• Microsoft is among the latest to cut middle management jobs.
• Tech giants like Intel, Amazon, and Google are also flattening structures for efficiency.
• Experts warn that while flattening can speed decisions, it is possible to take it too far.

Companies are shedding bloated layers of management in an attempt to reduce bureaucracy. Some employees are applauding the move, known as flattening the middle, in the hopes of getting faster and boosting efficiency.

Microsoft said Tuesday it's slashing around 6,000 employees. While the days since have made it clear many of those cut were individual contributor-level engineers, executives previously told BI one motivation behind the recent cuts was to increase managers' "span of control," or the number of reports per manager.

Intel announced a great flattening last month, emphasizing more time in the office, less admin, and leaner teams.

"The best leaders get the most done with the fewest people," said the chip giant's new CEO, Lip-Bu Tan, in a memo to staff.

Amazon has also increased the ratio of individual contributors to managers. They call it a "builder ratio." Google CEO Sundar Pichai told staff late last year that the company cut vice president and manager roles by 10% as part of an efficiency push. Meta has been at it for years, with CEO Mark Zuckerberg writing in a 2023 memo, "flatter is faster."

The risk is that these companies cut too many managers, leaving the remaining folks with too many direct reports.

But for now, it appears to be a risk companies are willing to take.

Agility and expertise

The logic of cutting from the middle to speed up is sound, management experts say.

"You can't go faster and be more connected to a larger ecosystem if you're having to go up and down a hierarchy for every decision," Deborah Ancona, a professor of management at Massachusetts Institute of Technology, told Business Insider.

While some companies have been trying for decades to zap management layers, there's a new urgency to do so. Businesses exist in "an exponentially changing world," Ancona said.

Dell executives explained this to employees earlier this month, when they began reorganizing managers to have more direct reports. The company, whose head count has dropped by 25,000 in two years, also pointed to the influx of artificial intelligence as a reason it needed to move faster.

Ideally, companies would remove layers and spread decision-making throughout the organization so that those closest to customers or technology, for example, could generate ideas and make decisions, Ancona said.

"You're kind of flipping the organization," she said. "Rather than all the ideas coming from on high, you have entrepreneurial leaders who are lower down in the organization coming up with new ideas."

Bayer CEO Bill Anderson is leery of having to run everything up the chain. After taking over the German biotech company in 2023, he began implementing what he calls a "dynamic shared ownership" setup that has cut thousands of managers. Staffers come together in "mini networks" for 90-day stretches to work on projects.

"We hire highly educated, trained people, and then we put them in these environments with rules and procedures and eight layers of hierarchy," Anderson previously told BI. "Then we wonder why big companies are so lame most of the time."

Fewer managers, more reporting, more meetings?

When middle managers are cut and layers condensed, inevitably, more workers report to fewer managers. The logistics of that vary, and the success in terms of morale has a lot to do with the starting point.

Amazon started flattening last year. In September, CEO Andy Jassy ordered a 15% increase in the ratio of individual contributors to managers by March. BI reported that senior Amazon Web Services managers received a memo in January instructing them to restrict high-level hiring and increase their number of direct reports.

An Amazon spokesperson told BI at the time that the memo may have been intended for one team, but does not apply to the company at large. The Amazon spokesperson also referenced a September memo from Jassy on the importance of reducing management layers.

An AWS manager told BI this month that the flatter structure has since put more burden on employees on her team to report on what they're doing day-to-day, in addition to their actual work, since managers have less time to inspect individuals' work.

Plus, this manager said they are spending more time in meetings as they took on a more diverse group of direct reports. The Amazon spokesperson also emphasized that the individual employee's anecdote does not represent the company as a whole.

Yvonne Lee-Hawkins was assigned 21 direct reports when she worked for Amazon's human resources. She told BI that she had to quickly learn new skills to handle the load, like asynchronous work strategies, but her teams' performance suffered as her number of reports grew from 11 to 21 employees.

Weekly one-on-ones — the subject of much debate among tech titans — became impossible, and she had to cut them in half.

At Microsoft, a half-dozen employees who spoke to BI about the manager flattening trend generally regarded it as a positive step to eliminate inefficient and unnecessary levels of managers. Some managers have as few as one or two reports.

Microsoft ended up with many management layers, the people said, because it often tried to reward good engineers by promoting them to become managers. Often, those engineers-turned-managers still spent most of their time in the codebase and weren't very effective as managers.

Meanwhile, larger groups of direct reports often work better for senior employees, who need less one-on-one time and can do more things in a group setting.

A Microsoft spokesperson did not comment when asked about these factors.

Gary Hamel, a visiting professor at London Business School who lives in Silicon Valley, told BI that pushing managers to take on more direct reports can reduce micromanaging, a common bane of corporate existence.

When managers have a lot of people to oversee, it pushes them to hire people they trust, mentor rather than manage, and give up a "pretty big dose" of their authority.

"Those are all hugely positive things," he said, even if they require "a fairly dramatic change" in how managers see their role.

How many direct reports is too many?

Nvidia CEO Jensen Huang famously has 60 direct reports. Managers at Dell have been told they should have 15 to 20. An AWS document viewed by BI in January mandated no fewer than eight per manager, up from six. An Amazon spokesperson told BI there are no such requirements companywide.

Gallup research indicates that the quality of a manager matters more than the number of direct reports in terms of how well teams perform. That's because more engaged managers tend to lead to more engaged teams. And small teams — those with fewer than 10 people — show both the highest and lowest levels of engagement because managers can have an outsize effect, for better or worse.

That may explain why some companies seem to thrive with dozens of direct reports per manager and others fail.

The nature of the work matters, too. When work is more complex, it can be harder for managers to oversee too many people.

Managing dozens of people gets harder when "life intersects with work," Ravin Jesuthasan, the global leader for transformation services at the consulting firm Mercer, told BI.

When employees have an issue, they often need someone to talk to about it.

"As a manager, you are the first port of call," he said.

That's one reason, Jesuthasan said, that having something like 20 direct reports would likely be "really hard." For most managers, the couple of dozen direct reports that many tech companies are aiming for is probably the limit, he said.

Strong managers can powerfully boost a company's ability to develop talent and its bottom line. A 2023 analysis from McKinsey & Company, for example, found that organizations with "top-performing" managers led to significantly better total shareholder returns over five years compared with those entities that had only average or subpar managers.

While flattening schemes may be successful at reducing bulk in the middle and speeding up decision-making, they can hinder future growth if they're not well-managed.

Jane Edison Stevenson, global vice chair for board and CEO services at the organizational consulting firm Korn Ferry, told BI that removing layers from a management pyramid can help elevate those high performers. But flatter companies may fail to develop leaders who can pull together the disparate parts of an organization.

At some point, she said, "You've got to start to make a bet on the leaders that are going to have a chance to build muscle across, not just vertically."

• Meta employees let go for "low performance" describe gaslighting, golden handcuffs, and toxicity.
• From medical leave to an offer to be rehired, former employees share details of their termination.
• Former employees detail their experiences and what led up to their firings.

Months after Meta's high-profile culling of low performers, the stigma associated with the job cuts still stings.

CEO Mark Zuckerberg's push to "raise the bar on performance" saw about 5% of its workforce — roughly 3,600 employees — laid off in a sweeping round of cuts in February. The company said it targeted its lowest performers, but some former employees pushed back on that.

