• Mark Cuban says data center developers are using healthcare-style contracts to confuse communities.
• He said healthcare contracts are "so big and so long" that insurers "get to be contractual terrorists."
• Data centers, which are crucial for the AI boom, have become a hugely contentious issue across America.

Healthcare has long used complex contracts that can obscure who pays for what. Mark Cuban says data center developers are using a similar playbook.

"A lot of communities have been really dumb with the deals they've negotiated for data centers," Cuban said on an episode of "The A16z Show" released on Wednesday. "They've fallen for the contractual terrorism in a lot of cases."

Cuban, who co-founded Cost Plus Drugs, a company that sells generic medicines at a transparent markup, previously coined the phrase "contractual terrorism" while discussing healthcare.

On the A16z podcast, he said massive healthcare insurers and their subsidiaries use sprawling, opaque agreements to obscure costs and make it difficult for employers to know where their money is going.

Cuban said those contracts are "so big and so long" that insurers "get to be contractual terrorists" — a formulation he also applied to communities' data center deals.

His comments come as the data center boom has become a hugely contentious issue in communities across the US.

Residents living near facilities have reported incessant low-frequency hums, diesel smells, intrusive lights, construction disruption, pressure on water and electricity supplies, and the loss of trees and open land.

In Northern Virginia, the country's data center capital, local opposition has helped reshape elections, prompted lawsuits, and pushed residents to demand tighter noise and zoning rules.

A University of Massachusetts Amherst poll of 1,000 Americans from September found that 65% opposed building an AI data center in their local community, including 49% who said they strongly opposed it.

The ghost data center backlash

The race to build AI infrastructure has also produced far more announcements than operating facilities.

Aterio, which tracks projects through satellite imagery, permits, and other records, counted 8,268 announced US facilities as of October 20. But only 2,126 were active, and 861 were under construction.

Thousands more had just been announced, leaving local governments to negotiate consequential long-term agreements for projects that may never progress beyond a press release.

That is a risk Cuban has warned about.

In a series of X posts in September, he laid out a 10-point playbook for communities considering their first data center, arguing that every promise from a developer should come with a measurable limit, independent oversight, financial backing, and a remedy if it is broken.

Cuban's advice is to settle those questions before construction begins.

"Resolve it before granting the approval that spends the last of the town's leverage," Cuban wrote at the time.

• Charles Lamanna, head of Copilot, Agents, and Platform, just took over Office and Teams.
• Teams within his new combined unit have adopted flatter structures amid the rise of AI agents.
• He's also enforced a stricter RTO for senior staff, and offered employees relocation to major hubs.

Microsoft's head of the Copilot platform, AI agents, and now Office is making good on his promise to reshape how his teams work.

"The most important thing my team will do this year won't be any single product or feature we ship: instead, it will be changing how we work," Charles Lamanna, the new boss of this key Microsoft unit, wrote in August.

Lamanna has started requiring senior staff to work in offices four days a week, and he's offering relocation to major hubs, according to people familiar with the changes. Teams within his organization have also reorganized to a flatter structure removing middle managers in line with Lamanna's views about how agentic AI is changing engineering.

"We're rethinking where managerial layers are critical and where flatter structures can enable faster decision-making, clearer ownership and greater autonomy," a Microsoft spokesperson said.

The moves offer an early look at how Microsoft is reworking its software empire for the AI age. Lamanna, who already oversees the company's Copilot and AI agent platforms, now controls the Office productivity software and Teams workplace chat service that are used by millions of workers.

Lamanna took over those groups earlier in October when the previous boss, Ryan Roslansky, announced plans to leave at the end of the year. Roslansky was seen as a successor to Microsoft veteran Rajesh Jha, who retired in July. Roslansky and Lamanna were two of the three executives who made up the new Copilot leadership team.

The changes come as CEO Satya Nadella reshapes the company around AI. Microsoft has been consolidating responsibility for its AI products, and a wave of senior executive departures has made way for a new generation of leaders.

"Span of control"

Lamanna's expanding responsibilities make him a key player in this leadership transition, giving him oversight of Microsoft's flagship workplace software and the AI tools that could change how customers use its products.

Recently, the Teams unit now under Lamanna reorganized by increasing the "span of control" for managers in engineering, according to two of the people familiar with the changes.

This refers to the number of employees who report to each manager, and it's become common for tech companies to flatten engineering organizations like this in the AI era, as coding tools and agents write more software.

The reorganization included converting managers to individual contributors, the people said. The idea is that agentic coding capabilities mean Microsoft needs fewer managers, according to one of the people. The team will now have a ratio of one manager per 12 individual contributors.

"Agentic coding capabilities have changed every level of the engineering function," Microsoft's spokesperson said. "Software engineers are no longer hand-assembling every line of code. Instead, they are now designing and directing the 'agentic coding machine' that helps build the product."

As AI makes it easier to build software, skills like making sound technical decisions, designing systems, ensuring quality and security, and understanding customer needs become more valuable.

"At the same time, some tasks that traditionally required managerial coordination, such as managing handoffs and orchestrating work across teams, become less central," the spokesperson said.

Other teams have implemented similar restructuring, the spokesperson confirmed.

RTO and relocation

Lamanna has also started reshaping where employees work. Last week, Lamanna's Copilot, Agents, and Platform (CAP) organization offered select US employees the option to relocate to offices at Microsoft's Redmond, Washington headquarters, San Francisco, or New York, according to three people familiar with the plans.

The people who spoke with Business Insider for this story asked not to be identified discussing internal matters.

Lamanna has also asked employees at the "partner" level and above — some of Microsoft's most senior staff — to start working in physical offices at least four days a week. The companywide standard is three days a week.

Have a tip? Contact this reporter via email at astewart@businessinsider.com or Signal at +1-425-344-8242. Use a personal email address and a nonwork device; here's our guide to sharing information securely.

• NATO has provided new details about its AI model for predicting Russian drone strikes on the front lines.
• Called Initiative 13 or Model 13, it works for about 31 miles behind the line of contact.
• The AI gives commanders a map showing the next day's expected Russian drone activity.

A new NATO artificial intelligence model gives Ukraine's forces a frontline map of where Russian drone activity is expected the next day, per the alliance's innovation center.

The system, called Initiative 13 or Model 13, "turns an array of disparate data into a clear map of expected activity" for commanders to plan their next day of operations, a spokesperson for NATO's Joint Analysis, Training, and Education Center, or JATEC, told Business Insider.

Business Insider reported on Ukraine's announcement of the program earlier this month, but JATEC is now providing new details on how it works.

Initiative 13 covers areas 5 to 50 km, or roughly 3 to 31 miles, of Ukrainian territory behind the line of contact, and divides that space into geographical cells, per JATEC.

"For each cell, the system determines the expected intensity of strike and reconnaissance activity of UAVs, the level of threat, the trend of its change, and the range of possible values," its spokesperson said.

The map shows a forecast for the next day, helping commanders to highlight risk areas so they can better deploy their forces and equipment, they said.

They added that the AI model is trained on battlefield data and considers previous Russian drone activity, what types of drones were used, when and how they were deployed, weather conditions, terrain, and distance to the line of contact.

"Initiative 13 does not replace data from the current situation and commander's decisions," the JATEC spokesperson added. JATEC is NATO's joint organization with Ukraine, based in Poland, that aims to glean lessons and innovations from the war.

Ukraine's defense ministry said in mid-October that Initiative 13, which it called Model 13, was "ready for integration into the Armed Forces of Ukraine's situational awareness systems."

The AI model, it said, could predict Russia's drone strikes with "high probability."

Kyiv's forces use a battlespace management system called Delta, which feeds data from the front lines onto an interactive map that commanders can access via laptop, phone, or tablet.

The drone-predicting AI was developed by Ukraine's forces and NATO's Allied Command Transformation, with support from JATEC.

Ukraine's defense ministry said in its initial announcement that another AI model, dubbed No. 32, for planning artillery deployments was also in the works but is still in prototype stages.

Artificial intelligence has increasingly been making its way into Ukraine's war, from helping mid-range attack drones identify and strike targets to sending uncrewed ground vehicles to autonomously complete supply runs.

