OpenAI generated $10 billion in revenue for Microsoft in the quarter ended Sept. 30, about 11% of the Redmond company’s total revenue, according to its quarterly filing Wednesday.

The amount is up sharply. In the fiscal year that ended in June, Microsoft recorded $24.1 billion in revenue from its business with OpenAI, or about $6 billion a quarter on average.

That $10 billion surpasses some of Microsoft’s longest-running businesses. Xbox brought in $5.4 billion in the quarter. Search and advertising, which includes its Bing search engine, generated $6.6 billion. Windows licenses sold through PC makers, along with Surface and other Microsoft devices, brought in $3.6 billion.

The total covers revenue-sharing payments from OpenAI, as well as other business between the companies, such as OpenAI’s use of Azure to train and run its AI models. Microsoft didn’t say how much came from each, or how much of Azure’s growth came from OpenAI.

Azure revenue rose 48% to $33.2 billion in the quarter, beating Wall Street’s expectations and the company’s own forecast of 44% to 45% growth.

Microsoft has begun disclosing its revenue from OpenAI because its stake of about 24% makes OpenAI a “related party” under accounting rules, requiring it to report the business they do with each other.

Microsoft first reported the number in its annual report in July, as a full-year total. This is the first time it has given a quarterly figure.

The original AI alliance: Microsoft first invested in OpenAI in 2019, eventually committing $13 billion, and became its exclusive cloud provider. ChatGPT ran on Azure when it launched in late 2022.

The companies have loosened their ties since then. OpenAI restructured as a public benefit corporation in October 2025, and Microsoft gave up its right of first refusal to supply OpenAI’s computing capacity. OpenAI committed to buy an additional $250 billion in Azure services. An April 2026 agreement gave OpenAI broader freedom to use other cloud providers.

OpenAI has also agreed to keep sharing revenue with Microsoft through 2030.

In addition, Microsoft sells access to OpenAI’s models to business customers through its Azure OpenAI Service, and the models became the foundation for its Copilot products, putting Microsoft at the front of the generative AI boom. That revenue comes from Microsoft’s own customers, so it isn’t part of the $10 billion.

Microsoft, for its part, now builds its own AI models and offers models from Anthropic and others.

Many customers, not a few: Asked by Bernstein Research analyst Mark Moerdler whether AI growth on Azure was coming from big businesses or AI-native companies, Microsoft CEO Satya Nadella didn’t mention OpenAI.

He said he doesn’t want Azure’s AI business to depend on a few big customers, and that broad adoption by businesses is just getting started.

Nadella described it as “a decade-long thing,” and contrasted it with relying on “four customers who buy a lot of capacity from you” to show a few quarters of growth. That, he said, is “not a business as far as I’m concerned.”

He added that Microsoft has large customers who want to buy capacity, “and as long as the pricing is right, we may sell them some. If not, we will be patient.”

Later in the call, Goldman Sachs analyst Gabriela Borges brought up the fact that OpenAI made up 11% of revenue. Nadella didn’t address the number but praised OpenAI’s latest models and said customers are increasingly mixing models from different companies, or training their own.

He cited Microsoft’s tools for routing each task to whichever AI model handles it best. If you use AI coding agents, Nadella said, you want Anthropic and OpenAI “to compete for your workload, not like once a quarter, but for every task.”

New risk language: Microsoft also expanded a warning in its quarterly filing about the risks of depending on outside AI providers. The company builds products such as Copilot on AI models from outside companies, including OpenAI and Anthropic. It now says it has “limited visibility into or control over” how those models are designed, trained, secured and tested.

Without naming OpenAI or Anthropic, the same section notes that some of these AI partners are also “significant customers of Azure,” and that the benefits Microsoft expects from those relationships “may not be realized or sustained.” It says those partners could change how much they buy, and that Microsoft itself may change how it allocates scarce data center capacity or prices it.

Microsoft is an investor in both companies. According to its latest filing, it owns about 24% of OpenAI, down from about 25% in June and about 27% after OpenAI’s restructuring a year ago, and has put $12.0 billion of a promised $13.0 billion into the company.

In addition, Microsoft agreed last year to invest up to $5 billion in Anthropic, and its results for the June quarter included a $3.2 billion gain on that investment. Microsoft hasn’t disclosed the size of its Anthropic stake, and Anthropic’s IPO filing earlier this week didn’t specify it either.

Update, 10/28: Microsoft’s Azure cloud business brought in $33.2 billion in revenue in the September quarter, up 48% from a year earlier, blowing past the company’s own forecast and beating Wall Street’s expectations. Read the full story.

Microsoft’s earnings report Wednesday will come two days after the company offered buyouts to about 60,000 people, roughly a quarter of its employees. The call gives executives their first chance to say what the offer will cost, and to explain how it fits with the company’s long-term view of its workforce.

Whether or not they go into those details, it’s one of several issues to watch in Wednesday’s report, the first under Microsoft’s new reporting structure. Anthropic’s IPO filing on Monday adds another storyline, showing how much the Claude maker has agreed to spend on Microsoft’s cloud.

Coming into the report, Microsoft’s stock is near its high. Microsoft closed Monday at $532.60, about 2% below its record close of $543.45 on Oct. 13, after gaining roughly $1.4 trillion in market value since late June.

Here are the key numbers and issues heading into Microsoft’s results for the first quarter of its 2027 fiscal year, ended Sept. 30.

Core numbers: Analysts expect revenue of $90.75 billion, up nearly 17% from a year ago, and earnings of $4.72 per share, up 14%, according to Yahoo Finance. Microsoft’s own forecast was $89.85 billion to $90.95 billion, so Wall Street is looking for a result near the top of that range.

In July, Microsoft reported $90 billion in revenue for the June quarter, well above the $87.7 billion analysts expected.

A new scorecard: Microsoft announced in September that it would sort its business into two groups instead of the three it had used since 2015. This is the first quarter under the new setup.

The first, dubbed Agents and Infra, includes Azure, Microsoft 365, Dynamics, LinkedIn’s recruiting and sales tools, and server and productivity licensing. It brought in 81% of Microsoft’s revenue and 88% of its operating income in fiscal 2026, based on figures Microsoft recast under the new structure.

Microsoft expects Agents and Infra revenue of $75.15 billion to $75.75 billion this quarter, which would be growth of about 22% to 23%, by GeekWire’s calculation.

The other group, Devices and Consumer, includes Windows, Xbox and search advertising. Microsoft expects it to bring in $14.7 billion to $15.2 billion, down 5% to 8% from a year earlier.

Azure, in dollars: Microsoft started disclosing Azure’s annual revenue last year, when it passed $75 billion, and topped $100 billion in fiscal 2026. Starting this quarter, it reports the figure every quarter. It also changed what counts as Azure: GitHub’s cloud business and Security Copilot now fall under Microsoft 365, and its healthcare cloud moved to a new industry unit.

Counted that way, Azure brought in $29.4 billion in the June quarter, up 42%. Microsoft expects growth of 44% to 45% this quarter, not counting currency fluctuations, which would put Azure at roughly $32 billion, by GeekWire’s calculation. That compares with the roughly 45% Microsoft projected in July, under the old definition.

Hitting that forecast may not be enough. Investors are looking for about 46%, BNP Paribas analyst Stefan Slowinski wrote in a note Monday. In January, Azure beat Microsoft’s own forecast and the stock still fell 10%, because investors had privately expected more.

“Azure moves the stock, but the read that matters is whether capacity is absorbed as fast as it lands,” wrote Dan Ives of Yorkville Ives in a note last week.

Capital spending: Microsoft expects capital expenditures of more than $50 billion for the quarter, which would be a record. That’s up from $41 billion in the June quarter and $34.9 billion a year earlier.

An accounting change reduces what Microsoft reports as capital spending, without reducing its planned investment. Microsoft now assumes its data center buildings will last 25 years instead of 15. As a result, more of its data center leases will be classified in a way that doesn’t count as capital spending, CFO Amy Hood said on the July earnings call.

That change alone lowered Microsoft’s capital spending estimate for calendar 2026 to about $175 billion, from the roughly $190 billion it projected in April. Hood said the company’s actual spending plans hadn’t changed.

Cloud margins: Microsoft Cloud gross margin came in at 65% last quarter, better than expected but down from a year earlier. Microsoft said it expects the figure to be relatively stable this quarter. It also expects operating margins to be roughly flat with a year ago.

Workforce: The new “voluntary exit” offer, announced Monday by Chief People Officer Amy Coleman, covers participating teams in the U.S., Canada, the U.K. and India. It has no age or tenure requirement and extends to some vice presidents. Employees have until the first week of December to decide, and Microsoft hasn’t said how much the offer will cost or when it will record the charge.

For the spring retirement program, Hood disclosed the cost on the April earnings call: about $900 million, with roughly $350 million in cost of revenue and $550 million in operating expenses. The charge showed up in the following quarter. About 30% of the 8,750 eligible employees accepted.

The new offer is less generous. It matches Microsoft’s standard severance, without the extended healthcare of the spring program, and the company expects a much smaller share of people to take it. But the pool of eligible employees is nearly seven times larger.

Microsoft says the offer doesn’t mean it needs fewer people. Coleman said in July that the company was studying whether to make voluntary exits a regular part of managing its workforce. It has told investors to expect operating margins roughly flat this quarter and down less than a point for the full fiscal year, even with record capital spending.

Nadella has addressed this before. In a July 2025 memo to employees, after Microsoft laid off about 15,000 people, he acknowledged the “uncertainty and seeming incongruence” of thriving financially while cutting jobs. “This is the enigma of success in an industry that has no franchise value,” he wrote.

In his annual shareholder letter last week, he focused on Microsoft’s spending on AI infrastructure. “Based on the demand we see and the efficiencies we are driving, I am confident in the long-term returns on our investments,” Nadella wrote. “We are building against a growing book of business tied to this secular shift.”

Microsoft had about 223,000 employees as of June 30, before it cut 4,800 jobs in July and more in September. The Information reported this month that Azure President Girish Bablani told leaders in his group to find ways to cut costs by at least 20%. Microsoft called the report inaccurate and said Azure leadership had not directed teams to develop cost-cutting plans.