Some said they had received positive ratings months earlier. Others said they were on medical or parental leave, mid-transfer between teams, or hadn't received a formal rating yet. Some speculated that conflict with a manager was the deciding factor. As of late April, laid-off employees received severance payouts.

The eight accounts below offer a sense of how employees who were let go experienced the cuts and their concern about how it could affect their future. BI has verified the performance records of these ex-workers, who asked for anonymity because they fear the "low performer" label could hurt future job prospects.

A Meta spokesperson reiterated what the company previously shared with BI earlier this year about its low-performer cuts: "Let me be clear, these were performance-based terminations," the spokesperson said. "Prior ratings were not downgraded. Simply because someone had a history of meeting or exceeding expectations does not mean they continue to consistently meet the bar."

The spokesperson added that Meta employees are held accountable to a goal-based culture of high performance.

"Just part of Mark's mandate"

A former employee in Meta's human resources division had returned to work after being on parental leave for nearly five months when his division lost close to 10% of its staff. His mid-year review put him "At or Above Expectations" — the middle tier — but his year-end rating was "Meets Most," which made him eligible for termination.

"I tried calling my boss immediately, but in my heart, I knew it probably wasn't an error — Meta wouldn't make a mistake that significant." He said the manager was distressed but repeated: "I can't say anything. It's just part of Mark's mandate. It's a hard year."

He said he didn't know why he had been downgraded. "We can't even see the feedback our managers wrote for us," he said.

"Once the shock wore off, there was actually a mix of emotions — including some relief at being freed from that situation." But he does worry about carrying the low-performer label.

"It's one thing to be let go in a restructuring," he said. "It's another to have your professional reputation potentially damaged by being mischaracterized as an underperformer."

Cut, then recruited

A former senior machine learning engineer at Meta described the shock of being laid off, only for a Meta recruiter to invite her to reapply three days later and skip the interview process.

"Same email address. Same person. No acknowledgment of what had just happened. It felt surreal," she said.

She said she joined Meta in early 2024 and earned a mid-year review of "At or Above Expectations." Initially rated "Meets All Expectations" in the January 2025 year-end review, she was downgraded to "Meets Most Expectations" after a second round of director-level reviews. She said she suspects "directors were under pressure to hit a layoff quota."

Before the layoff, she was dissatisfied with her manager and sought a transfer to a top-tier team. It was approved, but she was told the move could lower her rating, so she waited — only to be pink-slipped. She said she's not taking the recruiter up on her offer.

Ultimately, she's confident she can land or create her next opportunity.

"I'll be fine. I can get another job or start something of my own."

Cut after a leave for burnout

A software engineer who joined Meta in May 2024 to work on cross-platform content sharing said a promising start unraveled due to internal politics over a new feature he prototyped. He said he took a four-month leave for burnout in August and returned in December to complete two more projects.

Two weeks before the layoffs, he said, his new manager told the team everyone was "safe." Then came the termination email — and a performance rating of "Meets Some Expectations," low on Meta's end-of-year rating scale.

"How could they evaluate my performance when I'd only worked 10 weeks in 2024?" he said, adding that an HR director had said he was "too new to evaluate."

Although he's now fielding interest from other companies and expects to receive 16 weeks of severance, he still feels the sting of his rating. "The harmful 'low performer' label still feels wrong," he said.

"Everything changed when I moved to Reality Labs"

A product manager joined Facebook as a contractor in 2018, converted to full-time in 2020, and consistently received year-end "Greatly Exceeds Expectations" ratings label.

In 2023, when she moved to Reality Labs, "everything changed," she said.

"My manager would call late at night to question my capabilities and create doubt, even though teammates gave me positive feedback," she said.

A review in early 2024 dropped her two tiers to "Meets All Expectations." Feedback included accusations that she was "blinded" by a "desire for promotion," she said. In August 2024, she took a 12-week disability leave "because I was physically ill from all the stress." She returned in November and was laid off in February.

She's being "selective" in her hunt for a new job. "The experience damaged my health, affected my family, and forced me onto medication for anxiety. I need to walk away from toxicity at the first sign, not wait until it affects my health and family again."

A templated response

This technical program manager was laid off after seven years of good performance ratings at Meta.

A reorganization placed her under a new manager a few months earlier — a shift she felt exposed her. "When managers have to meet quotas and choose someone to cut, people with less history on the team are at higher risk," she said.

"I contacted my manager in shock" after the termination notice, she said. "I just wanted five minutes to talk to ensure I wasn't going crazy." The manager responded with a templated email — the same company-approved language she had been required to use the day before when laying off her own direct report.

She still doesn't know exactly what led to their selection. "Was it because I had significant equity vested when the stock price was low?" she said. " Was it because I expressed disappointment about Meta rolling back DEI programs on internal forums? I'll never know — but there's clearly more to this story than simply cutting the lowest performers."

"The Hunger Games, but for high performers"

After 14 years at Google and three at Meta, this program manager said she was laid off despite a strong performance record: "At Meta, I'd even received 'Exceeds Expectations' and the rare 'Redefines Expectations.'" She said her manager told her she was doing "a great job" at the midyear review.

"I was candid with my manager — sometimes critically, but always constructively — about where I thought things could improve," she said. "I was transparent about how stress in my personal life was affecting me: my husband had been laid off, and we were caring for a special-needs child."

She said the manager said not to worry, but she had noticed a shift in tone at Meta — for example, CTO Andrew "Boz" Bozworth saying a worker who alleged mistreatment should just quit.

"The compassionate tone of the 2022 layoffs — the regret, the messaging about hard decisions — feels like a distant memory," she said. "This round was cutthroat. Silent. Cold. The Hunger Games, but for high performers."

"I wrote a final email to my skip-level manager," she added. "Not to get my job back. Not even to get a reply (I didn't). I just wanted to be heard. To say: I know I did good work. I know I didn't deserve this. And I hope someone — anyone — in leadership still has the humanity to care."

"This was about filling quotas"

This employee spent four years at Meta working on internal security systems, including tools to manage insider threats. He spent the weekend before the cuts making sure those would run smoothly on the day of the layoffs.

He had received consistently strong performance ratings, and said his managers agreed he was on track, but his termination letter assigned him a "Meets Most Expectations" rating. He said his managers didn't look at the self-review he submitted. "That told me everything. This wasn't about performance," he said. "This was about filling quotas."

"My manager said he 'stood by' the rating but wished he'd had more time and freedom to talk things through."

In his view, the company culture had grown increasingly cutthroat, with employees clinging to projects to boost their ratings, jockeying within the performance curve, and seeing dissent increasingly punished.

"Many people stay because the compensation is hard to walk away from, not because they believe in the mission. I was one of them. It was a place that offered scale and learning, but at a steep emotional cost," he said, adding that his healthy salary would have made him "an easy target" for cost-cutters.

"Companies can quietly rewrite the story of your career"

An engineer was laid off after five months of leave for a serious health crisis while in the middle of disability-related negotiations. He said his final performance rating — "Meets Most Expectations" — felt like "a scarlet letter."

"I wasn't underperforming," he said. Peer reviews before the leave had been entirely positive, and he saved "screenshots of it all for my lawyer."

His compensation exceeded $500,000, making him the highest-paid person on the team after his manager. He said he worked 60 to 75 hours before falling ill, but he sensed resentment after raising burnout concerns and stepping away from a major project.

"And after the layoff? Nothing. She didn't even say goodbye."

He said Meta's branding of laid-off workers as low performers "sets a dangerous precedent — companies can quietly rewrite the story of your career, and you don't get to fight back."