• Shopify CEO Tobi Lütke converted his underused wine cellar into an AI server room.
• Lütke's AI assistant researches topics and pulls together his Fitbit and calendar data.
• Lütke used heart rate data to identify his most stressful meetings and told the people involved.

The Shopify CEO didn't have enough wine to fill his cellar, but he had plenty of uses for an AI system.

Speaking with Alex Heath on an episode of the Sources podcast released on Wednesday, Tobi Lütke said he consulted his “inner 16-year-old” before converting the underused room into a home for computing hardware.

"So now I have a rack cellar, and it has like one of those GB300s from Dell in it," Lütke said. "So it is actually fully training models."

The GB300 refers to Nvidia's Grace Blackwell Ultra computing platform, which Dell has built into powerful computers designed to develop and run AI models.

Lütke said the setup allows him to run AI locally, working with sensitive information he can't send to an outside lab. The agent researches topics, prepares briefings, and gathers data about his daily life, he said.

“I think it's probably the most compelling product I've ever encountered in my life," Lütke said. "It runs a script that sort of makes a collection of all sorts of things, including my Fitbit data, and, you know, obviously has my calendar."

"It has a little app. It knows my geolocation," Lütke added. "For my hobby, I do a lot of health-style analysis, but then given that I have all this data, like I have heart rate data for, you know, for this, and I figured out who my most stressful meetings are with."

Tech bosses are finding increasingly personal uses for AI. Business Insider previously reported that Google CEO Sundar Pichai uses agents to follow internal discussions and computing usage. Apple CEO Tim Cook uses AI to summarize emails, while OpenAI's Sam Altman has turned to ChatGPT for parenting questions.

Lütke has also pushed employees to put AI to work. In a memo made public in April 2025, he said that all Shopify teams must demonstrate why they can't achieve their goals with AI before hiring new employees.

Asked by Heath whether he told the people he discovered were stressing him out, Lütke said: “Yeah. Totally.”

"It's like: here's the funniest thing I just find out — guess what?" Lütke said. "And then they didn't think it's nearly as funny."

• Microsoft lowered its 2026 capex forecast to $170 billion, down from $175 billion.
• CFO Amy Hood cited delays in third-party capacity as Wall Street scrutinizes returns on AI spending.
• Azure revenue jumped 48%, beating expectations, while Microsoft's headcount fell 2%.

Microsoft reduced its forecast for capital expenditure, a rare move in an industry that mostly continues to splurge more on huge AI investments.

The software giant said capex for the 2026 calendar year will be $170 billion from a previous forecast of $175 billion.

Chief Financial Officer Amy Hood gave the forecast on the company's fiscal first-quarter earnings call on Wednesday.

She said Microsoft was expecting new capacity to arrive from third-parties, but some of these "GPU-as-a-Service contracts" have been delayed until the second half of the fiscal year.

This likely refers to other providers of AI cloud services that use Nvidia GPUs to run their offerings. Microsoft has partnered with some outside providers to meet rampant AI demand.

Hood's comments could refer to a recent spate of AI data center project delays, which have mostly been caused by a lack of electricity, along with local opposition.

The CFO said Microsoft expects capex to accelerate in the second half of its fiscal year, as the delayed capacity kicks in.

Wall Street has become increasingly concerned the huge investments cloud companies are making to expand AI infrastructure. Microsoft last quarter kept its forecast unchanged, notable as many cloud giants have been steadily increasing their AI spending forecasts.

Overall, Microsoft reported $92.3 billion in revenue, up 19%, for the fiscal first-quarter ending Sept. 30. Azure cloud revenue was up 48% to more than $33 billion, surpassing Wall Street expectations.

Microsoft share were up about about 3% in after-hours trading following the earnings call.

Earlier this week, Microsoft offered a buyout to more than a quarter of its workforce. A Microsoft spokesperson told Business Insider on Wednesday that the company expects 5% of eligible employees at most to accept the buyout.

"We expect very few people to select this offer because of the growth opportunities here at Microsoft," Hood wrote in an internal memo viewed by Business Insider. Microsoft's headcount slipped 2% during the fiscal first quarter, the company reported.

Have a tip? Contact this reporter via email at astewart@businessinsider.com or Signal at +1-425-344-8242. Use a personal email address and a non-work device; here's our guide to sharing information securely.

• Microsoft is offering buyouts to around 60,000 employees.
• Chris Williams, former Microsoft VP of HR, said waiting for a better offer may or may not backfire.
• He shared what employees who are offered a buyout should consider and what they may not realize.

Microsoft is offering buyout packages to some 60,000 employees, or more than 25% of the workforce. That's a huge number in any terms, and it sent a signal and shock wave through the industry this week.

I'm the former VP of HR for Microsoft, and although I've been away from the company for some time and have no insider information, having been on the inside at Microsoft and in the industry for 50 years, here is what Microsoft employees should be considering — and what they may not realize about taking a buyout offer.

Your buyout offer may or may not be the best offer you get

During a buyout, position, tenure, level, and other factors are taken into account for each bespoke offer. The offer is likely to include compensation, benefit continuation, and possibly some acceleration or continuation of equity vesting.

That said, Microsoft might not offer more generous buyouts than severance packages, but companies usually offer better terms to people who take a voluntary separation than people who get laid off. That is not only to induce employees to go voluntarily, but also because the costs to the company are less if these workers voluntarily leave. There is less legal risk, time, and expense. It's also simply easier for the company to manage.

If you're offered a buyout, I usually find it could be the best offer you'll get. If it's not, you could be settling for the same offer as someone in a layoff and miss out on collecting unemployment.

A buyout can force some very hard career choices

If you get a buyout offer, looking at your budget should be your next step. Does the offer provide you with enough to retire? Enough to support an extended job search? The buyout can be a good forcing function for some very hard career choices.

Spend time looking at your cash flows, and pay particular attention to benefits. For many, healthcare costs are a substantial part of their budget. Weighing the buyout offer against continued coverage is often a tipping point in the decision.

Also consider any equity vesting dates you may have. Compare when you might vest with and without the offer, as those dates can mean a significant change in your investments. Whether you should take the buyout is highly dependent on both the offer and your situation. If you had retirement anywhere in sight, this offer might be just the trigger you need to move on. Perhaps the best offer you'll ever see.

However, if you need to continue working, there's a great deal more to consider.

Taking a buyout can give you a head start in a tough job market

The tech industry job market has been rough for the last year or so, and shows all the signs of getting worse. If the buyout doesn't afford you the ability to retire, you should plan on an extended period of job searching, maybe six months or more. It's tough out there. Be prepared.

One thing to consider if you need to continue working is that taking the buyout might give you a short head start in the job market. Assuming Microsoft and others follow this up with more layoffs like they did the last time, the technical job market will be flooded with quality people. Of course, you can search for a job while you have one, but taking the buyout gives you time for a full-time job search.

In any case, if a job search is in your future, don't delay. This market shows no sign of getting better.

Finally, if you're offered a buyout, don't ignore it. It's a sign. I'd read it as a clear signal from the company about your value, at least from their perspective. You can bet on the odds, but if follow-up layoffs happen, I'd bet that those who declined the buyout are somewhere on that list.

Buyout or not, some things to do

This buyout is the latest signal of a broad realignment of the job market, at least in the tech sector. Even if you aren't offered it or don't think you're at risk of later layoffs, you'd be wise to keep all your options open.

Get your résumé up to date. Spend time honing it, emphasizing the clearest set of skills you can. When a market is contracting, the safest bet is having a particular skill someone is looking for. When times are tight, the generalist is often the first person let go.

Work with and expand your network. Have coffee or a Zoom chat with everyone you know. Peers in other companies, even other industries, are invaluable when things get tough. Especially when the world is changing, and predicting the next hot spot is so hard, having as many warm connections across industries is a huge asset.

Don't go asking for jobs as much as keep those connections alive. You never know when someone's manager will say, "We need someone like that." And your friend can say, "I know just the person."

Whether you are offered a buyout, or whether you decide to take it or not, having a network of people to connect with is your single biggest career asset.

What to do if you get a buyout offer

Budget first, and outreach next.