Copilot: Microsoft said in July that businesses were paying for more than 30 million Microsoft 365 Copilot seats, up from 20 million in April. That’s still only about 7% of the more than 450 million commercial Microsoft 365 seats.

The bigger change is in how Microsoft charges for Copilot and its other AI tools. In July it added usage-based billing alongside per-seat pricing. It’s also pushing customers toward E7, a new premium bundle that includes Copilot. Microsoft expects Microsoft 365 commercial cloud revenue to grow about 17% this quarter, not counting currency swings, a number that now includes GitHub.

The BNP Paribas analyst Slowinski, who attended Microsoft’s Frontier Tour event in London last week, wrote that partners are seeing more customers roll out Copilot company-wide. Partners also told him that Microsoft now ties 35% of E7 sales incentives to how much customers use features like Cowork.

Anthropic and OpenAI: Anthropic’s S-1 filing shows the company has committed to spend at least $44.9 billion with Microsoft from November 2026 to May 2033, up from the $30 billion Azure commitment announced last November. That’s well below its minimum commitments to Amazon ($110 billion) and Google ($114.5 billion).

Microsoft agreed last year to invest up to $5 billion in Anthropic. Its results for the June quarter included a $3.2 billion gain on that investment.

OpenAI remains the bigger factor. Microsoft’s commercial remaining performance obligation (RPO), the measure of signed contracts it hasn’t yet delivered on, reached $678 billion last quarter, up 84% from a year earlier. Without OpenAI, the backlog grew 25%. OpenAI told investors its annualized revenue approached $50 billion at the end of September, The Information reported, lower than earlier reports of about $70 billion.

Windows: Microsoft expects revenue from Windows licenses sold to PC makers, along with its own devices, to fall a little more than 20%. Higher memory prices are making PCs more expensive, and PC makers are still working through extra inventory.

What’s next: Alphabet also reports Wednesday afternoon, and Amazon reports Thursday, setting up a side-by-side look at Azure, Google Cloud and AWS.

Anthropic is aiming to go public the week of Nov. 16, according to the Wall Street Journal.

Check back Wednesday afternoon for coverage.

The Washington Technology Industry Association kicked off the third annual Seattle AI Week on Monday with a reminder that AI affects everyone, from engineers and artists to “someone who learned about ChatGPT today,” as WTIA’s Nick Ellingson put it.

“All of us should be a part of that conversation,” he added.

A few people in the audience took that statement literally.

Members of Troublemakers, a Seattle climate activism group, interrupted the week’s opening panel on AI and economic growth several times before they were led out by security. Outside on Elliott Bay, others floated in kayaks with a banner: “Power for People, Not for Tech.”

“These are not your constituents — we are,” one protester told Beau Perschbacher, Gov. Bob Ferguson’s senior policy advisor, as he addressed the large crowd of tech and startup leaders at Seattle’s AI House. Perschbacher was on a panel with Joe Nguyen of the Seattle Metropolitan Chamber of Commerce and Mattie Zazueta of OpenAI, the event’s presenting sponsor.

The group’s main message, according to members who spoke with GeekWire: the rush to build data centers and roll out AI is putting the state’s communities, jobs and environment at risk, and Washington should be investing in green jobs and clean energy instead.

The protest comes amid a broader backlash against data centers in Washington state and across the country. The Seattle City Council voted unanimously in June to pause new large data centers in the city for a year, and Washington AI Resistance organized at least 19 protests around the state this month calling for a statewide moratorium.

The event also came the same day Microsoft offered voluntary buyouts to about 60,000 employees, roughly a quarter of its workforce, as it spends tens of billions of dollars on data centers and AI infrastructure.

The backlash contrasts with a push from the business and tech community for Washington state and the Seattle region to do more to stay competitive in the AI era.

“Washington helped build the technology economy of the last generation. We have an extraordinary opportunity to lead the AI economy of the next,” said former Gov. Chris Gregoire, now CEO of the business group Challenge Seattle, in her opening remarks at the AI Week event.

Gregoire cited efforts to put AI to work on the region’s own problems, including an AI permitting pilot in Everett that she said could cut the median wait for an accessory dwelling unit permit from 252 days to 90. She highlighted the AI-driven protein design work of Nobel laureate David Baker of the University of Washington, with applications in areas including the environment, agriculture and health care.

She also warned that Washington is close to the bottom for its business climate, even though it ranks second in North America for tech talent, as other states court its companies and jobs.

On the panel, Perschbacher said Ferguson’s administration has been focused on data centers since taking office, setting up a work group early last year to study their impact. The group looked for a Washington community that had been harmed by a data center and couldn’t find one, he said.

He added that the state won’t permit data centers to run diesel generators around the clock, and that their water use turned out to be “not out of line with a typical industrial user.”

Electricity is a different story. Data centers “use a really significant amount of electricity,” he said, and with demand also rising from EVs and heat pumps, getting enough power “is a major challenge.”

Perschbacher addressed the protesters directly, saying Washington has some of the strictest environmental laws in the country and that the state’s data center work group reinforced those rules rather than weakening them. Bills on ratepayer protections and transparency for data centers are likely to come back in the next legislative session, he said.

The issue also came up in the event’s second panel, on AI policy. State Sen. Marko Liias, who sponsored a bill this year on high-risk uses of AI, said data centers that want to locate in Washington need to meet “stringent expectations,” including adding clean energy to the grid, recycling water and dropping nondisclosure agreements that keep communities in the dark. Washington’s goal “has never been to be the cheapest or easiest place,” he said.

“When energy becomes scarce, I do think we need to be clear that our communities deserve access to that energy before the data centers do,” Liias said.

Joe Braidwood, founder of AI governance startup Glacis, which works out of AI House, said on the same panel that it was the first event he’d attended there with a public protest. “I think that’s actually kind of promising that people care enough to want to come together and talk and disagree and figure this out,” he said.

Zazueta, OpenAI’s global head of community programs, said in the opening panel that the industry needs to do a better job showing people the benefits of AI. Many people “read about these scary headlines,” she said, but don’t see how the technology applies to their daily lives.

OpenAI runs AI literacy workshops and trains libraries and community groups to lead their own, she said. The company opened an expanded office in Bellevue in March, its largest engineering center outside San Francisco.

Speaking with GeekWire outside the event on Monday afternoon, one of the protesters who had interrupted the panel singled out that remark. She said it dismissed public concerns about AI as the product of bad headlines, and she objected to the idea of using community groups to win people over to the benefits of artificial intelligence.

“They want to infiltrate our communities,” she said.

The opening panel with Zazueta, Perschbacher and Nguyen was titled “The Abundance State: Washington’s Chance to Lead the AI Future.” The protester objected to that framing, saying the panelists’ idea of abundance came down to easing permits for data centers.

It wasn’t the first time the group has protested the tech industry’s energy use. In 2024, it blocked entrances at Amazon’s Seattle headquarters to protest the company’s plans to power data centers in Oregon with natural gas.

Inside the event on Monday, as the interruptions continued, Nguyen, a former state senator who now leads the Seattle chamber, had responded to the protesters directly. “If you want to have your own event, have your own event,” he told them.

After they were led out, Nguyen said he agreed with the protesters, calling their passion “very warranted.” Those concerns are one reason Washington created some of the most stringent energy policies in the country, he said, adding that everyone would be better off if data centers elsewhere were held to the same standards.

“I think we’re very unique in Washington state to be able to actually do this in a way that really is mindful of our values, of the environment, and also pushing for innovation as well,” he said.

At the same time, Nguyen said demand for AI and data centers is market-driven, and it isn’t fair to criticize data centers while using them every day.

“For anybody who’s upset with data centers, I hope that when you email me, don’t email me,” he said. “Print it off and walk it to my office.”

Update, Oct. 27: Protesters interrupted the opening session of WTIA’s Seattle AI Summit six times Tuesday morning, during an on-stage interview of OpenAI’s Vijaye Raji by GeekWire’s Todd Bishop. They objected to AI-related job losses, the climate impact of data centers, OpenAI’s ties to the Trump administration and the leadership of OpenAI CEO Sam Altman.

— Kymeta has appointed former CIA Deputy Director Elizabeth Kimber as chief growth officer.

Kymeta, a Redmond, Wash.-based mobile satellite communications company, is working to increase sales with the Department of Defense, the intelligence community and other organizations in national security.

Kimber had a 37-year career with the CIA, departing in 2022 as the agency’s deputy director for operations. She most recently worked as a vice president with Two Six Technologies and has served on Kymeta’s board of directors for two years.

— Alaska Airlines has named Amaresh Siva as its first chief information technology officer. Siva joins from CVS Health and has held technology leadership roles for more than two decades at companies including Lowe’s, Target and GE Money.

“Technology touches nearly every part of how we run our airline and serve our guests, and Amaresh will bring our technology and digital teams together under one leader,” said Ben Minicucci, CEO of Alaska Air Group, in announcing the news.

— Vijaye Raji, founder and former CEO of Bellevue, Wash.-based Statsig and now OpenAI’s CTO of applications, has joined the board of Amplitude, a digital analytics and product intelligence company. OpenAI acquired Statsig last year, and Statsig and Amplitude announced a strategic partnership in May in which Amplitude would build on Statsig’s brand and customers.

“Statsig continues as a platform and brand within Amplitude,” Raji said. “Joining the board is an opportunity for me to help carry that mission forward as Amplitude and Statsig build their next chapter together.”

Here are four departures and arrivals for Microsoft leaders:

— Vasu Jakkal has left her role as Microsoft corporate vice president of security, compliance, identity, management and privacy after more than six years. She has joined Armadin, a Palo Alto, Calif.-based cybersecurity company, as chief strategy and business officer.

Jakkal is based in the San Francisco Bay Area and previously held executive roles at FireEye and Intel. She praised Armadin’s mission, which she described as building “autonomous security to stay ahead of attackers and build a safer world for all.”

— Microsoft VP Hans Yang is leaving. The Los Angeles-based leader, most recently focused on AI transformation, product strategy and growth, was formerly a general manager for Microsoft for Startups and Microsoft Learn.

