Gartner encourages organizations to experiment with generative AI, which can create synthetic images and text with just a few words of human input.
“I think getting to 20% will be reasonable within two to three years for technologies like Copilot, because there’s going to be early adopters, and there’s going to be fast followers,” he said.
It might be easiest for companies to distribute Copilot to the mostly highly paid executives, whose time is precious, said Piper Sandler’s Bracelin. The tool could help them prioritize email messages and quickly understand documents.
But the top brass might end up causing headaches for tech support, Wong said. It might be wiser to first give Copilot to technically savvy employees who have drawn on generative AI for personal use and are familiar with shortcomings such as the potential to spout inaccurate information, Wong said. Microsoft acknowledges on its website that “the responses that generative AI produces aren’t guaranteed to be 100% factual.”
That hasn’t discouraged people from using ChatGPT, the chatbot from Microsoft-backed OpenAI whose language models are at the core of Copilot. After ChatGPT launched in November 2022, Microsoft and other large software companies moved quickly to incorporate similar generative features. Microsoft says prompts and responses in Copilot aren’t used to train language models and adhere to the company’s privacy standards.
Microsoft won’t only benefit from the new monthly Copilot fees. While setting up the tool, companies might end up using additional Azure cloud services, such as Purview for managing data, Wong said.
White House Senior Advisor Elon Musk walks to the White House after landing in Marine One on the South Lawn with U.S. President Donald Trump (not pictured) on March 9, 2025 in Washington, DC.
Samuel Corum | Getty Images News | Getty Images
Tesla shares fell in premarket trade on Monday after CEO Elon Musk announced plans to form a new political party.
The stock was down 7.13% by 4:27 a.m. E.T.
Musk said over the weekend that the party would be called the “America Party” and could focus “on just 2 or 3 Senate seats and 8 to 10 House districts.” He suggested this would be “enough to serve as the deciding vote on contentious laws, ensuring that they serve the true will of the people.”
Now tech billionaire’s reinvolvement in the political arena is making investors nervous.
“Very simply Musk diving deeper into politics and now trying to take on the Beltway establishment is exactly the opposite direction that Tesla investors/shareholders want him to take during this crucial period for the Tesla story,” Dan Ives, global head of technology research at Wedbush Securities, said in a note on Sunday.
“While the core Musk supporters will back Musk at every turn no matter what, there is broader sense of exhaustion from many Tesla investors that Musk keeps heading down the political track.”
Musk’s previous political foray earned him Trump’s praise in the early days, but he has since drawn the ire of the U.S. president.
The two have clashed over various areas of policy, including Trump’s spending bill which Musk has said would increase America’s debt burden. Musk has taken issue to particular cuts to tax credits and support for solar and wind energy and electric vehicles.
Trump on Sunday called Musk’s move to form a political party “ridiculous,” adding that the Tesla boss had gone “completely off the rails.”
Musk is contending with more than just political turmoil. Tesla reported a 14% year-on-year decline in car deliveries in the second quarter, missing expectations. The company is facing rising competition, especially in its key market, China.
Jonathan Ross, chief executive officer of Groq Inc., during the GenAI Summit in San Francisco, California, US, on Thursday, May 30, 2024.
David Paul | Bloomberg | Getty Images
Artificial intelligence semiconductor startup Groq announced Monday it has established its first data center in Europe as it steps up its international expansion.
Groq, which is backed by investment arms of Samsung and Cisco, said the data center will be located in Helsinki, Finland and is in partnership with Equinix.
Groq is looking to take advantage of rising demand for AI services in Europe following other U.S. firms which have also ramped up investment in the region. The Nordics in particular is a popular location for the data facilities as the region has easy access to renewable energy and cooler climates. Last month, Nvidia CEO Jensen Huang was in Europe and signed several infrastructure deals, including data centers.
Groq, which is valued at $2.8 billion, designs a chip that the company calls a language processing unit (LPU). It is designed for inferencing rather training. Inferencing is when a pre-trained AI model interprets live data to come up with a result, much like the answers that are produced by popular chatbots.
While Nvidia has a stranglehold on the chips required for training huge AI models with its graphics processing units (GPUs), there is a swathe of startups hoping to take a slice of the pie when it comes to inferencing. SambaNova; Ampere, a company SoftBank is in the process of purchasing; Cerebras and Fractile, are all looking to join the AI inference race.
European politicians have been pushing the notion of sovereign AI — where data centers must be located in the region. Data centers that are located closer to users also help improve the speed of services.
Global data center builder Equinix connects different cloud providers together, such as Amazon Web Services and Google Cloud, making it easier for businesses to have multiple vendors. Groq’s LPUs will be installed inside the Equinix data center allowing businesses to access Groq’s inference capabilities via Equinix.
Groq currently has data centers in the U.S. and Canada and Saudi Arabia with its technology.
Don’t miss Groq CEO Jonathan Ross on Squawk Box Europe at 7:45 a.m. London time.
Hidden among the majestic canyons of the Utah desert, about 7 miles from the nearest town, is a small research facility meant to prepare humans for life on Mars.
The Mars Society, a nonprofit organization that runs the Mars Desert Research Station, or MDRS, invited CNBC to shadow one of its analog crews on a recent mission.
“MDRS is the best analog astronaut environment,” said Urban Koi, who served as health and safety officer for Crew 315. “The terrain is extremely similar to the Mars terrain and the protocols, research, science and engineering that occurs here is very similar to what we would do if we were to travel to Mars.”
SpaceX CEO and Mars advocate Elon Musk has said his company can get humans to Mars as early as 2029.
The 5-person Crew 315 spent two weeks living at the research station following the same procedures that they would on Mars.
David Laude, who served as the crew’s commander, described a typical day.
“So we all gather around by 7 a.m. around a common table in the upper deck and we have breakfast,” he said. “Around 8:00 we have our first meeting of the day where we plan out the day. And then in the morning, we usually have an EVA of two or three people and usually another one in the afternoon.”
An EVA refers to extravehicular activity. In NASA speak, EVAs refer to spacewalks, when astronauts leave the pressurized space station and must wear spacesuits to survive in space.
“I think the most challenging thing about these analog missions is just getting into a rhythm. … Although here the risk is lower, on Mars performing those daily tasks are what keeps us alive,” said Michael Andrews, the engineer for Crew 315.