Carefully decide your near- and long-term financial future, and make the decision from hard facts. Then connect with people you know across industries, far and wide. Those connections will be excellent sounding boards and could well provide the path to your next role, or the friends you need in retirement.

When contacted about this story, a Microsoft spokesperson pointed Business Insider to its statement on the official Microsoft Blog, which read in part, "Today, we're introducing a new Voluntary Exit Offer (VEO) to eligible employees within participating organizations." They added that "the VEO is similar to a standard severance offer, which varies by market. The offer includes payment based on completed service and base pay, as well as career transition services, which includes career coaching and job search assistance."

• An Anthropic manager says Claude has made traditional coding interviews a poor hiring signal.
• His advice to juniors: target business problems and build a focused prototype over a weekend.
• Google and Amazon are also reshaping engineering work and hiring around AI fluency and impact.

Out with the coding test, in with Claude.

That's the shift Felix Rieseberg, an Anthropic engineering manager, says he has made to his team's hiring process, arguing that AI has rendered traditional coding interviews a poor measure of who will succeed at work.

"The coding interview is dead, at least for my team at Anthropic," Rieseberg wrote in a Tuesday post on X. "I've seen enough to know that it no longer provides signal."

Rieseberg elaborated in a blog post, writing that there's "no point" in giving candidates coding challenges or asking them about programming languages like assembly.

"Claude is now better at both," he said of the company's own AI model.

Coding challenges were once a "weak but useful" predictor of success, Rieseberg said, but the most impactful builders on his teams are not necessarily the people who excel at such tests.

"I'm seeing engineers who are comparatively new fly past people who are probably better coders and have been around much longer," he wrote. "Hunger beats tenure."

AI is writing most of the code at Anthropic, and the shift tracks with changes underway at other Big Tech companies. Business Insider's Hugh Langley reported last week that Google is expanding pilots for software-engineering interviews that allow candidates to use Gemini to assess their AI fluency and prompting skills.

Amazon, meanwhile, told engineers on Tuesday to prioritize customer and business outcomes over raw AI usage. The company calls the moment the era of the "Renaissance Developer," in which human judgment and machine intelligence combine to build products faster.

How junior engineers can stand out

The rise of AI coding tools has fueled fears that junior developer roles will become harder to get, as AI takes on work like debugging and building simple features — tasks that once helped early-career engineers learn the ropes.

In his blog, Rieseberg shared advice for junior engineers — or, more broadly, people without organizational status or power. Rather than just clearing backlog tickets, he said they should identify the business problems their company is trying to solve.

Then, they should come up with a bold idea that could materially advance those goals and build a working prototype over a weekend. The project should communicate one compelling idea, he said, rather than bury several good ones inside an overly ambitious product.

Newer engineers should also lean into their outsider perspective. They may be more likely to spot broken user experiences or question ingrained assumptions about what is possible, Rieseberg said.

He pointed to two relatively new Anthropic engineers who built a graphical interface for Claude Code over a weekend; the first public version shipped around eight days later.

Rieseberg said he has not yet solved the problem of how to assess candidates who build effectively with AI.

For now, he said he is scanning resumes and portfolios for small prototypes that show "taste, empathy, and principled thinking" — qualities that AI researchers and tech leaders are increasingly putting a premium on in the age of AI.

• A startup founder said that leaner teams may not be the right approach for every company in the AI era.
• Chatbase CEO Yasser Elsaid said he plans to double his 40-person team within six months.
• AI models are "extremely smart and capable, but not fully independent," he told Business Insider.

As some companies race to use AI to shrink their teams, one startup founder is going against the grain by planning to hire more people.

Yasser Elsaid, the founder of customer-service AI startup Chatbase, said the company now has 40 employees and plans to double its team over the next six months.

"I might have been wrong about building with a lean team in the age of AI," Elsaid wrote in a post on X this month.

Elsaid previously embraced smaller teams. In a LinkedIn post in January, he said Chatbase, a San Francisco-based startup that lets companies build AI agents for customer service and support, had stayed at 17 full-time employees for much of its first three years, with most hires in engineering.

However, in comments sent to Business Insider, Elsaid said that he now considers his bet on lean teams "pretty wrong."

"I think it's a big mistake for founders to live in a future that doesn't exist yet," he said.

Elsaid said that more capable coding models, including Anthropic's latest Claude model Opus 5.5, initially made him think Chatbase could get by with fewer people.

But Elsaid said that this conclusion only holds if a startup keeps the same roadmap it had before AI. Rather than accept that code had become a commodity, he decided Chatbase should set its sights higher.

"If building is easier, build harder things," he said. "Most moats in software come from someone doing something hard for longer, and being right. That hasn't gone away."

'Not fully independent'

That view runs counter to a broader push toward AI-powered lean teams.

Block cut nearly half its workforce in February, with CEO Jack Dorsey saying AI and "smaller and flatter teams" were changing how companies are built.

Coinbase, meanwhile, cut 14% of staff in May, citing engineers' ability to ship in days what once took teams weeks; it is now testing smaller AI-native pods and "one person teams."

The same logic has powered the "tiny teams" startup trend: OpenAI CEO Sam Altman has predicted a one-person billion-dollar company, while Cursor grew from $1 million to $100 million in annual recurring revenue in under a year with fewer than 50 employees, according to private market research platform Sacra.

But Elsaid said the best models are not yet independent enough to replace people. They still need humans to decide what to build, give them direction, and make sure the result works for customers, he said.

"Code is largely right or wrong: it does what you want or it doesn't," Elsaid said. "But what to build, how it should work and what the experience should feel like is what makes your product different."

He estimated that about 90% of Chatbase engineers' work is now product work: researching, talking to customers, forming an opinion, and turning that into a specification an agent can code.

The rest requires engineers who can understand the code and make sure different AI agents are not working against each other, he said.

Chatbase is also hiring customer-facing staff, particularly in sales and customer success, he said, because enterprise customers still need to be heard and persuaded.

The models are "extremely smart and capable, but not fully independent," he added. "Maybe that changes and one day you won't need humans in the building process at all. But that isn't today."

• Amazon shifts AI focus to business impact, not just raw usage, in new developer guidelines.
• The company warns that AI metrics can create bad incentives, including “tokenmaxxing.”
• Managers are told to examine teams at both the highest and lowest levels of AI adoption.

Amazon just issued new guidelines for software engineers in the AI era: Prioritize business outcomes over raw usage.

Terk Terkowitz, Amazon's vice president and distinguished engineer, said in a memo published Tuesday that AI is changing software development more quickly and fundamentally than earlier shifts such as cloud computing. But engineers should be judged by their impact on customers and the business, he wrote, not simply how much they use AI.

"We all need to embrace this change and take the time to learn, or we risk being left behind," Terkowitz said.

AI coding tools and agents are upending the software industry and changing what it means to be a coder. Amazon calls this the "era of the Renaissance Developer," where human judgement and machine intelligence combine to create new products and services faster and more efficiently.

"This is a time of change for our industry, and a time of change for what it means to be a software engineer," Terkowitz wrote in the memo.

The updated guidelines reflect Amazon's growing effort to make AI adoption part of how it assesses engineers and engineering managers. The company has tracked measures including AI usage, token consumption, and lines of code internally, and used leaderboards to rank teams by some of those metrics.

But the memo warns that using such measures as targets can produce bad outcomes, including "tokenmaxxing" rather than better software or customer experiences.

"Using these metrics could, if used as a target instead of a measure, drive negative outcomes," Terkowitz wrote in the memo.

Amazon has been experimenting with ways to measure and encourage AI use across its engineering organization.

Earlier this year, the company was tracking engineers's use of internal AI tools and some teams used leaderboards tied to measures such as token consumption. The programs created friction among some employees, who worried that pressure to use AI could reward wasteful activity rather than productive work.

The new guidance offers Amazon's clearest public acknowledgment of that risk. Rather than lay out a detailed AI playbook that could quickly become obsolete, Amazon said its revised developer guidelines focus on long-term impact on customers and the business.

Still, the company said AI should be "at the center of software engineering practice" and urged engineers to incorporate the technology into their daily work.

Leaders are expected to use AI tools on production work themselves, understand where they help and where they fall short, and explain that judgment to their teams, according to the memo.