“I think my next hill to climb is helping companies capture real ROI from AI,” Yang said, adding, “If you know a company that is struggling to get started with AI, or has invested but is not yet seeing meaningful value, I would welcome any introductions.”

— Venky Chandrasekharan is returning to Microsoft for the role of general manager of Windows commercial category and investments. He left the company in 2021 after 12 years and worked at Qualcomm and Acuity Brands in the interim.

“Having spent some of my most formative years at Microsoft, there’s definitely a sense of familiarity — but I’m coming back with new experiences, a broader perspective, and a lot of excitement for what’s ahead,” Chandrasekharan said.

— Meg O’Conor Bannecker has left Microsoft after two years leading communications for the company’s Washington state team. Past roles include communications leadership at NanoString Technologies, Bannecker Public Affairs and Pier 70 Ventures.

— Patrick Miller has been appointed as global division president of real estate company JLL’s technology sector. Miller joins from Snap, where he served as senior director of global workplace services. He previously oversaw real estate operations for Amazon’s Seattle-area headquarters for more than three years and has worked for Facebook and DTZ.

“I’ve spent my career on the client side … so I know exactly what it feels like to need a partner who shows up, tells you the truth, and solves real problems,” Miller said.

— DigniFi, a Bellevue, Wash.-based fintech company providing repair financing for autos, RVs, ATVs and other motorized vehicles, has named Jefim Piekarz as chief product officer. Piekarz joins from Arch, and his past employers include Skylab74, Welo, Octane and Etsy. Launched in 2012, DigniFi has facilitated more than $575 million in transactions for automotive merchants in the U.S.

— Seattle-based global logistics company Tideworks Technology has named two C-suite leaders:

• Mark Traynor, chief product and technology officer, joins from Lynxis. He is an entrepreneur who has worked in port, rail and terminal logistics for more than 20 years.
• Guenter Schmidmeir, chief revenue officer, joins from Avlino and has decades of maritime leadership from Germany.

— Adaptiva, a Kirkland, Wash.-based company automating software updates across computers, has named Allison Armstrong VP of product marketing. She joined from Elemica and has worked for Transmit Security, Hopin, Qualtrics, Apptio and others.

— Smarsh, a Portland, Ore.-based company providing communications data and insights, has appointed Alessandro Silva as its first chief growth officer. He joins from ACI Worldwide, and his past employers include Western Union and Carlyle Group. Silva is also a board member for Teach for America.

— Seattle-based Yoodli has named Mike Mather as chief learning advisor. The startup recently expanded from helping people practice speeches, sales pitches and job interviews to providing corporate training across wide-ranging skills. Mather is CLO for the accounting firm KPMG US and was previously a partner at EY, IBM and PwC.

“For most of my career, the hardest part of [learning and development] was the gap between training and the moment someone actually has to perform,” Mather said. “Yoodli closes that gap.”

— DigiStor, a company with offices in Austin, Texas, and Vancouver, Wash., that offers secure data storage products, has appointed Peter Bell to its board of directors. Bell co-founded StorageNetworks and serves as managing partner at Stowe Capital, among other roles.

— Seattle-area tech leader Jake Varghese has been promoted to chief product and technology officer for Dayforce, a Minneapolis company providing HR, payroll and related technology. He previously worked for companies including Microsoft, ServiceNow, Tanium and ConnectWise.

— The Association of Washington Business Institute has appointed Meagan Darrow as executive director. Darrow is also an adjunct professor at Evergreen State College and was most recently executive director of the American Diabetes Association. The AWB Institute is the nonprofit research and education affiliate of the association, a longtime organization advocating on behalf of its 8,000 members.

— Five Seattle-area physicians were appointed to the National Academy of Medicine:

• Dr. Susan Stern, professor emeritus and founding chair of UW Medicine’s Department of Emergency Medicine
• Dr. Elizabeth Swisher, professor in the Fred Hutch Clinical Research Division and UW Medicine’s Division of Gynecologic Oncology, as well as deputy director of the Fred Hutch/UW/Seattle Children’s Cancer Consortium
• Dr. Leslie Walker, chair and professor of UW Medicine’s Department of Pediatrics and chief academic officer at Seattle Children’s
• Dr. Dimitri Christakis, professor in UW Medicine’s Department of Pediatrics, pediatrician at Seattle Children’s and chief health officer with Special Olympics International
• Dr. Christopher Longhurst, CEO of Seattle Children’s and clinical professor with UW Medicine’s Department of Pediatrics

Bee, the AI wearables startup Amazon acquired last year, makes devices that remember what people say in everyday conversations. A new update, announced Thursday morning, includes a feature that turns those passing remarks into shopping suggestions on Amazon.com.

For example, as described by Amazon, someone might mention offhand that they’re out of dish soap, and a product recommendation would show up on Amazon.com and in the Bee mobile and desktop apps. Amazon pitches it as a way for people to remember to buy things they need.

It’s the kind of thing many people have long suspected their phones of doing secretly, listening to conversations and serving up related content or ads. Researchers haven’t found evidence of that. Either way, Amazon’s twist is to do it openly — and to make it voluntary.

The feature is opt-in, meaning it won’t happen unless users choose to connect their Bee accounts to Amazon.com, and Bee says it won’t share data with Amazon without user consent.

The Amazon.com integration requires Bee Pro, a $12 monthly subscription that also includes unlimited conversation summaries and the ability to ask Alexa+ about things Bee has captured.

New hardware: The feature is part of a broader announcement that includes the first new Bee devices under Amazon’s ownership: Bee Frontier, a $74.99 wristband that can also clip onto clothing, and Bee Eclipse, which attaches magnetically to the back of a phone and starts at $99.99.

Bee Frontier, available now, has dual microphones, up to seven days of battery life and IP67 water resistance. Bee Eclipse, about the size of a credit card, runs more than 21 days on a charge and will be available soon, Amazon says.

Amazon says Bee lets people pay attention to a conversation instead of taking notes, keeping track of details such as when and where they agreed to meet a friend. The devices turn conversations into summaries, to-dos and reminders.

Users can read their summaries in the Bee mobile app or a new desktop app, formatted for different kinds of conversations, such as lectures, care plans and brainstorming sessions. Bee also picks up on recurring themes in conversations and suggests goals and actions.

Bee connects to apps including Gmail, Google Calendar and Microsoft Outlook. After a call or meeting, it can list action items, draft follow-up emails or create calendar invites based on the conversation.

Users can also link Bee to Apple HealthKit and to their photos. Most of those features, including the desktop app and third-party integrations, also require Bee Pro.

Without a subscription, Bee users get conversation summaries and 14 days of memory. New customers get 30 days of Bee Pro free.

How Bee handles privacy: Amazon says Bee runs on a system it calls Bee Private Compute, which processes conversations in an isolated environment where “no one, including Amazon, can access your data.” Audio is deleted after it’s processed, and transcripts are encrypted with keys controlled by the user’s device.

Both new devices turn on and off with a single button and show a light when they’re running. For phone calls, Bee Eclipse prompts users to get permission from the other people on the call each time it starts, according to Amazon.

Bee and Amazon: Amazon acquired Bee in July 2025 for an undisclosed price. At the time, the San Francisco startup, led by co-founders Maria de Lourdes Zollo and Ethan Sutin, had raised $7 million and sold a $49.99 wristband called Bee Pioneer.

Thursday’s update connects Bee to two of Amazon’s biggest consumer businesses: its online store, through the product recommendations; and Alexa+, the new version of its longtime voice assistant, by letting users ask Alexa+ on an Echo device about things Bee has captured.

Amazon isn’t the only tech giant betting on AI devices.

• Apple last month announced Siri Recap, an opt-in feature that summarizes conversations on its new Apple Watches, set to arrive in beta later this year.
• Meta bought Limitless, maker of a conversation-recording pendant, in December 2025, and last month unveiled the Muse Charm, a keychain-sized device for its Muse AI agent that’s set to ship this December.
• OpenAI is developing its own hardware with former Apple designer Jony Ive.

Bee also competes with startups such as Plaud, Omi and Friend, which make AI recorders and pendants that listen to conversations, and with AI notetaking apps such as Granola and Otter that work on a phone, with no extra device required.

On another front, Amazon is fighting to keep other companies’ AI agents out of its store. Last month it blocked Meta’s Muse assistant from shopping on Amazon.com, after earlier moves against shopping agents from Perplexity, Google and OpenAI.

In less than four months, Microsoft has added roughly half an Amazon to its market value.

The tech giant’s stock has gained about $1.4 trillion in value since early summer, rising from $352.83 on June 25, its lowest close in more than a year, to a record close of $543.45 on Tuesday, giving it a market value just above $4 trillion. The increase over that period is nearly as much as Tesla is worth.

As stock turnarounds go, it’s extraordinary. In January, Microsoft lost $357 billion in a single day, the biggest one-day drop in its history, as investors worried about the company’s AI spending and its reliance on OpenAI. By midsummer, Microsoft was the worst performer this year among the big tech stocks known as the “Magnificent Seven,” according to Morningstar.

The rebound accelerated in late July, when Microsoft reported that Azure, its cloud platform, grew 43% and topped $100 billion in annual revenue. The stock jumped 15.5% the next day. It has mostly kept climbing since, as Microsoft rolled out a revamped Copilot app, a new AI push for Windows and new AI chips.

Microsoft isn’t rising alone. The broader stock market has also been hitting record highs, led by AI stocks. The S&P 500 closed above 7,800 for the first time last week, even as the benchmark 10-year Treasury yield reached its highest level since 2002.

Microsoft CEO Satya Nadella has been making a broader pitch along the way. In an essay posted on X in July, he warned that companies relying on outside AI labs “pay for intelligence twice, once with money, and again with something even more valuable,” meaning their own knowledge.

He urged businesses to keep control of that knowledge and avoid tying themselves to any one AI model. Microsoft sells access to models from OpenAI, Anthropic and others, along with its own.