He's exploring legal options. "Meta is telling one story," he said. "And for many of us, it's not the truth."

Have a tip? Contact this reporter via email at [email protected] or Signal at +1-408-905-9124. Use a personal email address and a nonwork device; here's our guide to sharing information securely.

• Amazon updated its pay model to better reward top talent this year.
• The change reflects a wider Big Tech trend prioritizing top performance and tighter cost control.
• Amazon's pay structure has faced criticism for a lack of transparency and competitive rankings.

Amazon is overhauling its compensation model to more clearly reward sustained top performance while reducing payouts for some lower performers.

According to internal guidelines obtained by Business Insider, Amazon is revising its compensation structure within pay bands to better recognize and reward long-term high performers.

Employees who earn a "Top Tier" performance rating for four consecutive years will now receive 110% of their pay range, exceeding the usual cap. In contrast, first-time Top Tier recipients will get 70% of their pay band, down from 80% last year.

"This approach ensures a steadier compensation progression," the internal pay guideline said, referring to the company's total compensation target (TCT) metric.

The changes put Amazon in step with other Big Tech giants that are streamlining employee rewards and tightening overall costs. Google has trimmed bonuses and equity for underperformers, Microsoft has introduced stricter review policies, and Meta is actively downsizing its lowest-rated employees.

Amazon's spokesperson told BI that the updated model "better distinguishes" those with consistent excellence.

"As always, employees' contributions drive the outcome of their annual compensation review," Amazon's spokesperson said. "What's different this year is that our approach to compensation changes now better distinguishes between newer high performers and those who have consistently exceeded expectations for their role and level."

'Overall Value' ratings

Amazon's 12-month pay cycle for corporate employees typically starts in April. Most rank-and-file employees receive their pay updates in early April.

Although Amazon considers various factors when determining employee pay, individual performance ratings—internally known as "Overall Value"—are a key driver. Employees are ranked across five performance tiers: Top Tier (TT), Highly Valued 3 (HV3), Highly Valued 2 (HV2), Highly Valued 1 (HV1), and Least Effective (LE).

The internal guidelines say performance "directly impacts" compensation, and the OV ratings are "used to generate the compensation recommendation for each employee."

This year, more weight is being placed on an employee's rating history. For example:

• A jump from HV1 to HV2 now yields 10% of the pay range, down from 20%.
• But someone dropping from HV3 to HV2 still receives 20%, highlighting a greater emphasis on past performance.
• Those who move up to HV3 will get 40% of their pay range, not the 50% previously offered.
• An employee with two years of Top Tier status now gets 90% of their pay band, down from 100%.
• But 3 straight years of Top Tier can now reach 105% of the range. Those who get 4 consecutive years of Top Tiers can get 110% of the range. Previously, it was capped at 100%.
• This year, first-time Top Tier recipients will reach 70% of their pay band, compared to 80% in prior years.

Amazon's pay structure has long sparked internal frustration, partly due to its lack of transparency and the competitive nature of employee rankings, as BI previously reported. Managers are instructed not to share individual OV ratings with employees, leaving staff to infer their performance based on changes in their compensation.

This year's guidelines also said Amazon will continue to run the pilot program that allows employees to take 25% of their new stock awards in cash. Stock compensation has traditionally accounted for a large portion of Amazon's total pay, but it has become less appealing to employees seeking more immediate financial returns.

Despite the recalibration, Amazon says most employees who showed improvement still saw increases this year. The company offers "multiple channels" through which employees can raise concerns about pay, the spokesperson added.

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

• Large language models like DeepSeek's R1 are "overthinking," affecting their accuracy.
• Models are trained to question logic, but overthinking can lead to incorrect answers.
• A new open-source framework aims to fix this and could give a glimpse of where AI is headed.

Large language models — they're just like us. Or at least they're trained to respond like us. And now they're even displaying some of the more inconvenient traits that come along with reasoning capabilities and "overthinking."

Reasoning models like OpenAI's o1 or DeepSeek's R1, have been trained to question their logic and check their own answers. But if they do that for too long, the quality of the responses they generate starts to degrade.

"The longer it thinks, the more likely it is to get the answer wrong because it's getting stuck," Jared Quincy Davis, the founder and CEO of Foundry, told Business Insider. Relatable, no?

"It's like if a student is taking an exam and they're taking three hours on the first question. It's overthinking — it's stuck in a loop," Davis continued.

Davis, along with researchers from Nvidia, Google, IBM, MIT, Stanford, DataBricks, and more, launched an open-source framework Tuesday called Ember that could portend the next phase of large language models.

Overthinking and diminishing returns

The concept of "overthinking" might seem to contradict another big break in model improvement: inference-time scaling. Just a few months ago, models that took a little more time to come up with a more considered response were touted by AI luminaries like Jensen Huang as the future of model improvement.

Reasoning models and inference-time scaling are still huge steps, Davis said, but future developers will likely think about using them differently.

Davis and the Ember team are formalizing a structure around a concept that he and other AI researchers have been playing with for months.

Nine months ago — an eon in the machine learning world — Davis described his hack of asking, referred to as "calling," ChatGPT 4 the same question many times and taking the best of the responses.

Now, Ember's researchers are taking that method and supercharging it, envisioning compound systems wherein each question or task would call a patchwork of models, all for different amounts of thinking time, based on what's optimal for each model and each question.

"Our system is a framework for building these networks of networks where you want to, for example, compose many, many calls into some broader system that has its own properties. So this is like a new discipline that I think jumped from research to practice very quickly," Davis said.

In the future, the model will choose you

When humans overthink, therapists might tell us to break problems down into smaller pieces and address them one at a time. Ember starts with that theory, but the similarity ends pretty quickly.

Right now, when you log into Perplexity or ChatGPT, you choose your model with a dropdown menu or toggle switch. Davis thinks that won't be the case for much longer as AI companies seek better results with these more complex strategies of routing questions through different models with different numbers and lengths of calls.

"You can imagine, instead of being a million calls, it might be a trillion calls or quadrillion calls. You have to sort the calls," Davis said. "You have to choose models for each call. Should each call be GPT 4? Or should some calls be GPT 3? Should some calls be Anthropic or Gemini, and others call DeepSeek? What should the prompts be for each call?"

It's thinking in more dimensions than the binary question-and-answer we've known. And it's going to be particularly important as we move into the era of AI agents where models perform tasks without human intervention.

Davis likened these compound AI systems to chemical engineering.

"This is a new science," he said.

Have a tip or an insight to share? Contact Emma at [email protected] or use the secure messaging app Signal: 443-333-9088

• The Abilene City Council approved a tax abatement for Crusoe's data center campus in Texas.
• Business Insider obtained a copy of the agreement via a public records request.
• Crusoe, the data center developer, could get an 85% property tax break.

The developer of the first Stargate data center in Abilene, Texas, could get an 85% tax break on billions of dollars worth of property, according to a new tax abatement agreement obtained by Business Insider through a public records request.

To qualify for the tax break, AI startup Crusoe will need to spend a minimum of $2.4 billion of a $3.5 billion "targeted investment."

President Donald Trump announced Stargate, a joint venture between OpenAI, Oracle, and SoftBank, at a White House press conference in January. That month, Business Insider reported that Oracle is the tenant of two data center buildings under construction on Crusoe's data center campus in Abilene, which is widely assumed to be Stargate's first site. The estimated cost for those buildings, $1.1 billion, will count toward the targeted investment.