Terkowitz said Amazon has found that metrics such as token use, AI-tool adoption, and code reviews can correlate with the outcomes it values. But he said the measures must be used carefully. More token use can raise costs, while a focus on lines of code can create unnecessary complexity.

He also recommended that managers pay particular attention to teams at both ends of the AI adoption spectrum: those in the bottom 10%, to understand why they are not using AI effectively, and those in the top 10%, to determine whether heavier use is translating into faster delivery.

Terkowitz said Amazon's most AI-forward "frontier teams" have shown stronger productivity and delivery. He also said teams and individuals that are not adopting AI or increasing output metrics are falling behind.

But his memo left room for engineers whose work may not yet benefit from current AI tools. In those cases, he said, employees should document why and work toward a solution by sharing their findings with peers and managers.

"An engineer who can explain where AI does not yet help their work is demonstrating the judgement we are asking for," the memo said.

Have a tip? Contact this reporter via email at ekim@businessinsider.com or Signal, Telegram, or WhatsApp at 650-942-3061. Use a personal email address, a nonwork WiFi network, and a nonwork device; here's our guide to sharing information securely.

• Amazon launched a series of new devices this month, including Alexa tablets and Bee AI wearables.
• Rising memory costs forced "difficult trade-offs" on Amazon's device prices, sales, and profit.
• Amazon is seeking more Alexa revenue from enterprise sales, licensing, and ads.

Amazon launched a flurry of new consumer devices this month. Behind the scenes, rising memory costs are adding pressure to its already money-losing hardware business.

An internal planning document obtained by Business Insider shows Amazon forecasting a $612 million gross loss this year for its first-party devices business, compared with a leadership target of a $438 million loss.

The document said rising memory costs were "putting pressure" on the business's profitability, forcing "difficult trade-offs" among sales volumes, average selling prices, and gross profit.

The projected $612 million loss was at the gross profit level, meaning it did not account for all operating expenses, such as R&D, sales and marketing, and administrative costs.

The figures offer a rare glimpse into the pressure on Amazon's hardware business as it absorbs rising component costs. Other tech hardware companies, including industry leader Apple, face similar pressure from surging memory chip prices.

An Amazon spokesperson declined to comment.

A new wave of AI devices

Despite the pressure, Amazon continues to expand its consumer hardware lineup. Even if the company loses money on device sales, it can make money in other ways, such as subscription services, as more customers use its gadgets.

In October alone, the company introduced new Kindles, Alexa-powered tablets, Ring smart locks, and AI wearables under its Bee brand. The August planning document identified Bee and new smart glasses as strategic products still in an investment phase.

Amazon has historically used low-cost hardware to pull customers into a broader ecosystem of shopping and digital content. Rising component costs could make that strategy more expensive, forcing the company to raise prices.

The clearest shift is in Amazon's tablets. Earlier this month, the company phased out Fire tablets after 15 years and replaced the budget-friendly lineup with pricier Alexa Tablets powered by Alexa+. The new devices cost between $229.99 and $499.99, a notable increase from previous generations.

Not just Amazon

The planning document was prepared in August, the same month Amazon raised prices on products including Echo speakers, Kindle e-readers, and Fire TV devices.

The pressure extends beyond Amazon. Apple has raised prices on some Macs and iPads, while Microsoft and Sony have increased console prices. Dell, HP, and Lenovo also face higher memory costs.

The AI boom is helping drive the squeeze. Huge data center investments have increased demand for memory chips, while manufacturers have devoted more capacity to high-bandwidth memory used in AI servers. That has tightened supplies and raised prices for components used across consumer electronics.

Amazon is experiencing both sides of the boom. It expects roughly $220 billion in capital expenditures this year, largely to expand AWS data centers for AI demand, while higher component costs squeeze its consumer hardware business.

The rising cost of Alexa+

The August planning materials show Amazon reassessing both Alexa+ revenue and operating costs. They called for updated forecasts, expansion through enterprise sales and advertising, and more detailed AWS cost modeling.

Amazon is now confronting the cost of operating Alexa+ at scale. Business Insider reported in July that internal forecasts put Alexa+'s AWS cloud costs at roughly $1.7 billion in 2026, nearly triple the prior year's figure. Amazon was working to lower those costs by relying less on Anthropic's AI models and routing more requests to its own, cheaper systems.

A separate project called Moonraker, intended to give Alexa more advanced AI-agent capabilities, was projected to require more than $100 million in GPU spending this year. Amazon internally considered delaying or scaling back the initiative as one option for easing cost pressure, Business Insider previously reported.

Have a tip? Contact this reporter via email at ekim@businessinsider.com or Signal, Telegram, or WhatsApp at 650-942-3061. Use a personal email address, a nonwork WiFi network, and a nonwork device; here's our guide to sharing information securely.

• Cybersecurity experts found severe PHP bugs that could allow attackers to hijack web servers.
• AI is being used to exploit vulnerabilities, making Americans more susceptible to scams.
• Experts stressed basic cyber hygiene and scam awareness, especially for older Americans.

In just two days, AI exposed critical flaws in a programming language behind 70% of all websites. That may have opened the door for a rush of scams — especially those targeted to vulnerable populations like older Americans.

Last week, cybersecurity researchers uncovered two severe bugs in the PHP programming language. Attackers could hijack web servers and execute remote code. This prompted the PHP Foundation to release emergency patches to prevent widespread hacks, and governments, cybersecurity agencies, and businesses quickly moved to do damage control.

Though some experts said there could have been more problems had these flaws not been immediately addressed, such a crisis can happen again. As AI becomes more advanced and widespread, Americans are becoming more exposed and susceptible to scams.

Scams have been pervasive for decades, starting with physical mail and landline telemarketing. But they've gotten significantly more complex and deceptive, now integrating AI in ways that even tech whizzes may not spot. Older Americans are often the main targets of fraud, losing billions in fraud. An AARP survey from April found that 4 in 10 older Americans have lost money to fraud.

With attackers pouncing on any vulnerability to steal personal details or get money, five experts told Business Insider that now is the time to get ahead on scam prevention. Their quotes have been lightly edited for length and clarity.

Craig Newmark, founder of Craigslist, who founded the cybersecurity awareness campaign Take9

What are some of the most common scams targeting older Americans?

Hackers are using AI to ditch the typos and make scams harder to spot. AI also makes it easier to produce large numbers of targeted scams. This includes AI voice cloning and scans of social accounts for personal details. These scams are on the rise and hurting real people, including older folks like me.

How should families react to all the recent headlines about system vulnerabilities?

Don't panic, but don't ignore them either. It shouldn't be up to regular folks to outsmart scammers. The best thing for families to focus on is basic cyber hygiene.

Organizations like Google and Apple are working hard to block scams, particularly on phones. It's just as important that we prepare loved ones with the tools to protect themselves as scammers become more sophisticated.

What are ways in which you work to make people more cautious about scams?

I've been fighting scams for 25 years. I take them personally. That's why I founded Take9 in 2024 to help people understand how to protect themselves and stay safe online.

Scams are overwhelming, and people often don't know where to begin. My advice is to start with the basics: Take 9 seconds to pause before clicking on unknown texts or links.

Once they've got that down, try multi-factor authentication, password managers, passkeys, and single sign-on. Family codes are an effective method to keep your loved ones safe, too.

Don Grant, certified financial planner at Sabre Wealth and CFP Board Ambassador

What are some of the most common scams targeting older Americans?

Scams often involve impersonation of an institution or government agency like the IRS, Medicare, or Social Security. Fake technical support scammers contact you with claims of a breach at your bank or financial institution.

A good friend and client is in his mid-60s, is a single Notre Dame graduate, and is a retired university English professor. He talked incessantly about his gorgeous, accomplished girlfriend. She has sent him photos, but he had never met her in person, nor via video chat.

Fast forward to earlier this year. They are no longer an item. He reluctantly admitted that she was attempting to scam him. He did send her about $1,200 to help get her 9-year-old daughter into a music school. When she hit him up about 10 times so she could move, he figured it out and cut the ties. From what we pieced together, this was an AI-assisted romance scam.

How should families react to all the recent headlines about system vulnerabilities?