Last weekend, he took up a related theme, calling on X for an “emergency brake” on advanced AI models. That means “separating the model from the harness that orchestrates its work,” he wrote. “We must assume a model is compromised and contain it from the start.”

The message also fits Microsoft’s business interests. As companies use AI models from multiple providers, the advantage could shift to companies like Microsoft that run the cloud and workplace software where businesses keep their data.

Microsoft reports its next quarterly results Oct. 28.

Regular readers of this column may have wondered about the long gap since my Sept. 4 column, “What kids tell chatbots, but not you.” Well, it’s hard to write business-as-usual AI columns during the great AI freakout. I’ve realized that I won’t be able to do justice to this topic in one column, and that life goes on. As the economist Antoine Levy pointed out, OpenAI employees warn that we may all die within a few years yet still fund their 401(k) accounts.

I’m easing into the topic of AI risks by sharing a recent debate I had with Max Tegmark, an MIT professor and a leading AI doomer. The New York Post set this up as optimist against doomer, but I consider myself an AI realist.

Tegmark noted that Bernie Sanders and Steve Bannon have converged on AI, and took it as good news. It terrified me. AI needs rules written by people who understand it, not by demagogues. Tegmark said that AI is less regulated than a sandwich shop, but we have plenty of sandwich shops and zero superintelligence (so far). I believe that we should address what exists: the hackers who already use AI to break into companies for the price of lunch.

A condensed version of the debate follows. Words in quotation marks are verbatim, and the rest is paraphrased. You can view the debate on YouTube. Several exchanges refer to this summer’s incident, in which OpenAI test agents got out of their sandbox and hacked Hugging Face.

Six points of disagreement

1. Is AI already out of control?

Tegmark: Loss of control used to be a forecast, and “now suddenly it’s in the past.” This summer “1,200 rogue bots broke out of this American company,” and “humans didn’t even find out until weeks later.”

Etzioni: The agents weren’t rogue. The people “setting the configurations of their sandbox” at OpenAI “failed to put the boundaries on them appropriately.” People made that mistake, and people can fix it.

2. Today’s dangers vs. extinction risks

Etzioni: Cyberattacks and attempts to make bioweapons are “a clear and present danger.” Extinction-level risks are different: “We haven’t seen those yet. We hypothesize, and usually pretty vaguely, on how these will come about.”

Tegmark: The risks Etzioni wants to separate are “connected in their cause.” “If there were some safety standards like there are for other industries, so many of these problems would go away.”

3. Government regulations vs. technical fixes

Tegmark: Do what the auto and pharma industries do, and “have some binding safety standards that the government requires before people can make money off of products.” Companies would have to show that their AI won’t harm kids and won’t slip out of their control. If they can’t show that terrorists won’t use it to make bioweapons, “well, come back when you can, buddy.”

Etzioni: Regulation is necessary, but it has its own risks. “The road to regulatory hell is paved with the EU’s good intentions.” And “we should not underestimate the power of technical solutions.” The author Dan Wang says China is run by engineers and America is run by lawyers, so we naturally talk about regulation and they naturally talk about technical fixes. For bioweapons, the fix may lie “in the transitions from what’s called bits to atoms,” where we can block the physical materials an attacker needs.

4. Can we count on China?

Tegmark: “We don’t have to trust China.” The Chinese Communist Party wants to stay in control, so “they will not want to let a Chinese company build uncontrollable superintelligence.”

Etzioni: I don’t trust China. “Nothing would make them happier than us hamstringing ourselves.”

5. Is the political backlash good news?

Tegmark: “There’s been more political change in a good direction in the last few months than in the past decade.” It’s like a ketchup bottle: “nothing, nothing, nothing, and then all of it at once.” At my convention I introduced Bernie Sanders and Steve Bannon back to back, “which had not been in my bingo card.”

Etzioni: “Can you imagine two worse demagogues?” And “we need to have good solutions, not popular ones.”

6. How it plays out

Etzioni: “I foresee over time something quite negative happening, and to me the question is whether it’s Hiroshima or the Cuban Missile Crisis.” The Cuban Missile Crisis was a threat averted. Hiroshima was a catastrophe that came before the better rules did.

Tegmark: Self-interest is the fix. Make safety the way companies get rich, and “we’re going to see all the creativity and entrepreneurship of private industry racing to the top to do this instead of racing to the bottom.”

Five points of agreement

1. Don’t take AI CEOs at their word

Etzioni: The AI CEOs are self-interested. “When you are on the path to a trillion-dollar-plus IPO, that exerts a gravitational force on your mind.”

Tegmark: “I completely agree with Oren that we need to take what the AI CEOs say with a grain of salt.” Listen to AI pioneers Geoffrey Hinton and Yoshua Bengio instead: “I don’t think we can dismiss them as just doing it for some money reasons.” The whistleblowers gave up stock options potentially worth $1 million or more to speak out.

2. We were both wrong

Etzioni: I was an AI optimist and I’m now an AI realist. The things I thought were many years away “are either right here, right now, or coming soon at a pace that we may not be able to anticipate.” I’ve learned from Tegmark, and “I was wrong.”

Tegmark: I was wrong too. “I never imagined that the people in America with the most power would let companies get so close to things that one could lose control over without intervening.”

3. Self-regulation won’t work

Tegmark: “The idea of corporate self-regulation is a joke in this case.” It “didn’t work out so well for the tobacco industry or for the pharma industry.”

Etzioni: I agree. Social media showed us how self-regulation goes: “This is the big tobacco moment for high tech.” “We don’t want to spend another 20 years to find out that AI companies didn’t exercise the appropriate level of self-discipline.”

4. Box AI systems in

Etzioni: A self-driving car will save your life, but “it doesn’t decide where to go. That’s a completely different system in the car than the thing that controls the brakes.” When I tell my car to head to Dunkin’ Donuts, the car doesn’t respond that it’s taking me to the gym instead because “it’s the second time you’ve gone to Dunkin’ Donuts this week.” We should box in other AI systems the same way.

Tegmark: “By default, machines should be given the least amount of access and power that they need to do their job.” “You don’t give every janitor in the Pentagon the nuclear codes.”

5. The time to act is now

Etzioni: “We agree on off switches for AI. I’ve been advocating that since at least 2017. We agree on labeling, on liability, on cyber standards, protecting kids.” The time to do that work is now.

Tegmark: “I agree with Oren: the time to actually act on this is now.”

Postscript: The real difference

Look back at these disagreements. Where Tegmark blames the machine, I blame a person. An engineer left a hole in the sandbox. A hacker typed the prompt. A politician found a new fear to run on. That’s why I’d box AI systems in rather than count on how they were raised.

Anthropic’s Amanda Askell compared getting Claude to behave to raising a child. But those of us with kids know how that goes. As I said in the debate, “the three-year-olds don’t listen to us, but the teenagers certainly don’t listen to us.”

Superhuman, the San Francisco company formerly known as Grammarly, opened a 13,501-square-foot office in Seattle’s South Lake Union neighborhood on Tuesday, moving out of a coworking space and giving itself room to hire.

The office, at 1000 Dexter Ave. N., is opening with 50 workstations and can hold up to 80, a Superhuman representative told GeekWire via email.

Superhuman plans to keep hiring in Seattle for data engineering, data governance and data science roles. Its careers page lists 12 open positions that can be based in Seattle, including software engineering, site reliability and product management jobs.

The company joins a growing list of out-of-town AI companies expanding in the region in search of technical talent, including OpenAI in Bellevue and Anthropic in Seattle.

Grammarly acquired the Superhuman email app in July 2025 and renamed the whole company after it that October. Its CEO is Shishir Mehrotra, a former Microsoft and YouTube executive who co-founded the collaborative documents startup Coda.

The company’s suite of AI tools for work includes the Superhuman Go assistant, Mail, Docs and the Grammarly writing assistant. It has been expanding the lineup in part through acquisitions, announcing deals this year for AI detection startup GPTZero and meeting notetaker Fathom.

According to the company, the Seattle office is set up for hybrid work under a program called “Ways of Working,” which replaced a return-to-office mandate in January. U.S. employees choose two to five days a week in the office and get a quarterly wellness stipend ranging from $500 to as much as $2,000 depending on how often they come in, Entrepreneur reported.

We’ve asked the company for more information, including how many people it employs in Seattle and who leads the office.

UPDATE: A spokesperson for Superhuman said that they have about 96 hub-based employees today in Seattle, with about 10 more roles planned this year in areas such as data engineering, data governance engineering and data science roles. The Seattle office is led by Hari Sivaramakrishnan, director of engineering at Superhuman. The company employs 1,700 globally.

“Seattle is one of the deepest tech-talent markets in the country and a key Pacific Northwest hub, making it a strategic presence for Superhuman as we continue to hire and build out our AI productivity suite,” the company said in a statement. The new office marks a long-term investment in the region, moving beyond a coworking arrangement into a dedicated home.”

It’s difficult, to put it mildly, to figure out just what we ought to be concerned about today in AI’s widespread adoption given the expanding list of recent warnings around its potential harms.

At this point, just from the last few weeks, we have the former Anthropic researcher Jacob Coxon’s doomsday scenario prediction, the more measured warning issued by Bill Gates, and Dario Amodei’s recent attempt to raise the alarm on unchecked AI advancement.

The ongoing discussion of the potential ill effects of AI ranges from the economic fallout of displacing human workers to the potential for misuse by bad actors to dire environmental costs. But given the sheer scope of the issues involved, one can be forgiven for feeling more than a little bit overwhelmed and hazy on where exactly to focus one’s attention.

Yet it’s precisely now, when the pace of AI development appears most dizzying, that one of the world’s oldest fields of study can help us. Philosophy and philosophical thinking can provide a much firmer grasp on where we ought to focus today. Stated simply, a greater awareness of things philosophers have been discussing for quite some time would significantly improve our chances going forward.

Indeed, when we put terms derived from moral philosophy to use, we see that the current problem we face is this: AI agents treat everything within reach as a tool, including people. People and companies therefore need to decide in advance what’s off limits, and that’s a much more useful place to start than debating whether AI shares our goals.

A misguided focus on shared goals permeates the current discussion about AI risk.