Since then, Crusoe has registered two more data center buildings with a state agency on the site. No tenant is listed for one of those buildings; Oracle is listed as the tenant for the other.

Crusoe has agreed to build six new data centers, each with a minimum square footage of 100,000, on Lancium's land in Abilene. These will join the two it is already constructing for Oracle.

Data centers create jobs—although some say it's not enough

A local news outlet first reported the abatement after the Abilene City Council approved it in February, naming the site's landowner Lancium as the tax break recipient.

The abatement requires Crusoe and Lancium to create 357 new full-time jobs with minimum salaries of $57,600. The does not specify what kind of jobs need to be created, though the hiring timeframe indicates that construction jobs will likely count toward the quota.

Critics of tax abatements for data centers have said that they don't create enough jobs to justify the tax breaks. Proponents like to point out that the sites can employ thousands of skilled laborers during the construction period.

Property tax abatements are common financial incentives used by local governments to encourage companies to operate in their cities and towns. Specific property tax breaks for data centers vary widely depending on locality. Loudoun County, Virginia, home to the world's largest concentration of data centers, does not abate property taxes for data centers. In New Albany, Ohio, a fast-growing data center, data centers have been granted 100% property tax abatements.

Stargate's expansion plans

Oracle executive chairman and CTO Larry Ellison has said the first Stargate data center is in Abilene, though scant details have emerged about the massive AI infrastructure initiative since Trump's announcement.

In February, OpenAI CFO Sarah Friar wrote in a LinkedIn post that the company was evaluating additional locations in Texas, Pennsylvania, Oregon, and Wisconsin for Stargate.

Crusoe announced it is developing a data center campus in Abilene, though it has not publicly confirmed that it is connected to Stargate. Public filings in Texas show that Oracle is Crusoe's tenant, Business Insider reported in January.

To finance the Abilene development, Crusoe entered into a $3.4 billion joint venture with private credit company Blue Owl. JP Morgan has lent $2.3 billion in construction financing to the project.

Lancium declined to comment. Crusoe, Oracle, and the Development Corporation of Abilene did not respond to requests for comment from BI.

Contact Ellen Thomas at [email protected] or on Signal at 929-524-6964.

• Amazon faces uncertainty as tariffs disrupt forecasting and supply chain operations.
• Amazon's suppliers and sellers say there has been very little guidance.
• Amazon's stock was among the hardest hit on Thursday following Trump's tariff announcement.

The tariff mayhem is throwing Amazon into uncertain territory.

Forecasting, for example, has become nearly impossible for some teams. One of Amazon's largest supply chain units recently warned about the challenges of making its second-quarter projections due to tariffs, according to an internal email obtained by Business Insider.

The "volatility and uncertainty" from the new round of tariffs were simply too high to derive any meaningful numbers, the email said.

Amazon employees, alongside suppliers and sellers, are scrambling for answers as President Donald Trump's whipsaw trade policy roils the country's largest e-commerce retailer. On Thursday, Amazon was among the hardest-hit stocks when roughly $2.5 trillion was wiped out of the S&P500 Index over Trump's aggressive tariff plan.

Amazon has given little guidance or financial flexibility so far, according to multiple employees, suppliers, and sellers, who mostly spoke on the condition of anonymity because they were not authorized to talk to the press. Tension is intensifying as concerns of a prolonged trade war and potential recession loom.

Amazon's spokesperson didn't respond to a request for comment.

'Large risk'

The same email from Amazon's supply chain team said that the near-term impact of tariffs will ultimately be captured and reflected in a later forecast. But the exposure to tariffs and a global trade war is "a large risk" that can set back Amazon's retail business going forward, it added.

Some Amazon employees have been in direct contact with its suppliers, commonly known as first-party vendors. These companies sell their products wholesale to Amazon, which then resells them to shoppers.

These vendors said Amazon isn't willing to pay more for their products, even if the tariffs would increase the suppliers' costs. According to a March email seen by BI, an Amazon employee encourages vendors to seek further cost savings from their own manufacturers or through government subsidies.

"We understand the challenges posed by the current economic and trade environment," the email said. "However, we believe there are alternatives to direct cost increases that haven't been fully explored."

Some vendors told BI that Amazon also seeks "margin agreements" that guarantee the same margin after a vendor increases its prices. That way, Amazon would maintain its profit margins from its suppliers, even if it buys the products at a higher price.

In some cases, Amazon is pausing shipment orders from vendors to monitor the market. One shipping company recently told a vendor that "as per Amazon's request," it was holding the pick-up schedule of inventory to "mitigate the impact" of tariffs, according to an email seen by BI.

Alan Adams, president of Navazon, a vendor software company, told BI that tariff discussions with Amazon employees have been ongoing for months. He said both Amazon and the suppliers are pursuing ways to adapt to the new market conditions, but the constant policy changes make finding a long-term solution difficult.

"We are all in a wait-and-see mode with a tremendous amount of uncertainty across different categories," Adams said.

Raising prices

Trump imposed sweeping tariff increases on most countries this week. The changes are expected to increase prices across a variety of goods.

Truist Securities' Youssef Squali said the tariffs will likely have an adverse effect on e-commerce companies, including Amazon. Import costs will likely eat into their margins, though it will still take time to fully measure their impact on each individual company, he wrote in a note Friday. Amazon's stock is down roughly 10% from Wednesday.

Third-party merchants who sell on Amazon told BI they will likely have to raise their prices due to the tariffs.

Charles Chakkalo, founder of JoeyzShopping, who sells home and kitchen items, said he anticipates over 50% tariffs on his products. To counter, he will have to raise prices, while leveraging his unit volume to lower manufacturing costs.

Oscar Babarin, managing director of marketing agency Hawke Media, said a number of his clients are feeling the impact deeply. Some of them are scaling back their business, while others are more aggressively pursuing market share, he said.

However, some sellers, are excited about the elimination of the de minimis exemption that allowed tax-free shipments of Chinese imports valued at less than $800, according to Oliver Scutt, board member of Merchant AI. Those sellers expect less competition from Temu and Shein following the change, he said.

Still, most sellers and vendors said they feel helpless against the complexity of trade policies. On Thursday, as the market plunged, one supplier emailed an Amazon manager to ask for additional guidance, only to receive very little support.

"Rest assured, we are looking into it," the Amazon manager said.

Do you work at Amazon? Got a tip? Contact this reporter via email at [email protected] or Signal, Telegram, or WhatsApp at 650-942-3061. Use a personal email address and a nonwork device; here's our guide to sharing information securely.

• Amazon projects Rufus to indirectly contribute over $700 million in operating profits this year.
• Rufus' product recommendations increase consumer spending on Amazon's marketplace.
• Amazon plans to expand Rufus globally and enhance its AI model for better service.

Amazon has rosy projections for its Rufus AI shopping assistant.

According to an internal planning document obtained by Business Insider, Amazon expects Rufus to indirectly contribute over $700 million in operating profits this year.

The outlook is part of a metric called "downstream impact," an internal financial figure Amazon uses to measure a product or service's potential to generate additional consumer spending across Amazon's vast offerings.

For example, Rufus' product recommendations can lead to more purchases on Amazon's marketplace — and an increase in its DSI. Rufus, a free service that answers everything from product details to special promotions, generates no direct revenue.

Based on this metric, Rufus lost an estimated $285 million in 2024. By 2027, however, it is expected to reach $1.2 billion in DSI profit contributions, the document showed. The estimates, after operating costs and server payments, include the income from ads placed within Rufus' responses to user inquiries.