A security breach at an institution doesn't automatically mean that your accounts have been hacked. However, it does mean it's time to change your login credentials.

Set up account notifications for withdrawals, transfers, password changes, and anything to do with your accounts. Never respond to an unsolicited email from a site or financial institution with which you do business. Don't call the number provided in the communication or click the provided link. Take the time to log into a verifiable site and call a verifiable number with questions.

What are ways in which you work to make people more cautious about scams?

Communication and education are critical. I have open conversations about financial fraud, anonymously citing actual situations where a scammer attempted to separate them from their money. Recounting others' experiences seems to really drive the point home. We also insist that all clients designate a trusted contact we can reach if we notice unusual activity or have concerns about possible financial exploitation.

Removing the stigma that they are "stupid" if they are scammed is paramount. Many victims are embarrassed to admit they've been deceived, and that embarrassment can prevent them from seeking help.

Alex Jacquez, Groundwork Collaborative's senior vice president of policy, advocacy, and research

What are some of the most common scams targeting older Americans?

The FTC recently released its Protecting Older Consumers report, describing fraud in 2025. The top fraud types by incident report numbers for older people include online shopping, business impostors, government impostors, and investment scams. These, plus romance scams, are the top fraud types by total dollars lost for older people.

Scammers use AI to create deepfakes, clone voices, and send phishing messages. These types of fraud are not new, but AI makes them increasingly convincing and enables them to be carried out on a larger scale. Scammers also use AI in predictive targeting to find the most susceptible individuals.

In March 2026, Interpol warned that "AI-enhanced fraud is 4.5 times more profitable than traditional methods" because AI agents can complete the fraud on behalf of the scammers.

How should families react to all the recent headlines about system vulnerabilities?

Families should see the recent headlines about system vulnerabilities as part of a larger problem. Companies collect lots of information on users and fail to protect it. Now, consumers are exposed to data breaches and sophisticated scams.

What are ways in which you work to make people more cautious about scams?

We put out material for consumers and policymakers highlighting how companies are using more sophisticated techniques and technologies to rip them off.

While there are best practices people can follow to remain alert and vigilant against common scams, especially in the AI age, we believe that any real fix must be systematic and undertaken by policymakers. We need companies to better protect consumer data, stop collecting irrelevant data, and stop using it to squeeze consumers for every last dime.

Amy Nofziger, senior director of victim support, AARP Fraud Watch Network

What are some of the most common scams targeting older Americans?

Impostor scams remain the number one category reported to the AARP Fraud Watch Network Helpline. They will say there is a problem with your account and need to remotely access it, or they will say they are with the courthouse and you missed jury duty and demand payment to avoid being arrested.

AI hasn't changed human nature, but it has changed the scale, speed, and sophistication of scams. Criminals are still exploiting the same emotions they always have: fear, trust, loneliness, hope, curiosity, and urgency. We used to tell people to look for red flags like misspellings and poor grammar in emails, but with today's AI tools, those clues are often gone.

How should families react to all the recent headlines about system vulnerabilities?

It's understandable to be concerned by headlines about system vulnerabilities and data breaches, but the key is to focus on what you can control. The goal isn't to live in fear. It's to stay informed, slow down when something feels off, and ask for help. We don't have to do this alone.

What are ways in which you work to make people more cautious about scams?

AARP Fraud Watch Network helps individuals and families protect themselves from scams. It's a free resource available to everyone. We provide fraud prevention information, online tools, the award-winning Perfect Scam podcast, and a Fraud Watch Network Helpline that receives more than 100,000 calls each year.

Ali Tager, vice president of external affairs, National Cryptocurrency Association

What are some of the most common scams targeting older Americans?

Some of the most common online finance scams are relationship scams and impersonation. Giveaway scams are when scammers impersonate public figures, promising "free" money or crypto if users start by sending a small amount first. In phishing scams, fraudsters pose as a friend or a crypto exchange and ask victims to click links or share login details to steal private information.

Pump and dump scams occur when groups inflate the price of lesser-known crypto coins via online hype. When enough people join and spike the price, fraudsters quickly sell off their holdings for a profit, leaving others in the dust.

How should families react to all the recent headlines about system vulnerabilities?

Stay alert, not alarmed. Whether it's your bank, your Venmo account, or your crypto account, security should always be top of mind. Protecting yourself starts with recognizing common scams and practicing smart online habits.

For crypto holders, guard your private key or recovery phrase like you would your bank account number and password. No legitimate company will ever ask for it, so never share it, and keep physical backups in secure locations. Watch for small clues like odd sender addresses or sloppy spelling, and remember that deepfakes and fake websites can look real at first glance.

What are ways in which you work to make people more cautious about scams?

A scammer's greatest weakness is an informed community. The NCA is dedicated to helping everyday Americans understand and use crypto with confidence. Research shows that 72% of crypto holders cite scams and security as a concern. However, 76% of crypto holders say it has positively impacted their lives overall, while just 3% report negative impacts. Of those, about one-third are scam/fraud-related.

• Nate Soares said he stopped adding to his 401(k) because AI may kill us within a decade.
• The MIRI president said humanity has a more than 50% chance of dying within a decade.
• Elon Musk said AI could make retirement savings irrelevant by creating an age of abundance.

Saving for retirement is a bet on the future. Nate Soares, the president of the Machine Intelligence Research Institute, isn't so sure AI will let us get that far.

"I do not save for retirement right now. I have a 401(k) from when I worked at Google, but I haven't contributed for a decade," Soares told The Atlantic CEO Nicholas Thompson in a recent interview for Thompson's "The Most Interesting Thing in AI" series.

Soares' decision reflects a grim view of the AI race.

"Godfather of AI" Geoffrey Hinton and Anthropic researcher Evan Hubinger have, respectively, put the odds of AI killing everyone within a decade at 10% and more than 10%. Soares is more pessimistic: He told Thompson he believes there is a greater-than-50% chance AI causes human extinction in less than a decade.

Soares, who is also the coauthor of "If Anyone Builds It, Everyone Dies," has spent more than 12 years working on AI alignment, the field focused on ensuring powerful systems act as intended. Before joining MIRI, the Berkeley-based nonprofit focused on AI risks, he worked as an engineer at Google and Microsoft.

Soares said that a sufficiently capable AI could develop goals at odds with humanity's, evade oversight, improve itself, and build the automated infrastructure needed to operate without people.

From there, he said, it could reshape the planet around its own objectives, potentially making it uninhabitable for humans — a warning he shared with Business Insider last year.

A 401(k) for the apocalypse?

The question of how AI will shape the future has also prompted Elon Musk to question whether retirement savings will matter. Musk's vision of the future, however, is radically different than Soares'.

In January, the Tesla and SpaceX CEO said people should not worry about "squirreling money away for retirement" in 10 or 20 years, predicting advances in AI, robotics, and energy will make retirement savings "irrelevant." (Personal finance experts and AI gurus told Business Insider the opposite: You really should be saving for retirement.)

Musk has said that supercharged productivity could create an age of abundance, in which everyone has access to a "universal high income," cheap goods and services, and better healthcare and education.

Musk has also said money could become irrelevant in a future where AI and robots outperform humans at virtually every task.

For Soares, the prospect of a vastly capable AI is an existential danger, not a route to plenty.

Echoing calls from tech leaders to "pace the frontier" of AI development, Soares said governments should halt the training of general-purpose AI systems. They should focus on developing technology for applications like cancer research and safer self-driving cars, he said.

The solution, he told Thompson on Wednesday, "is to not build the thing that would kill everybody."

"I do everything I can to stop this eventuality," he said. "I don't think it's inevitable."

• Simile raised $300 million to build models that simulate consumer behavior.
• CEO Joon Sung Park said users can define an audience and ask simulated consumers questions.
• Park said CVS and companies in finance and entertainment are already working with Simile.

Companies could soon have a new way to find out your shopping habits — by asking your digital clone.

Joon Sung Park, Co-Founder & CEO of Simile, an AI startup, told show host Molly O'Shea on the Sourcery podcast released on Monday that the company is building simulations of people that businesses can quiz about everything from products to marketing campaigns.