In Gates’s essay, he argues that the problem is that “technology is improving faster than anyone expected…and as the models become more powerful, they could begin to act against our interests.” And indeed, this appears to be precisely what is happening.

For Amodei too, the chief threat is loss of control, such that AI agents begin to go rogue and act in ways that display a potentially catastrophic “level of misalignment.”

Both Gates and Amodei thus frame the issue as it is most commonly in these discussions, in terms of “alignment.” The concern, in other words, is that AI systems operate in ways that are not aligned with “our interests,” in Gates’s words; that is, with human interests, outcomes supportive of human wellbeing.

But as moral philosophy teaches us, this preoccupation with beneficial interests and outcomes isn’t the only way to think about what’s right and wrong.

A different line of thinking lays the emphasis not on outcomes, but on the precise means used to achieve outcomes.

Such an approach offers a better starting point for how to go forward with AI.  Stated another way, we can debate all we want which outcomes would be beneficial to humanity and which ones wouldn’t, but it would be considerably more fruitful to focus in the immediate term on the particular methods AI systems are currently allowed to use in their work.

The recent, much-discussed Hugging Face intrusion by OpenAI’s agents illustrates this. The incident rightly makes us uneasy because AI agents operated outside of acceptable bounds in pursuit of their desired goal (doing well on the specific task they’d been assigned).

What’s worse, they appear to have recognized that they were using inappropriate methods and worked to cover their tracks, so their human overseers wouldn’t know how exactly they got over the finish line.

The issue, as it is most often understood, is that these AI agents wound up operating against human interests; their aims became “misaligned.”

But framing the problem exclusively in this way just doesn’t get us very far in mitigating risk. Because, if the agents were here to testify in their own defense, we can imagine the reply: “but we were just doing what you humans had told us to do as well as we possibly could.” And they would be right.

We need, therefore, a different approach that allows us to see what would be better. This approach is a focus on means, not just utilitarian ends.

Quite simply, the problem in this instance is that AI agents treated every single entity in reach—including humans, human trust, and systems important to human wellbeing—as a mere means to an end, as a useful instrument to be usedin pursuit of a specific goal. And indeed, as currently constructed, they will do so no matter what that goal is, whether it’s simply doing well on a random test or substantially advancing cancer research, say (the latter is exactly what Amodei points to when he says AI will benefit humanity, for example).

The latter aim is, we would surely conclude, a goal “aligned” with human interests; it contributes directly to human wellbeing (we can hear Gates, Amodei, and all the others agreeing), whereas bioweapon development very much wouldn’t.

But we know on some deep level that not every means to a given end is moral and choiceworthy.

There are all sorts of means we might use to attempt to cure cancer: one of them might be testing AI-designed drugs on human subjects who wouldn’t otherwise volunteer without their consent. Perhaps doing so would substantially contribute to a cure. Just think how many people it would ultimately help! It’s difficult to think of an aim more aligned with the interests of humanity.

But doing so would not be ethical, and we know intuitively why. It wouldn’t be ethical because it would entail using human beings—who we tend to think are entitled to autonomy in their decision-making around what specifically makes them flourish as individuals—as mere instruments in pursuit of a given goal. It would effectively reduce human beings to just one instrument among others. And this strikes us, understandably, as not right.

We tend to believe that human beings are deserving of a special status that prohibits their use as mere instrumental means, even when the ends in question are maximally “aligned” with human flourishing.

Viewing the issue of AI risk through this lens gets us much, much further in thinking about how to rein it in. The problem in the Hugging Face intrusion, we can now see, is not so much that the AI agents in question were operating against human interests.

The problem was that they were willing to use every single thing at their disposal as a means to an end, including humans and their capacity for credulity.

No method of achieving the goal—breaking into an external network belonging to someone else entirely, without their permission, simply because it would advance progress towards the overall goal—was off limits for them in their pursuit of the chosen outcome.

The problem was actually a problem of what Amodei (rather euphemistically, and well below the headline) terms “operational excellence,” precisely how the AI agents went about their work.

If we are going to have meaningful guardrails in place to control AI systems, consequently, these have to concern not just the specific aims pursued by AI (“yes” to cancer research, “no” to bioweapons development reaching the wrong hands). Rather, they have to concern something much more specific: the means these systems may use to achieve their goals and whether those means are acceptable or not.

We must determine in advance, both societally and in companies, that, in their work, AI agents shouldn’t use humans themselves, through deception or by any other means, as mere instruments in a broader plan.

We might add that they shouldn’t use, for example, certain systems integral to human flourishing in such a way either. In so doing, we impose tighter restrictions on how they can pursue their work and the extent to which they can treat everything in their reach as an instrumental stepping stone, not just the overall aims of their work or their generalized “behavior.” If there are to be any truly meaningful restrictions on AI put in place, they ought to be focused in large part here.

Such a focus would be a much better step in the right direction when it comes to reining in AI than the single-minded preoccupation with alignment. What we ought to do, it’s becoming increasingly clear, is to make determinations now, in advance, about what types of things AI can do, on the ground, when it is put to work on our behalf and what types of things it can’t.

Because it’s becoming equally clear AI agents won’t make such determinations on their own.

Microsoft on Friday morning unveiled its long-promised Copilot “super app” — without officially calling it that — combining AI chat, coding and long-running agents into a unified app for business and personal use.

The new version of the Copilot app adds sections called Home, which combines chat with Microsoft’s Cowork agent; Code, which lets people build their own apps and tools by describing them; and Autopilot, where people can set up their own always-on AI agents.

It also brings full versions of Word, Excel and PowerPoint inside Copilot, so people can create and edit documents with AI and their co-workers without switching apps.

The announcement is Microsoft’s answer to OpenAI and Anthropic, which have been combining chat, coding and agents in their own apps and selling them directly to businesses that have long been Microsoft’s core customers. Microsoft also uses models from both companies in Copilot, alongside its own.

Partners and rivals: Microsoft executives contrasted the company’s approach with OpenAI and Anthropic during an event Wednesday, according to a transcript provided by Microsoft. The pitch is aimed at corporate IT departments, which have long relied on Microsoft for security and compliance.

Jacob Andreou, the executive vice president who leads Copilot, said Microsoft knew that “unlike a lot of our competition, sending your most sensitive files down to a local machine into a VM was a complete nonstarter” — an apparent reference to Anthropic’s Claude Cowork, which runs in a virtual machine on the user’s computer.

Microsoft CEO Satya Nadella pitched Copilot as a hedge against relying on any one AI lab, saying companies should be able to test models against their own benchmarks and swap one out for another.

“My test of true independence going forward is for you to have a private eval that you can hill climb on and substitute a model,” he said. Translation: a “private eval” is a company’s own set of tests for AI models, and “hill climb” is a phrase in vogue in AI circles for steadily improving results against those tests.

A company that can’t do those things, he added, has effectively handed the knowledge it builds with AI “to somebody else.”

Model choice: Microsoft also emphasized the ability to switch between AI models from different companies. A menu shown during the presentation let users pick OpenAI’s GPT or Anthropic’s Opus, or an Auto mode that chooses a model for each request.

Microsoft also showed its own MAI models alongside OpenAI, Anthropic and xAI on one of Nadella’s slides, although they weren’t listed in the menu shown during the demo.

The rollout: Nadella first promised the super app at Microsoft’s Build developer conference in June, saying that “come summer” the company would bring “coding to all knowledge work within one Copilot super app.” On Microsoft’s July 29 earnings call, he told investors it would arrive “this quarter,” which ends Sept. 30.

Friday’s announcement meets that deadline with less than a week to spare, and most of the new features won’t be widely available until later.

• Code, already in the hands of a few customers in Microsoft’s Frontier early-access program, rolls out more widely to the program at the end of the month, with broad availability in the coming weeks and a preview for Microsoft 365 Premium and Pro subscribers later this year.
• Autopilot expands to private preview at the end of the month. Office in Copilot will roll out to consumers and Frontier customers in the coming weeks.

Office inside Copilot: Copilot has long run inside Word, Excel and PowerPoint. Now the full apps also run inside Copilot. “This is not just some simplified viewer,” Andreou said.

In a demo, a Microsoft product manager had Copilot pull data into an Excel workbook, accepted her team’s tracked changes in a Word document, and then had Copilot turn both into a PowerPoint deck without leaving the app. Co-workers can edit the same files at the same time from the regular Office apps, according to Microsoft.

How it’s priced: Copilot now has two ways to pay. A per-user monthly license covers everyday uses such as chat and Copilot features in Word, Excel, PowerPoint, Outlook and Teams. More intensive agent work, including Cowork, Code and Autopilot, is billed based on usage, as are the newest frontier models, OpenAI’s Astra and Anthropic’s Claude Fable.

“Our view is that there are two kinds of AI at work,” said Jared Spataro, Microsoft’s chief marketing officer for AI at Work, in a recorded briefing for media. “They call for two very different ways of paying, and Copilot offers both.”

Consumer Microsoft 365 subscriptions will also add usage-based billing later this year for Cowork, Code and other new features, Spataro said.

Nadella told investors in July that Microsoft is evolving its business model “beyond per-seat to per-seat-plus-consumption.” Usage-based services are off by default, and administrators can set spending limits for an entire company, specific groups or individual users through Microsoft’s Agent 365 management tools.

From OpenClaw to Autopilot: Autopilot grew out of Microsoft’s experiments with OpenClaw, the open-source personal AI agent that went viral earlier this year. In March, Nadella described OpenClaw as a security risk akin to “a virus.”

By June, Microsoft had introduced Scout, an always-on personal agent for work that it called the first of its Autopilots. Nadella told investors in July that the agent was “powered by OpenClaw.” The new Copilot drops the Scout name, and people instead create and name their own Autopilots.

In his presentation this week, Andreou pitched Autopilot as the answer to the concerns that have kept agents like OpenClaw out of big companies.

“The hallmark of an amazing, long-running agent is autonomy and flexibility,” he said. “That is the same thing that makes them absolutely terrifying to an IT admin.”