The bullish forecast may explain why Amazon is aggressively expanding Rufus and its other AI search offerings. Amazon's leadership previously told employees that such AI projects are "absolute" top priorities.

Rufus, which launched in February 2024, has received mixed reviews from early users, but Amazon continues to invest in this space. Just last week, Amazon unveiled another AI-powered shopping discovery app called Interests. Last year, the company disclosed that Rufus answered "tens of millions" of customer questions in its first six months of operation.

It also shows how Amazon justifies its investment in one of its flagship AI products, even as DSI has received more scrutiny lately.

Amazon's spokesperson declined to comment.

'Positive lift on downstream spending'

To be sure, Rufus' DSI forecast is small for Amazon, given the company's $68.6 billion in total operating income last year. But early results indicate Rufus has huge potential with more room to grow.

The document said initial results showed customers using Rufus had a "positive lift on downstream spending," and the impact only increased with more engagement.

Amazon has significant expansion plans for Rufus.

This year, Amazon expects $711.7 billion worth of products on its site to be eligible for Rufus' many features, like product recommendations and comparisons, up from $164 billion in 2024, according to the document. That number is expected to reach $849.8 billion in 2027.

Amazon also wants to roll out Rufus to at least 13 international marketplaces this year, the document said. Currently, Rufus is available in the US, the UK, and India, and a handful of European countries.

The document said Amazon plans to fivefold the size of the AI model powering Rufus, internally called Shopping LLM, to improve the quality of its answers. The increased model parameters to Shopping LLM were also factored into DSI, but it is unclear when this will be implemented. With the planned improvement, Amazon hopes to launch a new service that leverages Rufus to provide a wider selection and "seamlessly purchase on behalf of the customer," it added.

Rufus is already becoming more prevalent across Amazon. In recent weeks, some Amazon sellers who used the Rufus app discovered an update that auto-pops up Rufus when they use trigger phrases like "how to" or "what is" in the search bar.

For some products, Rufus automatically launched with its own response before search results showed up. Amazon also started testing ad placements within Rufus' responses, Adweek reported in September.

'Personalize more'

Amazon isn't the only retailer with a homegrown AI shopping assistant. Others, like Walmart, Target, and Instacart, have launched their own AI-powered search apps in recent months.

These apps may still be in their infancy. A recent survey by e-commerce analytics firm Profitero showed that just 10% of US customers used a retailer's AI chat assistant when searching for products. By comparison, 37% of shoppers used the search bar on the retailer's websites or apps, and 29% used promotion and deal pages.

For Rufus, the other challenge is negative reviews.

Andrew Hamada, a former Amazon employee who now runs seller agency Reason Automation, wrote on Linkedin last month that Rufus misses basic catalog information and often gives inaccurate answers.

"In our experience, Rufus rarely works," he wrote.

There's also the question of DSI's true value.

Amazon's DSI is not the most scientific metric, and several Amazon employees who spoke on the condition of anonymity because they were not authorized to speak to the press, have raised questions about the fuzzy math behind the numbers.

For instance, it was never clear whether Prime Video or Amazon's shipping team should claim more DSI for the Prime membership revenue, which offers both services as a perk, one of the people said. Recently, Amazon has been deemphasizing Alexa's DSI metric, the Wall Street Journal previously reported.

Still, Amazon's retail CEO Doug Herrington remains a strong supporter of Rufus. During an internal all-hands meeting earlier this year, Herrington said AI-driven "personalization" is a major focus area for Amazon, mentioning Rufus and other search features as examples.

"AI is really allowing us to personalize more and more of the shopping experience on Amazon," Herrington said.

Do you work at Amazon? Got a tip? Contact this reporter via email at [email protected] or Signal, Telegram, or WhatsApp at 650-942-3061. Use a personal email address and a nonwork device; here's our guide to sharing information securely.

• Adobe is moving away from its diversity hiring targets.
• It had set goals in 2020 to increase minority representation in leadership positions.
• Adobe joins firms like Google and Meta in reducing diversity, equity, and inclusion efforts.

Another tech giant — Adobe — is scaling back its diversity hiring goals.

At an internal all-hands meeting on Monday, Adobe's chief people officer, Gloria Chen, said Adobe would no longer have diversity, equity, and inclusion hiring targets, internally dubbed "Aspirational Goals," according to a recording of the meeting obtained by Business Insider.

These goals were set in 2020 to "increase global diversity and inclusion" at leadership levels, Adobe's company blog said. The goals included increasing female leadership representation to 30% globally, doubling underrepresented minorities in leadership roles, and doubling Black representation as a percentage of US employees by 2025.

"We will discontinue the practice of setting aspirational representation goals while continuing our focus on fair and consistent hiring practices," Chen said, adding Adobe had never actually hired based on such quotas.

Adobe is the latest in a long list of companies, including Google, Meta, McDonald's, and Deloitte, to reduce its DEI initiatives. The change follows executive orders issued by President Donald Trump in January aimed at ending DEI programs within the federal government and its contractors.

In an email to BI, Adobe's spokesperson said, "We've always believed in and remain committed to Adobe for All, which is our belief in creating a company culture where all employees are empowered to make an impact. While some programs and policies are changing, our values are not."

'Need more transparency'

During Monday's meeting, Chen said executive orders could be "complex" to interpret and Adobe was "evaluating" many internal programs, activities, and practices to ensure the company complied with them.

Chen added that Adobe didn't believe the DEI pullback applied to countries outside the US for now. As of November, Adobe had a little over 30,000 employees worldwide, 50% of which were in the US, according to a company filing.

Adobe also removed all diversity mentions in its latest proxy statement, a change first spotted by Michelle Leder, the founder of Footnoted, a site that analyzes Securities and Exchange Commission filings. Adobe mentioned diversity 22 times in the previous year's proxy. This year's proxy also deleted a chart on director diversity.

Some Adobe employees shared their frustration in an internal Slack channel, according to screenshots seen by BI. One employee said Adobe's unique culture of embracing different perspectives, called Adobe for All, now seems "lost." Another said it was "heartbreaking" and asked for additional guidance from the leadership team.

"I think we all need more transparency around this issue," one of the people said.

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

• Shopify has acquired Vantage Discovery, a startup building AI search for retailers.
• Vantage Discovery was founded by two former Pinterest engineering leaders.
• Shopify has been acquiring startups in an effort to bring on more AI talent.

Shopify has acquired Vantage Discovery, a startup that builds AI-powered search functions for retailers.

Cofounders Lance Riedel and Nigel Daley both previously worked in engineering at Pinterest. Riedel built out Pinterest Shopping, while Daley worked in engineering infrastructure.

They started Vantage Discovery in early 2023 to bring a "Pinterest-like capability to any retailer" using generative AI, Daley said in an interview with Business Insider in February.

Vantage Discovery uses LLMs to enhance retailers' search functions, allowing shoppers to see more personalized, relevant results when searching through a store's product catalog.

"In the past, machine learning and AI had been held by some of the much bigger companies," he said. "Now with Gen AI and technology like Vantage Discovery, we can bring that same power to any retailer, from the smallest mom-and-pop shop to massive enterprises."

Riedel said in a LinkedIn post that Vantage Discovery would integrate its "revolutionary technology with Shopify's commerce platform." Daley confirmed the news in an email but declined to comment further on the deal.

"Vantage Discovery's search platform and mission-aligned team will play a key role in supercharging our work for both merchants and buyers," a Shopify spokesperson said to BI.