Park said the company is "creating a foundation model of human behavior that basically can take as input a description of a person or a group of people and then simulate what these people might do in different contexts," Park said.

"So they get access to this web interface where they can basically come in and define two things," Park added of existing users. "They can say who they want to talk to and what questions they want to ask. And really they can ask anything, and right now they can ask any number of questions."

Park said Simile raised $300 million in total at a $2 billion valuation.

Simile’s pitch comes as retailers turn to AI to get shoppers to spend. Amazon and Walmart have rolled out AI shopping assistants that answer questions, compare products, and recommend items, while retailers use customer data to personalize what shoppers see. Executives at both companies have said their AI tools are helping boost online sales.

Park said CVS has worked with Simile to better understand consumers, including their responses to store layouts and messaging. The startup also works with financial institutions and entertainment companies, Park said.

Simile maintains simulated panels for different industries and collects additional data when customers want to understand a specific population. Park said Simile's partnerships with organizations, including Gallup, which have access to actual humans who have agreed to participate, provide it with data from more than 40 million people globally. He said the company also uses behavioral information, such as transaction and mobility data, to train its models.

“If we can create simulations of people and be able to ask any questions at any time," Park said, "That becomes a superpower for these companies."

• Anthropic shows strong profit potential, with revenue growing way faster than expenses.
• Anthropic's revenue per employee exceeds $6.6 million, outpacing estimates for Meta and Google.
• Anthropic has a surprisingly diverse customer base and huge cash reserves.

For years, one of the biggest unanswered questions about the artificial intelligence boom has been surprisingly simple: Can the companies building the world's most powerful AI models actually make money?

After spending hours poring over Anthropic's IPO filing on Monday, we finally have an answer: Yes. And potentially lots of it.

A couple of Chinese AI labs went public earlier, revealing appalling financial results that sowed doubt about the industry's economics. Now Anthropic has provided a much more compelling demonstration of how this business can work at enormous scale.

The numbers are remarkable. Anthropic generated $11.6 billion in revenue during the second quarter of 2026, up 1,368% from a year earlier. During that same period, spending on computing and infrastructure increased 300%, while other expenses rose 741%

So, the company is bringing in money much faster than its most important costs are rising.

For every dollar Anthropic spent on computing in the second quarter, it generated $2.08 in revenue, up from just 40 cents in the first quarter of 2025. That's a more than fivefold improvement over five quarters.

The result: Anthropic recorded $974 million in quarterly operating profit, representing an 8% margin. Its adjusted operating profit margin reached 19%, compared with -129% a year earlier.

"Frontier-lab unit economics can turn positive at scale," Mark Mahaney, a top tech analyst at Evercore ISI, wrote in a research note on Monday.

Better than Meta and Google

There's more evidence that Anthropic has built an unusually efficient business.

Over the year ending September 30, revenue increased roughly 1,200%, while its employee count grew about 170%. Revenue per employee jumped from $1.4 million to more than $6.6 million.

For comparison, Mahaney estimated that Meta generates about $3.2 million in annual revenue per employee, while Google generates about $2.5 million.

That comparison isn't perfect. Anthropic relies heavily on cloud providers, whose employees handle much of the computing infrastructure and distribution work. The improvement is still striking.

Customer surprise

Another surprise: Anthropic doesn't depend as much on a handful of enormous customers, as some investors had feared.

The company had more than 11,900 customers generating at least $100,000 in annual revenue as of September, compared with over 950 a year earlier. More than 1,900 customers spend at least $1 million annually.

No single customer accounted for more than 10% of revenue during the first nine months of 2026. Its five largest customers represented about 21% of third-quarter revenue, down from roughly 36% in the first quarter.

"Concentration may be better than feared," Mahaney wrote.

Cash cushion

Anthropic has a formidable financial cushion. Its June 30 balance sheet showed more than $130 billion in cash and short-term investments.

And Mahaney estimated that the company has burned through just $9 billion since it started more than five years ago. That's a long runway.

Now, don't ask me whether all this makes Anthropic worth $2 trillion, $1.5 trillion, $1 trillion, or considerably less. If I knew that, I'd be making a lot more money as an investor. And even professional investors don't know.

An $800 billion caveat

Some enormous caveats: Much of Anthropic's revenue comes from customers paying for what they use, rather than signing long-term contracts. Meanwhile, some of its computing agreements require minimum payments whether or not it needs all the capacity.

Mahaney estimates that Anthropic has entered into nearly $800 billion in computing arrangements, although only about $266 billion was fixed or guaranteed as of June.

That's still a staggering amount.

"The commitments assume revenue compounding that is not contracted," Mahaney wrote.

Translation: While the company has committed to spending hundreds of billions of dollars on AI computing infrastructure, it hasn't secured the future revenue needed to pay for all of it.

Importantly, most of the potential $800 billion exposure isn't fixed. Many arrangements depend on capacity being delivered, can be adjusted, or include cancellation provisions.

In-house compute and capex

Another wrinkle: Anthropic warned Monday that its spending will increase as it shifts toward leasing AI chips and operating more of its own computing infrastructure, rather than relying entirely on Amazon, Google, and other cloud providers.

That could make future cash flow considerably less attractive, even if operating profits continue to improve.

Still, the central takeaway from Anthropic's IPO filing is difficult to ignore.

For years, investors have wondered whether selling AI could become a genuinely profitable business, rather than an extraordinarily expensive scientific experiment.

Anthropic has now shown that it can. The next question is whether it can keep doing so as the bills get much, much bigger.

Sign up for the AI Insider newsletter here. Reach out to me via email at abarr@businessinsider.com

• Microsoft is offering buyouts to about 60,000 employees.
• The offers are individualized to each employee based on their level, location, and tenure.
• Business Insider reviewed offers to show what employees get when they leave the company.

As Microsoft sends out its buyout offers to more than a quarter of its workforce, specific offers reviewed by Business Insider reveal exactly what employees get if they choose to leave the company.

The offers are individualized for each employee based on their level, location, and tenure. Those reviewed by Business Insider confirm that the US buyouts align with the company's US standard severance offers. A Microsoft spokesperson said about 60,000 employees would be eligible for the buyout.

Microsoft uses a levels system to represent seniority, from entry levels starting at 59 to senior levels at 70, or vice president. The company offered buyouts to employees on select teams in the US, Canada, India, and the UK at level 70.

Offers reviewed by Business Insider show employees at internal levels 64 and below were offered one week of base pay per six months of service, and employees at levels 65 to 67 were offered two weeks of base pay per six months of service.

Buyout offers also include continued regular stock vesting and health insurance. The standard stock-vesting severance offer is six or 12 months for levels 67 and below, based on years of service.

Eligible employees will have until early December to decide whether to take the offer and exit at the end of the year, and there are no plans for another buyout at this time, a Microsoft spokesperson said.

Standard severance is capped at 39 weeks of base pay for levels 67 and below, while the standard severance offer for level 68 includes 12 months of salary.

Earlier this year, Microsoft rolled out a voluntary retirement program for employees. About 7% of Microsoft's US workforce, or about 8,750 people, were eligible at the time. As Business Insider previously reported, Microsoft's voluntary retirement program offer also aligned with the company's standard severance offer, including continued regular stock vesting for six or 12 months, based on years of service.

Earlier this year, another tech giant, Google, offered buyouts to staff in its business unit, Business Insider previously reported.

As more tech companies offer buyouts, some human resources professionals previously told Business Insider that this approach could soften job cuts, and also that when they're made available to workers of all tenures, companies risk losing talent.

Have a tip? Contact this reporter via email at astewart@businessinsider.com or Signal at +1-425-344-8242. Use a personal email address and a non-work device; here's our guide to sharing information securely.

• Microsoft is offering buyouts to around 60,000 employees.
• The offers will go out later today, and resemble Microsoft's standard severance offer.
• The buyout offer could help Microsoft cut costs and reshape its workforce for the AI era.

Microsoft is offering buyouts, according to a memo Chief People Officer Amy Coleman sent to employees Monday.

"The Voluntary Exit Offer reflects an important part of how we're evolving our approach to change: creating greater flexibility and, where possible, more ways for employees to shape their next chapter," Coleman wrote.