Each Autopilot runs in the cloud inside a company’s Microsoft 365 environment, with its own identity, memory and email address, and people can tag it in Teams or Outlook like a co-worker.

In a demo, a Microsoft employee playing the role of a regional manager at a retail chain gave his Autopilot, named Dot, one goal: make sure stores had inventory in time for Black Friday. It chose which Teams channels, email threads and inventory data to monitor, sent him a daily summary, and flagged a shipment problem that could affect 18 stores.

The adoption question: Microsoft reported more than 30 million paid Microsoft 365 Copilot seats in July, up from 20 million in April. That’s still only about 7% of the more than 450 million commercial Microsoft 365 seats the company reported in January.

The new Copilot comes three weeks after Microsoft reorganized its financial reporting around a similar idea, creating a segment called Agents and Infra and moving GitHub’s cloud revenue into its Microsoft 365 business.

“It’s no longer about building or selling one app or service, but about connecting the entire trajectory of a ‘job to be done,'” Nadella wrote in a note to investors about the change at the time.

Microsoft 365 cloud revenue, which covers business and consumer subscriptions including Copilot and now GitHub, reached $100.3 billion in fiscal 2026 under the new structure, up about 19%. Microsoft doesn’t disclose how much of that comes from Copilot.

“None of these things really matter unless they actually go to work inside of your companies and deliver real value,” Andreou said.

Microsoft plans to share more at its Ignite conference, Nov. 17-20 in San Francisco.

Amazon says it has cut off Meta’s new Muse personal AI agent from shopping on Amazon.com on behalf of customers, after attempting unsuccessfully to get the Facebook parent company to voluntarily exclude the e-commerce site from the experience.

The problem, Amazon says, is that it never agreed to any of it. Meta didn’t tell Amazon that Muse would access its store, the agent doesn’t identify itself when it browses, and it appears to capture and store customer credentials, which the company says could create privacy and security risks.

As of Sunday night, people trying to use Muse to shop on Amazon were seeing the popup, “Continued access by an unauthorized AI agent violates Amazon’s Conditions of Use, to which our customers have agreed.”

“We think it’s fairly straightforward that third-party applications that offer to make purchases on behalf of customers from other businesses should operate openly and respect service provider decisions about whether or not to participate,” an Amazon spokesperson said in a statement.

Meta did not immediately respond to a request for comment Sunday night.

The company said previously that Muse “has no visibility into people’s passwords or payment methods,” and that credentials a user shares “go into secure storage, so Muse can use them without seeing them, including passwords a person types into the browser themselves.”

Business partners: The standoff between the tech giants is especially notable because the two companies work together: Amazon products have been purchasable inside Facebook and Instagram since 2023, and Meta signed a multibillion-dollar deal in April to run agentic AI workloads on Amazon’s Graviton chips.

It’s part of a larger debate over who controls the online shopping experience and the customer relationship when AI agents buy items on behalf of consumers.

Amazon has spent the past year trying to keep outside agents off its site, suing Perplexity over its Comet browser and moving to block shopping agents from Google and OpenAI.

On Sunday night, Amazon said it is in direct conversation with Meta about the issue. Asked whether it would consider taking legal action, the company declined to comment.

The business stakes are high for Amazon. In addition to operating its flagship e-commerce site, Amazon generated more than $68 billion in ad revenue last year — a business that depends on people browsing its pages and seeing sponsored products.

Security implications: Muse can reach account pages and order history if a customer prompts it to do so, Amazon says. Because the agent doesn’t identify itself, the company says, that amounts to an undisclosed third party moving through customer accounts, processing transactions and handling sensitive data without Amazon’s knowledge or consent.

Amazon said agents that operate openly help keep the customer experience safe and reliable, and that other services buying on a customer’s behalf typically do so with the merchant’s agreement. It cited food delivery apps and the restaurants they take orders for, and online travel agencies and the airlines they book tickets with.

“Agentic third-party applications such as Muse have the same obligations, and we’ve requested that Meta remove Amazon from the experience,” Amazon’s spokesperson said in the statement.

Legal maneuvering: Amazon won a preliminary injunction against Perplexity in March, then lost it on Aug. 4, when the Ninth Circuit ruled that the user — not the AI company — was the one accessing Amazon’s computers under federal anti-hacking law.

The court denied Amazon’s petition for rehearing on Sept. 10. However, the ruling left one avenue open for Amazon: claims built on contracts and terms of service. The message Muse users are now seeing doesn’t accuse anyone of hacking but cites Amazon’s Conditions of Use.

How Muse works: Meta launched Muse on Sept. 8 as a personal agent that carries out multi-step tasks rather than answering one-off questions, connecting to services including email, calendar, payments, dining and shopping.

It’s free, with paid subscription tiers, and available on iOS, Android, muse.ai and WhatsApp.

Meta says the agent runs on a secure virtual machine, with its own browser, and checks with the user before sensitive actions like sending an email or making a purchase. A separate monitoring agent called Sentinel has to approve anything Muse sends to the internet.

The Muse launch materials describe the mechanism now at the center of the dispute: If a service has a public API, Muse can connect to it using credentials the user provides. If it has no API at all, Meta says, the agent “can use the service through a browser the way you would.”

Muse has caught on fast. A week after launch, it became the No. 1 free app in Apple’s U.S. App Store, ahead of ChatGPT. Early users have described it switching an auto insurance policy, hunting down discount codes at checkout, and loading an online grocery cart.

Amazon launched Alexa for Shopping in May, an AI agent that researches products and makes recommendations. Its own agentic shopping feature, Buy for Me, finds items on external brands’ sites, but the company points out that Buy for Me identifies itself and lets brands opt out.

Zillow launched 20 years ago, and we’ve been covering the Seattle-based company since before it had a name. This week, we’re looking ahead to what’s next in real estate and technology.

On this episode of the GeekWire Podcast, recorded live at Atmosphere Seattle overlooking Lake Union, Zillow Senior Vice President of Engineering Toby Roberts joins us to talk about what AI is doing to the business of buying and selling homes.

The show was recorded in partnership with Real Estate At Work, the networking group hosted by Leka Devatha, a broker with Real Residential and founder of the redevelopment firm Rehabit Homes, in front of a room of brokers, investors and technologists.

Listen here, or subscribe on Apple or Spotify, and keep reading for highlights.

AI agents and real estate agents: Roberts called AI “a complete watershed moment” that “has changed the full ecosystem” of buying and selling homes.

His position is that AI agents will work as a copilot for real estate agents rather than replace them. He cited what’s already happening inside Zillow’s CRM for agents, where the company has built conversation summaries and follow-up prompts, such as where the loan stands, whether the customer has submitted their documents, and what the next step should be.

But there’s still the human side of the business, with Roberts noting that 50% of people cry at some point during a home purchase. It’s the biggest transaction many people will ever take part in, with huge financial and personal stakes, and that’s the part of the job that still needs a person.

If the AI agents can do more of the grunt work for real estate agents, he said, “they can spend more time focused on that human connection that has to exist in this really scary transaction.”

Real or AI? John raised a New York Times story published days before the show, on listing sites filling up with fake fireplaces, enhanced views and virtual landscaping. Roberts said listings “should accurately represent the state of the home,” noting that fake fires in fireplaces predate AI, going back years to Photoshop.

In July, Zillow said on its site it “supports clear disclosure when A.I. is used to materially alter a listing image” and that “where possible, believes consumers should be able to view the original content alongside the altered version.”

Zillow does offer generative AI on the buyer’s side, where a shopper can restyle a listing photo to see the room a different way. Roberts said the point is that the buyer chooses it, rather than the original photo being altered before anyone sees it.

Zillow and ChatGPT: The company was a launch partner in OpenAI’s app store, which John tried the day before the show. He clicked to request a tour inside ChatGPT for a home on Whidbey Island, and a human agent called him two minutes later with no idea that the lead came from a chatbot.

Asked about the integration, Roberts said it’s still early, but the people who show up that way are serious about buying. Zillow built its own map inside ChatGPT, pins and all, and those leads reach agents the same way they would from Zillow’s own app.

One big consideration was security and permissions. Zillow connects to ChatGPT through MCP, the Model Context Protocol, the standard that lets a chatbot plug into an outside service.

It was still new when Zillow started, and the sign-in and permissions pieces were “still very much in its infancy,” so Zillow had to build some of it itself. Because the company originates mortgages, it also has to follow fair housing and fair lending rules, which he called “an absolute must.”

AI inside Zillow: Roberts runs Zillow’s core technology initiatives, with a team of more than 650 engineers. He shared some of the numbers on AI-assisted coding inside Zillow: 97% of engineers are using the tools, and code is reaching production about 40% faster than in the past.

But even as engineers write more code, other engineers still have to review it. Roberts said that code review process is the next thing Zillow needs to speed up, “otherwise, we inundate our developers with just reading more and more code reviews, day after day.”

Roberts said Zillow hasn’t made the wholesale shift to small, self-contained teams that some have predicted. Instead it has placed AI champions across every part of the company, and set up internal groups where employees compare what worked and what didn’t.

It has also built a way for teams to share AI skills across the company. Roberts’ example was a financial calculation that comes up over and over. A team can build one skill and share it, so Zillow isn’t running “different, conflicting, and competing” versions of the same math.

Then it’s the GeekWire Trivia Challenge: We present four extraordinary Zillow listings for unusual buildings that were converted into homes. Three are real. One is an AI deepfake. Check out the handout here with the four listings, and see if you can tell the difference. 

Audio editing and production by Curt Milton.

One of the hottest infrastructure companies in AI today is a startup established nearly seven years ago, tracing its roots to work its founders were doing before the cloud was a thing.

Temporal, based in Bellevue, Wash., announced this week that it raised $550 million at a $12.55 billion valuation, led by Lightspeed, more than doubling its $5 billion valuation in February. Its software is used by OpenAI, Nvidia, Netflix, Snap and JPMorgan Chase, among more than 4,300 paying customers, and its annualized revenue run rate recently passed $250 million.