Shopify has been on an acquisition spree. It acquihired six startups in 2024 with the goal of bringing on AI talent.

"These have been very tactical, thoughtful AI hires and we want to continually be thoughtful, proactive, and judicious on thinking about the cash," CFO Jeff Hoffmeister said during the company's earnings call in February.

Have a tip? Contact this reporter via email at [email protected] or Signal at mlstone.04. Use a personal email address and a nonwork device; here's our guide to sharing information securely.

Affected staff say more than 100 employees working to protect U.S. government networks were ‘axed’ with no prior warning

© 2024 TechCrunch. All rights reserved. For personal use only.

Cheaper, lighter, and denser: The trifecta defines an ideal battery. No one has devised a perfect cell quite yet, but one stealthy startup thinks it has found a new material that solves at least two of those challenges. Daqus Energy has been quietly operating for the past four months, refining a compound known as TAQ […]

© 2024 TechCrunch. All rights reserved. For personal use only.

• Meta maintains internal blocklists of employees who are ineligible for being rehired.
• BI spoke with several former employees who said they were surprised to learn they were on these secret lists.
• Some former Meta managers said it could be nearly impossible to get employees off the list.

A senior engineer caught up in Meta's 2022 mass layoff of 10,000 employees thought finding another role at the tech giant would be straightforward.

They had worked at the company for more than four years, consistently receiving "Exceeded Expectations," one of the stronger ratings on Meta's performance scale. The year prior, the engineer had earned a coveted promotion to a senior technical role. Their work had been praised by their skip-level manager, who told them in an email that their contributions were "crucial" to the company's success. They were told that their layoff was simply a business decision and that multiple hiring managers now wanted them back on their teams.

After applying to nearly 20 positions in the year after they were let go, the engineer noticed a pattern: hiring managers would express interest in bringing them back and set up a screening call with a recruiter, the first stage in Meta's multi-step hiring process. Then, the recruiters would ghost them.

Frustrated, the engineer asked a hiring manager what was happening. They were told Meta's recruiting team had forbidden the manager from contacting the engineer because they were deemed "ineligible for rehire."

"That was the first time I had a real indication that I was on some kind of list," the engineer, who is unnamed because they are actively trying to get rehired at Meta, told Business Insider.

Meta maintains internal lists that bar some former employees from rejoining the company, according to five former employees including two managers that BI spoke to. BI learned that even those who had good performance records may be prevented from being rehired. While such lists aren't illegal, they are unusual, employment and human resource experts said.

The former employees and managers confirmed Meta uses multiple systems to track rehire ineligibility, including a "non-regrettable attrition" designation and a "do not rehire" flag, though the exact mechanisms and number of affected employees remain unclear. While BI couldn't view a single, comprehensive list, multiple internal communications we reviewed showed that managers encounter systemic barriers when attempting to rehire certain former employees.

For the last few years, tech companies have been aggressively pursuing efficiency after years of explosive growth. Following widespread layoffs across Silicon Valley since 2022, tech giants are now implementing stricter performance management systems while simultaneously competing for top talent in AI and other high-growth areas. For former employees seeking to return these hidden rehiring restrictions create an invisible barrier in an already competitive job market.

Interviews with five former employees across different divisions, along with internal emails and messages viewed by BI, revealed multiple instances of workers who were laid off and discovered they were barred from rejoining Meta after applying for new positions at the company despite good performance records in their previous stints. These workers often learned about their status from third-party contractors rather than Meta itself.

BI has kept names and other identifiable details private to protect sources' confidentiality but has verified the identities and employment histories of these individuals.

Employees who violate workplace policies, such as stealing confidential data, receive permanent bans from future employment at Meta. The company also maintains lists that sometimes include underperforming employees. However, two former managers who spoke with BI say the bans extend beyond these clear-cut cases: managers often have broad discretion to add names to these lists, without documented performance issues.

"If a manager didn't like you, it wasn't hard to put someone on a list," one former manager said.

In a statement to BI, a Meta spokesperson said: "There are clear criteria for when someone is marked ineligible for rehire that are applied to all departing employees and there are checks and balances in the process so that a single manager cannot unilaterally tag someone ineligible without support."

Meta says its decision to bar an ex-employee from rehiring is based on a multitude of factors.

"We determine, at the time of separation, the reason for the employee's departure — policy violation, performance termination, voluntary resignation etc. — and that, along with the last rating prior to separation and any other recent performance signals, determines whether an employee is eligible for rehire or not," the Meta spokesperson told BI.

One former manager, however, said that they and other managers at the company were able to put people on lists by "just filling out a form" and "putting in any real issue."

It would "take minutes to get someone marked as 'non-regrettable'," they told BI. "The manager would get an email asking if it was non-regrettable attrition or not."

Blockedlisted without an explanation

After being cut during Meta's 2022 layoffs, a former hardware engineer with a record of being rated "Exceeded Expectations" was approached by a staffing agency about a contract position with the same team they built and managed at Meta.

In emails viewed by BI, the staffing agency was initially enthusiastic, noting the ex-employee had the qualifications for the role. But after they applied, the responses grew increasingly vague.

First, the agency said the role wasn't aligned with their experience. When the applicant pressed for more information, the agency revealed that Meta's HR team had marked them as "ineligible to be hired," the emails viewed by BI showed. The agency declined to provide additional details, saying the information was "kept confidential by Meta HR."

The staffing agency did not respond to a request for comment from BI.

Last month, Meta cut nearly 4,000 workers in a move that CEO Mark Zuckerberg characterized as performance-based trimming.

Another former employee, who was also laid off in 2022, was approached months later by a different staffing agency for a contract position at Meta that was similar to their previous role at the company. When they applied, the agency told them that Meta said that they were not eligible for rehire without providing a reason, according to an email between the staffing agency and this former employee viewed by BI.

"I got really frustrated because they didn't tell me the reason," the former employee said. "They gave this information to a third party, the contract company, but not to me." Their former managers did not respond when they reached out for clarification. The second staffing agency did not respond to requests for comment from BI.

Meta's practices appear to be unusual even within the competitive tech industry landscape. Barring former employees with no major policy violations from being rehired without notifying them explicitly isn't standard practice at other major tech companies, according to industry veterans familiar with hiring systems.

"It's incredibly uncommon. This is very, very rare," Laszlo Bock, Google's head of people operations from 2006 to 2016, told BI.

"I've actually never heard of a company having a 'do not rehire' designation for former employees, because if an employee was a decent or good performer, you'd much rather hire somebody who actually knows your company and culture than somebody else." Bock added that he had never seen the practice formalized at any major corporation thus far.

A mechanism outside normal recruiting tools

In the summer of 2024, Meta connected the senior engineer mentioned earlier in this story with one of its lawyers from the company's employee relations team after the engineer contacted HR.

After reviewing their employment history, the lawyer confirmed that there were no HR violations that might warrant a hiring restriction, the engineer told BI. When the engineer pressed for an explanation about the recruiting block, Meta's lawyer told them they couldn't comment.

In a follow-up email to Meta's HR department viewed by BI, the engineer asked Meta to confirm whether they were on a recruiting blocklist and, if so, to provide the reason. The company said it couldn't provide a definitive answer, and that they could try applying to open roles on Meta's careers website.

Over the next several months, the senior engineer contacted multiple Meta directors and managers who appeared interested in rehiring them, according to messages viewed by BI.

A division director told them that they had witnessed similar cases where current Meta employees, who had been given a month to find new internal roles after their programs had been canceled, discovered that they were blocked from transferring because they happened to be on a list.