Microsoft's voluntary exit offer is the company's latest attempt to slim its workforce without another round of involuntary cuts. After spending heavily on AI infrastructure and reducing headcount this year, Microsoft is offering remaining employees a chance to leave, testing whether voluntary departures can cut costs.

Eligible employees in the U.S., Canada, India, and the U.K. at level 70, which includes senior leaders at the company, and below, will receive an email later Monday with the terms of the offer. Only some teams are participating, excluding sales organizations, some engineering, and some corporate functions.

Around 60,000 employees — or more than a quarter of the company's workforce — about will be eligible, a spokesperson told Business Insider, and the terms will be similar to Microsoft's standard severance.

The US severance package offered when Microsoft laid off employees earlier this year included a minimum of 60 days of base pay, during which employees will remain on the payroll, and, for most employees, up to a maximum of 39 weeks of base pay, based on seniority and tenure.

Employees at internal levels 64 and below received one week of base pay per six months of service, for example, and employees at levels 65 to 67 will receive two weeks of base pay per six months of service, according to offers reviewed by Business Insider.

Eligible employees will have until early December to decide whether to take the offer and exit at the end of the year, and there are no plans for another buyout at this time, Microsoft's spokesperson said.

The buyout offer follows Microsoft's Voluntary Retirement Program, which offered a retirement buyout to about 9,000 eligible employees based on age and tenure.

About one-third of eligible employees took the buyout, in line with expectations, a person familiar with the turnout told Business Insider in June. That buyout allowed Microsoft to lay off a lower percentage of its workforce compared to last year, the person said at the time.

Read the memo:

Introducing a Voluntary Exit Offer

Today, we're introducing a new Voluntary Exit Offer (VEO) to eligible employees within participating organizations. This offer gives eligible employees who may be considering a future outside the company a new choice to make that transition now with support. It also gives participating orgs another way to adapt their workforce to changing needs.

The Voluntary Exit Offer reflects an important part of how we're evolving our approach to change: creating greater flexibility and, where possible, more ways for employees to shape their next chapter. This builds on what we learned earlier this year through the Voluntary Retirement Program.

What employees should know

This offer will be available to eligible employees in participating teams in Canada, India, United Kingdom and United States at level 70 and below. This is a time-limited, voluntary program that gives those who are eligible the opportunity to leave Microsoft with financial support and career planning resources. Some organizations are providing the Voluntary Exit Offer, and some are not. Participation varies based on each business's priorities, customer commitments, and workforce needs.

Eligible employees will receive an email with a link to their Voluntary Exit Offer details within the next few hours.

Looking ahead

The Voluntary Exit Offer is new for us. We'll learn from it. At the same time, we'll keep helping people focus on the work that matters most, build new skills, and, where it makes sense, find new opportunities across Microsoft. There is no one-size-fits-all approach for every team or every employee. These are a few of the ways we're trying to meet different business needs while continuing to create opportunities for our own people to grow and contribute.

My promise is to continue to provide clarity, listen and learn, and keep improving.

Thank you for everything you continue to do for our customers, Microsoft, and for one another.

Amy

Have a tip? Contact this reporter via email at astewart@businessinsider.com or Signal at +1-425-344-8242. Use a personal email address and a nonwork device; here's our guide to sharing information securely.

• Micro1 pledged to spend $1 billion on licensing companies' data for AI developers.
• The startup snags Slack messages, Canva files, Google Docs, and as much other data as it can get.
• By selling company data, CEOs have a new revenue stream that popped up this year.

As AI labs go on the hunt for real-world data, Micro1 is throwing everything at getting it for them.

The startup is one in a new class of fast-moving data brokers, gobbling up companies' internal data and packaging it for the likes of OpenAI, Anthropic, and other model developers. The executives who sell their companies' data to Micro1 find themselves suddenly participating in a new sector of the economy, where terabytes of mundane workplace interactions can be worth hundreds of thousands of dollars.

Ilya Brotzky, the CEO and cofounder of the recruiting startup VanHack, told Business Insider he was shocked when he filled out a Micro1 survey about his company's data and received a "substantial" offer.

"I didn't believe it, to be honest," he said. "I felt like it was a little too good to be true."

Micro1 says it pays $100,000 to $2 million for a single midsize company's data. It recently announced plans to spend $1 billion on the project over the next year. Brotzky and the other executives Business Insider spoke to for this story wouldn't share the exact amounts their companies received.

For Micro1, the new focus on workplace data matches a broader evolution in the industry, as Business Insider has written about in The AI Data Grab, a series about how tech giants are increasingly turning to workers, users, and the physical world to get data to train AI.

First, AI companies gobbled up text from the internet and corpora of old books. Then they realized that humans grading AI responses could improve their models. And now, companies are using complex simulations, built on real-world data, to improve the tenacity and skill of their AI.

Gartner estimated that spending on AI data is likely to surge from less than $1 billion last year to almost $6.5 billion in 2027. Micro1's chief revenue officer, William Almeida, told Business Insider in September that his company has been explaining to prospective customers why their data is suddenly so valuable for those simulations.

"Every time I talk to companies, I tell them, 'Hey, I see this as a booming business,'" Almeida said.

Slack DMs, Google Docs, Canva files: Micro1 wants all it can get

Micro1 has three ways of working with AI model developers.

First, it can take a company's raw data, transform it to protect privacy, and license it to a lab. In addition, Micro1 provides an in-progress AI agent to companies and gets feedback on how it works. The third route is the simulation: Micro1 takes a company's data up to a certain timestamp, freezes it, and then provides that "environment" to a model developer trying to teach its AI how to work from existing context.

Micro1 pays more for larger, more complex datasets, as well as for data that represents a unique type of work or knowledge. It usually wants to license a company's entire data corpus, so that an AI model can replicate what it would be like to be a human employee.

"The connection, the diversity, the traceability of all of these data points, they matter a lot when you think about model training," Almeida said.

Brotzky also received an offer from AI data startup Mercor to buy his company's data, though he opted for Micro1. Once Brotzky dug into Micro1's privacy process, he said he felt comfortable sharing all the data they requested. That included VanHack's Google suite, Confluence data, Slack messages, and customer relationship management software.

Grant Gooley, the CEO of the marketing agency MyPortal Marketing, also views his deal with Micro1 as an innovation. He has sold internal design files from Figma and Canva, as well as checklists from Monday.com.

"When we were building these processes years ago, I didn't think that we would be using it to support AI learning," Gooley said. "But doing that, it's very fulfilling."

CEOs pitch the deals to their workers

Micro1 and its rivals are full steam ahead. Still, there are open questions. Former and current employees aren't likely to get any compensation from — or a voice in — their work becoming AI fodder. Ifeoma Ajunwa, a law professor at Emory University, previously told Business Insider that data sales are an "ethically gray area."

When pitching to potential customers, Almeida tells them this business is in the early days.

"If you wait too much, you won't get the premium that the first companies are getting," he said.

VanHack's team got a confidence boost from the Micro1 deal, Brotzky said, because of the additional revenue stream it represents

Gooley said he's kept MyPortal Marketing's employees in the loop about the deal with Micro1 and hasn't heard any complaints.

"I told my team that this is something that you're going to tell your grandchildren about one day, because we are literally supporting the future growth of marketing agencies," Gooley said.

Have a tip? Contact this reporter via email at scouncil@businessinsider.com, or over text, Signal, Telegram, or WhatsApp at 415-757-8198. Use a personal email address, a non-work WiFi network, and a non-work device; here's our guide to sharing information securely.

• A version of this story originally appeared in the AI Insider newsletter.
• Sign up for Business Insider's weekly AI Insider newsletter here.

The AI data center boom looks like it's running into serious trouble. Projects are being delayed, canceled, or abandoned. Electricity shortages are holding everything up.

But there's a twist: A lot of these data centers were probably never going to get built in the first place.

Bernstein Research analyzed thousands of planned US data centers and concluded that only about 35% of the announced capacity is likely to become reality.

The industry has a wonderful name for some of the more dubious projects: "Bragawatts."

These are enormous data centers announced by companies that often have little experience building them. Some haven't secured electricity, construction permits, or even customers.