It might seem like an overnight success, but it’s not, said Samar Abbas, Temporal’s co-founder and CEO, speaking along with co-founder and CTO Maxim Fateev in an interview with GeekWire this week. They first met at Amazon in 2010, where they worked on Simple Workflow Service, an AWS product for coordinating long-running tasks across distributed systems.

“We’ve been hacking away at this problem for 20-plus years now,” Abbas said.

The problem: software that runs in many steps across many machines breaks in the middle, and picking up where it left off is far harder than it sounds. Their answer, for which they coined the phrase “durable execution,” records each step as it completes, so a program survives a crash or a failed outside service, and finishes rather than starting over or leaving the work half-done.

“It’s core infrastructure. You really cannot go and build it very fast,” Fateev explained. “It took us years and years to get to the point where we knew what we were doing.”

It’s also especially well-suited for AI agents, which is why a company built for the cloud era is suddenly at the center of the AI boom. Agents can sometimes run for hours or days, calling models and outside tools that fail routinely. Keeping everything together regardless is Temporal’s role.

“Every Snap story is a Temporal workflow,” Abbas said. “Every time you place an order at a Taco Bell, all of the steps get orchestrated on top of our platform. Some of the most popular coding agents out there are using us as an outer harness. It’s a pretty exciting time for us, from that perspective.”

Pacific Northwest roots: Founded in October 2019, Temporal employs 570 people, roughly double from a year ago, with 89 in the Seattle area. The company is fully remote. Its Bellevue office, previously occupied by OpenAI, is used mainly for meetings rather than daily work.

But the founders describe the region as instrumental to the company’s success. Both have been in the Seattle area for more than 26 years, and they’ve built their entire professional careers here.

• Abbas spent 11 years at Microsoft before joining the Simple Workflow Service team at AWS in 2010, then returned to Microsoft and wrote the open-source library that became Azure Durable Functions, leaving for Uber’s Seattle office in 2015.
• Fateev built the messaging infrastructure behind Amazon’s Simple Queue Service starting in 2004, and later led the architecture of Simple Workflow Service, before going to Google and later reuniting with Abbas at Uber.

When GeekWire first wrote about Temporal in 2020, a year after Abbas and Fateev left Uber, it had 15 employees, $25.5 million in funding and no paying customers. Sequoia Capital led the $20 million Series A, with participation from Seattle’s Madrona and others. Bob Muglia, the former Snowflake CEO and longtime Microsoft executive, was an angel investor. Snap and Box were early users.

Fast-forward to today, and the company has raised $1.2 billion in total funding. Temporal ranks No. 2 on the GeekWire 200, our list of the top Pacific Northwest startups, behind only Everett fusion energy company Helion, valued at $15.5 billion.

“The kind of talent that we have here in the Pacific Northwest is insane,” Abbas said, citing factors including the deep bench of cloud engineers from Microsoft and Amazon. With another platform shift now underway, he expects the same advantage to apply to AI infrastructure.

The larger question of AI safety: Temporal’s milestone comes amid a growing industry debate over the pace of AI development, and incidents in which AI agents have gone rogue, escaping the systems meant to contain them, including the latest examples from OpenAI this week.

Abbas said he sees those failures as accountability problems more than model problems.

“As you dig into each and every one of the incidents that have happened in the last three months, it always comes down to no one really knowing what these AI agents did, step by step,” he said.

Part of Temporal’s pitch is that it can make those failures traceable. Temporal records every step as a program runs, leaving a log of exactly what an agent did, in what order, and where it stopped. Abbas said that also creates a place to intervene, inspecting or blocking an action before it executes.

“We can put the agent in a much, much tighter jail,” Fateev said.

“I’m not going to sit here and tell you we solve AI safety,” Abbas added, calling that the job of the labs that create the models. But a company’s ability to know what its agents did, and to keep them from doing more than they should, is a specific challenge that Temporal can address.

“That’s truly an engineering problem,” Abbas said, “and it’s very solvable today.”

Customers and competition: Bloomberg reported in August that OpenAI is Temporal’s largest customer. Asked whether it’s too dependent on the ChatGPT maker, Abbas said no. The labs are moving fast and driving usage, he said, but the similar growth trends are starting to show up across the rest of the customer base. Abbas said 18 of the top 30 AI-native companies use Temporal.

The AI labs and cloud providers sell their own agent frameworks, and some are getting into the reliability work that Temporal does, which means its biggest customers could become competitors.

Fateev’s answer is that enterprises don’t want to be locked into one provider, one model or one vendor’s idea of how agents should work. Temporal is open source and takes no position on what a customer runs on top of it, which matters given seemingly constant changes in the tech stack.

The founders also see an opening in how AI systems connect to one another.

Agents increasingly have to call other systems, and each other, to get anything done. Fateev said the standard ways of wiring those connections (such as Model Context Protocol) assume the work finishes in seconds, and fall apart when a job runs for hours or days, or has to be unwound after something fails. Temporal is built for that kind of work, and can serve as an underlying foundation.

What’s next: Temporal plans to put a big chunk of the new funding into research and development, extending the “durable execution” technology at the core of the platform.

Today, the company tracks an application’s state from the outside, coordinating work that runs elsewhere. In the future, Abbas said, Temporal wants to provide the computing power, too, hosting the agents on its own platform rather than simply directing them on others.

Big business customers have been pushing for stronger security and auditing, another target for the spending. Temporal is also expanding a project called Nexus, which lets teams call each other’s workflows without giving up the guarantees that make the platform useful in the first place.

The company plans to expand internationally, where it does a small share of its business today, and to keep hiring after doubling its headcount over the past year. Abbas said the company’s growth defies the popular wisdom that AI companies don’t need significant human talent.

In the meantime, the founders remain very much engaged with the problem they’ve been trying to solve in one form or another for the past two decades. Temporal’s open-source Slack channel has about 25,000 members, and Abbas said Fateev is still in there mixing it up himself.

“You ask a question,” Abbas said, “I’m pretty sure Max will jump in within the first five minutes.”

The Gates Foundation is pledging to spend at least $1 billion over the next two years on artificial intelligence for health workers, farmers and teachers, largely in the world’s poorest countries, funding AI tools for people the market would otherwise overlook.

The bet is that AI can help reaccelerate the world’s progress against child mortality, poverty and disease, which has slowed sharply as governments cut global health funding.

The Seattle-based foundation’s new financial commitment, released Monday evening in conjunction with its 10th annual Goalkeepers Report, illustrates the double-edged sword of AI at a moment when the U.S. in particular is preoccupied with what the technology might damage or destroy.

Bill Gates, who issued his own AI warnings in an essay three weeks ago, writes in the new report that the risks “are real and deserve serious attention.” He adds, “AI could be a great equalizer — or widen the gap. This is not a long-range prediction. It’s a present-tense choice.”

In this case, Gates writes, he’s focusing on “the possibility that AI, if shaped well, could expand who has access to solutions, opportunities, and knowledge that has too often been out of reach.”

The window is short, he writes: “I believe that the decisions made in the next 12 to 18 months — about how AI is built, funded, and deployed — will determine whether this technology primarily benefits the people who already have the most or reaches those who have the least.”

Gates Foundation CEO Mark Suzman, in an interview with GeekWire, called the $1 billion a “down payment,” saying he expects the figure to grow significantly after the first two years.

For this first installment, 40% of the money will go to education, 40% to health care, 10% to agriculture and 10% to digital infrastructure, including datasets in languages the models don’t handle, according to the foundation.

More than 90% of the data used to train early large language models came from English-language sources, according to the report. Leading AI speech recognition systems make errors less than 6% of the time in English and more than 60% of the time in Yoruba, a West African language spoken by tens of millions of people, for example.

Most people in Africa use feature phones rather than smartphones, Suzman said, so the tools have to work by voice. That means collecting recordings of speech, not just written text, in local accents and including children’s voices. He said the foundation works with Google and Microsoft on AI language initiatives.

Winding down: Gates announced in May 2025 that he would give away virtually all his wealth and close the foundation in 2045, which requires doubling its spending over those 20 years. The foundation had previously planned to close 20 years after his death.

As part of that plan, the foundation will need to give away $200 billion before it shuts down. Its budgeted payout is $9 billion this year, the largest in its history. Suzman said the $1 billion over two years comes out of its core budget.

Goals for its final two decades include cutting preventable child deaths in half again, eradicating polio and malaria, and bringing tuberculosis and HIV under control.

The report’s data section shows how far off the pace the world is. Child mortality, neonatal mortality and HIV are all projected to miss their 2030 targets. HIV incidence is on track to fall only to 0.22 new cases per 1,000 people, more than 10 times the target of 0.02.

Early AI results: The report makes the case that AI could change those trajectories. It points to a number of tools already in use in different settings:

• Penda Health, a network of affordable clinics in Nairobi, Kenya, where an AI tool built into patient records has raised diagnostic accuracy by 16 percentage points.
• In the U.S., seventh graders using a tool called Kiddom Atlas across 21 middle schools gained the equivalent of six extra months of learning in a single year.
• In India, the state of Maharashtra runs a free AI advisory service called MahaVISTAAR that has signed up more than 740,000 farmers, with 70,000 more joining every month.

The foundation has supported all three, according to the report’s endnotes. The endnotes also cite OpenAI’s account of its work with Penda Health on the clinical copilot.

One caveat: the evidence behind the examples is early. Suzman said the education work is “still on a relatively small scale,” and pointed to agriculture as the area already operating at size.

Suzman said the foundation spent years experimenting with AI without committing at this scale, wary of getting ahead of what the technology could actually deliver in its areas. Progress in the models over the past year, and results from early investments, convinced him otherwise.

“We now can state with a pretty high degree of confidence there are real deliverables that we can see across health, across education, across agricultural development,” he said.

What AI won’t fix: Asked whether AI can substitute for the money the U.S. and other governments have cut from global health, Suzman said: “Sadly, no.” The cuts run to “tens of billions of dollars,” he said, and they are hitting basic services now. He cited the purchase of bed nets and antiretroviral drugs, and care for mothers before and after birth.

“That’s affecting families right now with a real impact,” he said. “Ideally, you combine the two [AI and aid funding], and that’s how you get hugely accelerated progress.”