Another director, who was actively trying to rehire the senior engineer, expressed frustration about the block and said that it was preventing them from rehiring multiple people they hoped to recruit.

One hiring manager who wasn't allowed to hire the senior engineer told them in a private message that Meta asked them not to discuss specific reasons for not moving forward. When the senior engineer asked a different hiring manager about getting off the list, they were told that even a sign-off from a vice president wouldn't be enough.

In an email viewed by BI, a hiring manager expressed frustration that they couldn't figure out why someone they wanted to rehire had ended up on a block list. It seemed, they said, that they were running into a mechanism outside normal recruiting tools. They wrote that they had not seen anyone successfully get off a list and be able to interview at the company again.

Employment attorney Ashley Herd told BI that companies maintaining such lists isn't illegal, but it could pose legal risks if it disproportionately affects protected groups.

"For it to be illegal under nationwide federal law, it would have to be discriminatory based on some sort of protected characteristic," Herd said. She added that it's "really a terrible practice because you're missing out on talent, especially people that know your business."

Mitchell Epner, a partner at Kudman Trachten Aloe and Posner LLP who practices employment law, told BI that it's common for severance agreements to include provisions barring former employees from reapplying. None of the former Meta employees BI spoke to said their agreements contained such language.

"The rights that an individual has vis a vis an employer are extraordinarily limited," Epner said. "Outside of discrimination against protected classes, employers can do virtually anything that they want... for any reason at all or no reason at all."

Meta's block lists aren't foolproof. In an incident that resulted in a lawsuit filed last year, Meta accidentally rehired a former employee as a contractor despite him being on a "Do Not Hire" list after he was accused of stalking and harassing a coworker for over a year.

Despite their experience, the senior engineer told BI that they would still return to Meta if given the chance.

"It's the worst company I've ever worked for," they said. "But they also pay the best. If I could get in there for a couple more years and make bank, I would do it."

Have a tip? Contact this reporter via email at [email protected] or Signal at +1-408-905-9124. Use a personal email address and a nonwork device; here's our guide to sharing information securely.

• CentralReach has tapped investment bank William Blair to weigh a potential sale, BI has learned.
• The company makes electronic health records software for autism care.
• One person with knowledge of the efforts said CentralReach could seek a valuation of over $1 billion.

Autism care software maker CentralReach has been exploring a sale, Business Insider has learned.

CentralReach has tapped the investment bank William Blair to pursue a potential sale, according to people with knowledge of the efforts.

Founded in 2012, CentralReach sells electronic health records software to providers caring for autism and related intellectual and developmental disorders. The company also builds AI-powered tools within its platform to help providers automate tasks like reviewing clinical notes.

CentralReach and William Blair didn't respond to requests for comment for this story.

Fort Lauderdale, Florida-based CentralReach has been publicly quiet about its financials to date. It's backed by venture and private equity firm Insight Partners, which invested an undisclosed sum into CentralReach in 2018.

Two people with knowledge of CentralReach's efforts said the company has soared as a market-leading maker of software for autism care. One of those people said CentralReach is doing about $75 million in earnings before deductions like taxes and interest.

That person said they expect CentralReach to seek a valuation north of $1 billion, based on current market standards.

A number of healthcare companies have been looking to sell themselves after a slow year for company combinations in 2024.

Investors and bankers told Business Insider in February that healthcare companies that are profitable and can demonstrate stable revenue will be far more attractive to buyers, especially private equity buyers.

Medical practice tech maker ModMed announced Monday that it sold a majority stake to PE firm Clearlake Capital from Warburg Pincus, which first invested in ModMed in 2017. The company didn't disclose the terms of the acquisition, but the Financial Times reported the day before that the deal valued ModMed at $5.3 billion.

CentralReach has also made multiple acquisitions in the past few years, most recently buying Behavior Science Technology, which built a platform for measuring and tracking the efficacy of autism therapy, in September. In its 13-year history, CentralReach has made 14 acquisitions.

CentralReach said in a September release that its tech is used by more than 175,000 healthcare professionals globally.

• BI obtained internal documents showing some of the teams and roles that were hardest hit by Meta's cuts.
• The data shows teams under Tom Alison, Facebook's head, had 335 job cuts.
• Meta began cutting thousands of jobs last week, which has affected roughly 5% of its workforce.

Newly obtained internal documents show some of the teams that were hit the hardest in Meta's recent job cuts.

The tech giant began cutting thousands of jobs last week, which has affected roughly 5% of its workforce.

Business Insider obtained 30 records sent to impacted workers along with their separation agreements, which offered a picture of how Meta is reshaping its workforce.

The materials include the number of employees affected, broken down by Meta's business groups, their teams, and their job titles. They also list the vice president or senior manager to whom each group reports. Additional records may exist, but BI has not been able to verify any.

Meta said in the documents that it provided this information under federal law to help affected employees decide whether to sign their separation agreements.

The records list 3,115 employees hit by the cuts, a figure that appears to represent a significant majority of the 5% of employees — equivalent to about 3,600 workers — who were laid off.

Among the 30 group documents reviewed by BI, the teams reporting to Tom Alison, the head of Facebook, saw the highest impact, with 335 employees affected.

The next-largest cuts came from Meta's Horizon team, which works on its virtual reality platform. A total of 244 employees were affected in the group, reporting to Vishal Shah, the vice president of Metaverse under the Reality Labs division — the unit overseeing Meta's VR and augmented reality efforts.

Other heavily impacted teams included those under Carmine Arabia, the vice president of devices at Reality Labs, where 195 roles in business analytics, engineering, technical sourcing, and technical program management were eliminated.

The group previously led by Lori Goler, Meta's former head of people, saw 189 administrative roles cut.

Additionally, 186 roles tied to Meta's data center strategy, design, engineering, and construction teams were eliminated. The group reported to Rachel Peterson, Meta's vice president for data center strategy.

Meanwhile, the organization under Peng Fan, its vice president of engineering for monetization, saw 180 cuts, primarily software engineers.

Fan recently said in an internal memo that Meta planned to expedite the hiring process for machine learning engineers through February and March. The move came as Meta CEO Mark Zuckerberg said the company would backfill roles cut in its effort to target "low performers."

The group reporting to Alex Himel, the vice president of augmented reality, had 141 employees cut; they mostly worked on wearables.

Meta declined to comment.

Two of the seven groups most affected by the cuts are tied to Meta's Reality Labs unit, which was recently reorganized and positioned as a core part of the business.

In a November memo obtained by BI, Meta's chief technology officer, Andrew Bosworth, called 2025 a make-or-break year for the company's metaverse ambitions.

Bosworth said this year was critical for proving whether the metaverse would be a visionary feat or a "legendary misadventure."

Editor's note: We've updated this story to further clarify that the data referenced in the 30 documents is a portion of all the data that may be available.

Are you a Meta employee? Got insight to share? Contact the reporter Jyoti Mann via email at [email protected] or via Signal at jyotimann.11. Reach out from a nonwork device.

• xAI plans to operate a large data center to support X in Atlanta.
• The data center will have about 12,000 GPUs and $700 million worth of equipment.
• xAI set up a massive data center in Memphis last year.

xAI has been quietly setting up a data center in Atlanta, expanding its footprint beyond the massive data center in Memphis.