Then there are the "ghost projects." Developers submit multiple requests for electricity connections, hoping at least one location works out. In Texas, these requests have reached more than five times the state's existing electricity supply.

No wonder the grid looks overwhelmed.

This doesn't mean the AI boom is fake. Far from it.

My colleague Eugene Kim exclusively reported this week that even Amazon, one of the world's most experienced data center builders, is struggling to get enough electricity to power its expensive AI equipment.

But there's a big difference between Amazon's very real problems and an inexperienced developer announcing a gigantic project that may never happen.

The AI data center boom is real, and so are the challenges of building all these facilities. But the sheer number of speculative projects may be making those challenges look bigger than they truly are.

Sign up for the AI Insider newsletter here. Reach out to me via email at abarr@businessinsider.com

• Pilots have faced the risk of deskilling for decades because of their reliance on autopilot.
• Steven Shaw of King's College London said they can show workers how to stay sharp with AI.
• "Even experts are at risk of deskilling," he said.

Pilots have relied on autopilot for decades. Their experience could offer a lesson for workers increasingly relying on AI.

Steven Shaw, an assistant professor at King's College London and coauthor of a widely cited study on "cognitive surrender," drew a parallel between pilots' reliance on automation and workers outsourcing their thinking and delegating their work to AI.

"We've had autopilot in planes for a very, very long time," Shaw said, speaking at a recent OxGen Summit panel. "So they learn to fly manually first, and then they learn to use autopilot strategically."

Some federal aviation guidance encourages pilots to occasionally switch off autopilot to maintain their manual flying skills.

There is a long track record of this. In a 2013 safety alert, the Federal Aviation Administration warned that "continuous use of autoflight systems could lead to degradation of the pilot's ability to quickly recover the aircraft from an undesired state."

The agency reiterated the importance of manual flying in a 2017 alert, recommending that airlines encourage pilots to practice without autopilot periodically when conditions permit.

"Even experts are at risk of deskilling," Shaw said, referring to the dangers of repeatedly outsourcing cognitive tasks to AI.

Shaw's warning comes as some tech leaders, AI researchers, and politicians have cautioned that overreliance on AI could weaken workers' underlying cognitive abilities.

Speaking at More Perfect University's "Students vs. Billionaires" town hall in Maine on Thursday, Rep. Alexandria Ocasio-Cortez warned that "AI's goal is to take your thoughts away from you, your ability to think away from you so that you become dependent on a machine."

OpenAI CEO Sam Altman said in July that one of AI's most overlooked risks was "cognitive atrophy," while Arthur Mensch, CEO of French AI startup Mistral, said last year that the technology's biggest risk could be "deskilling."

Think before you prompt

Early research is showing how readily people outsource their thinking to AI.

Vivienne Ming, an AI researcher and theoretical neuroscientist, told Business Insider in March that one of her experiments, involving 72 participants, found that roughly 90% to 95% either relied on AI to generate answers or used it to validate their assumptions, rather than challenge and deepen their thinking.

"The overwhelming trend is substitution," she said.

Similarly, Shaw's research suggests that the danger is that people increasingly accept AI's answers without thinking critically.

In experiments involving 1,372 participants, people accepted AI recommendations nearly 80% of the time even when the answers were deliberately incorrect.

Shaw calls this phenomenon "cognitive surrender."

His advice for avoiding it is straightforward: develop your own judgment before asking AI for help.

"Just think first and then go to the prompt," he said during the OxGen Summit panel.

Starting with AI, he warned, can narrow the range of ideas people consider independently.

Other experts have offered similar advice. Geetha Rajan, a former AI adoption leader at PwC, previously told Business Insider that she writes her own first drafts before using AI to check facts and challenge her work.

Ryan Deiss, CEO of DigitalMarketer, advocates a "10-80-10" rule: humans handle the first 10% of a task, AI takes on the middle 80%, and humans return for the final 10% to exercise judgment and refine the result.

• Everyone complains about AI-generated junk.
• But decades of sloppy human-written software may be a much bigger problem.
• This week's "Black October" freakout was caused by human slop, not AI.

The internet has a slop problem. And for once, we can't blame AI.

Most of us spend hours online every day without thinking much about the software holding everything together. We assume somebody, somewhere, has made sure it all works properly.

So it probably came as a shock this week to discover just how vulnerable some of that software is.

The cybersecurity freakout dubbed "Black October" has inspired warnings about AI-powered hackers, digital catastrophe, and even advice to stockpile cash and canned food.

At first glance, AI looks like the villain. But this is largely a human-made mess.

Here's the problem.

The internet runs on decades of software, some of it containing security weaknesses that have been overlooked, tolerated, or simply left unfixed.

For years, that was a manageable gamble. Fixing software takes time and money. Exploiting obscure weaknesses often required considerable expertise. Companies could convince themselves that the risk wasn't worth worrying about.

AI is changing that calculation. Someone who once needed years of training to hunt for weaknesses can now get help from an AI model. These tools can also examine software at speeds that humans can't match.

Suddenly, all those problems we've been ignoring look more urgent.

"Over the last 10 years the risk acceptance process has morphed from a way to accept risks to enable the business, to a way to justify never fixing security issues," Joel de la Garza and Zane Lackey, partners at a16z, wrote on Friday.

That's quite an indictment. The result is a recent surge in published vulnerabilities.

This looks like AI is creating new security problems. But it's actually AI exposing weaknesses that humans left behind.

Slop irony

The irony is delicious. We've spent the past couple of years complaining about "AI slop." Turns out human slop can be pretty consequential, too.

"The funny thing is everyone was saying AI slop code will cause a cyber crisis but it's actually all the old human slop causing it and the AI is just cleaning it up," software engineer Ben Kaufman wrote on X on Thursday.

"Humans have been the ones writing and reviewing each other's code for decades," tech founder Anant Jain wrote on X. "If we're being intellectually honest and holding ourselves to the same bar we hold AI to, we'd have to admit that we haven't been all that good at it either."

Humans are sloppy

Anyone who's worked with other humans for a decade or two knows exactly how this happens.

It's Friday afternoon. You've almost finished a project. You could spend another few hours making it perfect, or declare it 90% done and head off to that dinner party. We've all been there.

AI agents don't have dinner parties. They don't get tired on Friday afternoons. They can keep searching, testing, and trying again. That persistence is now colliding with an internet built by busy, imperfect humans who often settled for "good enough for now."

A tragedy of the commons

There's another wrinkle here, though, and it's an important one.

Much of the software underpinning the internet is freely available for anyone to use. Some of it is maintained by volunteers or small groups of experts, even as enormous technology companies depend on it.

In 2020, the cartoonist Randall Munroe captured this absurdity in an xkcd comic. Modern digital infrastructure still rests precariously on tiny foundations.

It's funny because it's uncomfortably close to reality. Take the specific PHP problems from this week. These flaws have existed for years, and the PHP Foundation acknowledges that limited resources and disagreements over fixes contributed to delays in fixing them.

Tom Occhino, chief product officer at Vercel, brought up Munroe's cartoon during my recent interview with him.

He explained that AI is finding weaknesses even in software that the industry has treated as dependable for decades.

This is partly a version of the tragedy of the commons: Everyone benefits from a shared resource, but no one has sufficient incentive to take responsibility for maintaining it.

And let's be clear: The volunteers keeping essential software running aren't the villains here. The companies benefiting from their work have a much greater responsibility.

Occhino thinks AI could help solve the problem it is exposing.

He described an idea being discussed at Vercel: Create a foundation or coalition that funds efforts to identify vulnerable software and replace it with safer alternatives.

"We need a consortium. We need a coalition," he told me.

He suggested companies could contribute computing resources and AI tools to the effort. Nothing has been formally announced, but the idea is compelling.

Indeed, OpenAI's Patch the Planet effort is an example of what's possible.

Rather than simply using AI to create more software, the industry could use it to repair what already exists.

Occhino put it nicely: Companies need to spend less time chasing new ideas and more time asking how to make existing systems "just work really, really well."

That's a better response than panicking about the imminent collapse of the internet.

AI didn't create decades of neglected software. We did. Perhaps it's time to clean up our own slop.

Sign up for the AI Insider newsletter here. Reach out to me via email at abarr@businessinsider.com