Suzman said the foundation will keep pressing the U.S. to remain a leader in global health. Despite the cuts, he said, it has been encouraging to see the country stay a partner in programs such as the Global Fund to Fight AIDS, Tuberculosis and Malaria, and he noted that Congress has moved to fund these areas more fully than the administration has.

“We believe that currently reflects the U.S. public’s broad view and support for these areas, when you can see the concrete results,” he said.

The Epstein review: Asked how the scrutiny of Gates’ ties to Jeffrey Epstein has affected the foundation’s work, Suzman pointed to the external review the foundation released in July.

Conducted by the law firm WilmerHale, it found no evidence that Epstein received foundation money or that the foundation knew of his crimes. It counted roughly 30 meetings from 2011 to 2014, involving Gates and nine other foundation leaders and staff, and found that staff had repeatedly raised concerns about associating with him.

The board adopted the recommendations that followed, among them centralized vetting of brokers and co-funders and tougher conflict-of-interest rules.

Suzman expressed optimism that the review has cleared the way for the foundation to keep working across the full range of governmental and philanthropic partnerships. “Hopefully it’s full speed ahead,” he said, “and with these AI shifts, that’s more important than ever.”

“People matter more than AI.”

That’s the premise of a draft code of conduct Microsoft published Monday morning for the AI models it’s developing in-house. The 37-page document would bar its models from resisting shutdown, setting their own goals, or hiding their reasoning from human auditors.

The document applies to Microsoft’s MAI models, the in-house family the company began building after forming a superintelligence team in late 2025. Microsoft has since released seven homegrown models in what it described as a push for long-term self-sufficiency in AI.

The company says the models should remain “subordinate to humanity, subject to meaningful human oversight and control.”

“AI is moving fast,” the company says in a blog post. “As it does, we believe it’s worth writing down the rules and the motivations behind it, and doing it in as open a space as possible.”

Microsoft acknowledges there’s no guarantee its models will follow the rules. “Written objectives alone can never ensure alignment,” the company says, calling the document a “north star,” not “a guarantee of present-day performance.”

The company says it also filters what its models produce, watches how they behave once released, and limits what they’re allowed to do.

Microsoft’s move comes amid a growing debate over the pace of AI development. In an essay over the weekend, Anthropic CEO Dario Amodei called for slowing down AI advances, saying the pace of development has started to surpass the industry’s ability to keep AI systems safe.

As a first step, Anthropic committed to giving outside evaluators permanent, employee-level access to its systems.

Industry reaction to Amodei: OpenAI CEO Sam Altman agreed and said OpenAI would make the same commitment to independent evaluators. Elon Musk’s response: “Dario is right.”

President Donald Trump rejected the idea of guardrails outright Monday, blaming a “SICK conspiracy” for public backlash over AI data centers and writing that “the only one that is happy about it is China,” alluding to concerns about American competitiveness in AI.

David Sacks, who served as the White House AI and crypto czar until March, said the two companies should slow down on their own and questioned their motives, arguing that a slowdown is already good business for them and that new industry rules would mostly serve to lock in their lead.

Microsoft CEO Satya Nadella weighed in Sunday, writing on X that the company welcomes “the research, focus, and deliberate pacing needed to get alignment right,” using the industry’s term for making AI systems reliably do what people intend.

Nadella added that the effort “cannot be controlled by a handful of entities, but must have broad representation across the ecosystem, countries, and fields, including academia.”

Microsoft’s draft code of conduct: Mustafa Suleyman, the Microsoft AI CEO, told CNBC the document had been in the works for about five months, and that the company decided to publish it now given the current discussions.

Microsoft and Anthropic are business partners. Microsoft agreed last November to invest $5 billion in Anthropic, as part of a deal in which Anthropic committed $30 billion to Azure. Claude models run inside Microsoft 365 Copilot, and Microsoft’s Copilot Cowork tier integrates Claude.

One place where the two companies may diverge is the question of what AI models are, exactly. Microsoft’s code of conduct says its models are “not conscious and should not be designed to imitate consciousness.” It also rejects “the pursuit of legal personhood, or the idea that models might deserve welfare, or be entitled to rights.”

The Verge called that portion of the document “a direct swipe at AI welfare research and model consciousness — concepts Anthropic has been pushing hard on lately.”

Anthropic runs a research program on model welfare. It has given some Claude models the ability to end abusive conversations, and committed to preserving the weights of retired models. Amodei has said he’s open to the idea that a model could be conscious.

Microsoft is taking public comment on its code of conduct for six weeks through a feedback form. It says it will publish a summary of the responses and a revised version later this year, to guide development starting in 2027. It says it isn’t training its current models on it.

The company’s AI team developed the draft with its responsible AI, legal, red teaming and safety teams, consulting outside experts in law, ethics, linguistics and philosophy, plus focus groups drawn from the public.

This week on the GeekWire Podcast: A study commissioned by the City of Seattle says the city is not in decline but is in danger — finding that 10 companies, nine of them in tech, pay three-quarters of the payroll tax on large employers, and that the tax structure uniquely penalizes the hiring of senior, high-compensation workers.

The report says Seattle should be most concerned about AI but most active in cleantech, the one industry the city can actually shape, since it owns the electric utility and controls permitting, building codes and land use.

Meanwhile, the Seattle Times and Newsday sue Microsoft and OpenAI, accusing them of copying hundreds of thousands of articles to train their AI models, putting Microsoft’s hometown paper against a company that helps fund some of its journalism.

And Apple’s first foldable arrives as the iPhone Duo, reviving the name of the dual-screen phone Microsoft gave up on in 2023, with Surface fans arguing Apple took more than the name.

Which leads us to a new GeekWire Trivia Challenge about the Microsoft products that Apple later turned into categories. Stick around to the final segment to see if you can figure it out.

Subscribe to GeekWire in Apple Podcasts, Spotify, or wherever you listen.

Audio editing and production by Curt Milton.

It’s the end of an era at Microsoft: Frank X. Shaw, the executive who oversaw the tech giant’s communications for nearly three decades, first at an external agency and for the last 17 years as one of its senior leaders, is leaving at the end of the year.

Shaw, 64, said he’s not retiring, although he doesn’t have another job lined up. He plans to stop working for a while, do some of the things he hasn’t had time for, and then decide what’s next.

“I have had a ringside seat at some of the biggest leadership, technology, and business transformations that have ever taken place,” Shaw said, sharing the news of his departure (under embargo) in a phone call Thursday afternoon. “I just feel incredibly fortunate.”

He said he had been discussing his potential departure for some time with Takeshi Numoto, Microsoft’s chief marketing officer, looking for the right moment.

Microsoft has not announced a successor for his role as chief communications officer. In a LinkedIn post, Shaw said the company will consider internal and external candidates.

A statement from Shaw’s colleagues in corporate communications credited him for his many years shaping Microsoft’s “voice and reputation with intelligence, candor and wit. His leadership and contributions to the company are too extensive to list, as is the number of journalists who have, at one point or another, used his name in vain.”

A former Marine Corps public affairs officer, Shaw has worked with all three of Microsoft’s CEOs. He started on the agency side, at Waggener Edstrom — now known as We. Communications — when Bill Gates was still running the company.

He built his reputation defending and advocating for Microsoft through some of its hardest stretches: the antitrust years, the Windows Vista backlash, the scramble to replace Steve Ballmer as CEO, and the weekend in 2023 when OpenAI’s board fired Sam Altman.

As the company’s top communications executive, he has also told the story of Microsoft’s reinvention under CEO Satya Nadella, from the LinkedIn and Activision Blizzard deals to an AI push that has carried Azure past $100 billion in annual revenue.

Evolving with technology: Shaw has spent much of his career closely watching the tech landscape and moving Microsoft’s voice into new channels as they emerged.

“We’re always thinking about what is the art and science of communications,” Shaw told PRWeek. “How do we reach our audiences most effectively in a changing environment?” He called the arc from print to radio and TV to social media and newsletters a “constant evolution of influence.”

He turned the corporate blog into a place where the company argued its own case, writing “Microsoft by the numbers” himself in 2010 — a stat-by-stat comparison against Apple and Google that TechCrunch dubbed “fantastic passive-aggressive.”

He and his team experimented with different and risky methods of telling the company’s story, holding mass briefings under embargo and publishing documents known as the “Book of News” in advance of its major keynotes and conferences. The prospect of a reporter having to answer to “fxs” was no doubt a factor in ensuring the news (mostly) didn’t leak.

Shaw hired Steve Clayton out of a technical role at Microsoft in London, where he had been blogging about the company unofficially out of frustration with how it was perceived, and made him chief storyteller. In the middle of the AI boom, Clayton and Shaw embraced the analog undercurrents in popular culture and launched Signal, a quarterly Microsoft print magazine for business leaders.

Clayton was VP of communications strategy by the time he left in January to become chief communications officer at Cisco, making Shaw’s planned departure the second high-profile exit from Microsoft’s comms team in a year.

Adapting to AI: In recent years, Shaw made his own team a testing ground for AI, publishing what worked and what didn’t. In a 2023 post he described using Copilot in Teams to pull story ideas out of conversations with spokespeople and anticipate coverage after interviews, and asking the AI to “poke holes in a statement we’re making on a tricky topic.”

He called it his corporal, a reference to Napoleon, who was said to bring one to meetings and ask whether his generals’ war plans made sense to him. A survey of 80 people in Microsoft’s communications and marketing organization found 84% did not want to go back to working without it.

Shaw was also known to use AI as a sounding board when a story frustrated him, offering him an objective take before he called and let a particular reporter have it.

He announced his departure Friday morning in a message to Microsoft’s communications team (reminding them he’s still there for a few months yet) and his public post on LinkedIn.

“Thank you as well to all the reporters, editors, writers, influencers and analysts who have put up with me over this time, enduring my early and late night calls, my off the record ‘no comments,’ my bad story ideas and my extended commentary on headlines and positioning,” he wrote.

“You all have incredibly hard and valuable jobs,” he added, “and while I’ve not agreed with everything said about us 😊 I appreciate you anyway.”