Elon Musk's AI startup plans to operate a large data center with X. The two companies are combining hardware to total roughly 12,000 graphics processing units, the Nvidia-designed chips used for most AI computation, according to a Business Insider tally of the equipment listed in the companies' agreement with Develop Fulton, one of Atlanta's economic development agencies.

In December, X and xAI signed similar agreements with Develop Fulton. Under the agreements, Develop Fulton orchestrated a municipal bond process to finance the $700 million in chips, cables, and other equipment going into the single facility. Fortune initially reported on the Atlanta data center. The size and scale of the data center have not been previously reported.

Representatives for X and xAI did not respond to a request for comment.

Inside the Fulton County data center

The Atlanta data center has sizable computing power, according to a data center solutions architect, and expert in AI hardware, who asked to remain nameless to comment on the documents. It's comparable to a data center a hyperscaler like Google or Amazon might set up.

X representatives described it as an "exascale" data center capable of computing "trillion-parameter AI." But the Georgia facility pales in comparison to the reported capacity of xAI's Memphis project, nicknamed Colossus, which Musk has called the largest data center in the world.

The Georgia facility will house an estimated 12,448 Nvidia GPUs. The vast majority of these are Hopper generation H100 GPUs, which cost between $277,000 and $500,000 for each rack of eight chips, according to the documents.

Roughly 3% of the chips are Nvidia's less-powerful A100 GPUs, which cost $147,000 for an equivalent configuration of eight chips. X is contributing all of the A100s, along with 11,000 H100s.

Neither of these chip designs requires liquid cooling, which has been a point of tension for Musk's companies in Memphis. When operating at full capacity, Colossus is expected to become among the largest consumers of water in the city.

In addition to H100 chips, xAI is contributing Mellanox switches and optics — high-bandwidth networking equipment that can facilitate chips working together faster, also purchased from Nvidia. Documents submitted to Develop Fulton by X indicate that the facility will be used to "develop and train artificial intelligence products."

Of the $700 million in accelerated computing hardware going into the facility, $442 million is allocated to X, and $258 million is allocated to xAI. The two companies will receive tax abatements that are estimated to be worth about $10 million over the course of ten years, to be split in proportion with their hardware investments, according to a representative at Develop Fulton. Kwanza Hall, chairman of the board at Develop Fulton, said the organization estimates the project will have an overall economic impact of more than $241 million.

The data center architect estimated the Atlanta facility would require 20 megawatts of total power, which it could realistically get from the power grid. xAI has requested 150 megawatts from the Tennessee Valley Authority for the Memphis facility, a representative for Memphis Light, Gas and Water said during a city council meeting in January.

X and xAI's partnership

The Atlanta facility is an example of Musk ostensibly pooling his resources to benefit both X and xAI. According to the records, X contributed 90% of the hardware for the data center, and xAI 10%.

xAI launched its chatbot, Grok, on X's platform in November 2023, but has since spun it off as an additional stand-alone app. On Monday, xAI released the latest version of the chatbot, and Musk said it has "more than 10 times" the compute power of its predecessor.

The equipment will be used to train large language models and semantic search products for the X platform, according to the documents. X has about 16 employees in the area, based on a review of LinkedIn profiles. xAI has one worker stationed at the Georgia facility and two additional employees listed as "X Corp Partner," the company's internal org chart shows. The deal with Develop Fulton states that 24 jobs will be maintained at the facility and none will be added.

Musk is trying to position xAI as a major competitor to Big Tech giants like OpenAI and Google, even drawing some talent from Tesla. The company built its Memphis data center in just 122 days, according to Nvidia — in record time for a data center of its size. xAI has also brought in hundreds of data annotators to train its chatbot over the past year with an eye to hiring thousands in the months to come, BI previously reported.

In February, Musk and a group of investors submitted a $97.4 billion bid to buy the nonprofit that controls OpenAI, but the billionaire later said he would withdraw the bid if OpenAI remains a non-profit entity. In response to Musk's initial offer, OpenAI CEO Sam Altman told reporters that the company is "not for sale."

"He obviously is a competitor," Altman said of the bid. "He's working hard, and he's raised a lot of money for xAI and they're trying to compete with us from a technological perspective, from getting the product into the market, and I wish he would just compete by building a better product."

Do you work for xAI or one of Musk's companies? Reach out to the reporter via a nonwork email and device at [email protected] or through the encrypted messaging platform Signal at 248-894-6012.

Have a tip or an insight to share? Contact Emma at [email protected] or use the secure messaging app Signal: 443-333-9088

• Meta is betting on humanoid robots as it creates a product group.
• In an internal memo, Meta's CTO announced the hiring of former CEOs from Cruise and The RealReal.
• Marc Whitten joins as VP of robotics. John Koryl is VP of retail, aiming to boost smart glasses sales.

Meta's next big bet is humanoid robots.

The company has hired former CEOs from Cruise and The RealReal to lead its robotics ambitions and boost sales of smart glasses, according to an internal memo seen by Business Insider.

Meta's chief technology officer, Andrew Bosworth, announced the hiring of former Cruise CEO Marc Whitten and John Koryl, a former CEO of the luxury-goods reseller company The RealReal, in the memo shared with employees on Friday.

Bosworth said Whitten would be the vice president of robotics and lead the new product group, while Koryl is joining as the vice president of retail, where he's set to focus on "supercharging" retail initiatives.

Explaining Meta's rationale for entering the consumer humanoid robotics space, Bosworth said it would leverage the company's open-source large language model, Llama, and help Meta create more value with its mixed- and augmented-reality programs.

"The core technologies we've already invested in and built across Reality Labs and AI are complementary to developing the advancements needed for robotics, like hand tracking, material simulation in Gemini, Aria, low-bandwidth, always-on sensors, and more," he wrote.

Meta didn't immediately respond to a request for comment.

In a memo from last month, Bosworth announced a series of reorganizations, which showed Meta is making Reality Labs a bigger part of its core business.

Brent Harris, the vice president of Meta's wearables business group; Brett Vogel, the company's vice president of product marketing and the metaverse; and the product marketing team now work directly under the team of Meta's chief marketing officer, Alex Shultz.

Bosworth also told employees in November that 2025 is "most critical" for the company's metaverse bets and that it could be a "legendary misadventure" if the company didn't boost sales.

Read Bosworth's full memo:

Welcome Marc Whitten, VP Robotics, and John Koryl, VP Retail

I'm excited to announce two exceptional leaders joining Reality Labs, reporting to me:

We are establishing a new Robotics product group and Marc Whitten will be joining as our new VP Robotics to lead this group. Marc is a seasoned product leader with a proven track record of driving innovation, having led teams at Unity, Amazon, and Sonos, as well as being one of the founding team members on the Xbox team. He most recently served as CEO of Cruise.

Our new Robotics product group will focus on research and development in the space of consumer humanoid robots with a goal of maximizing Llama's platform capabilities. The core technologies we've already invested in and built across Reality Labs and AI are complementary to developing the advancements needed for robotics, like hand tracking, material simulation in Gemini, Aria, low-bandwidth, always-on sensors, and more. We believe that expanding our portfolio to invest in this field will only accrue value to Meta AI and our mixed and augmented reality programs. This work is moving quickly and we'll share more updates soon.

John Koryl, VP Retail, joins us to focus on supercharging and growing our first-party retail initiatives. It's no secret that products like Quest, Ray-Ban Meta, and AI wearables require a different kind of customer experience beyond the standard retail offerings — our recent Meta Lab concept in LA is proof of that. We see an opportunity to build more direct expertise in the space and then share what we have learned with all of our retail partners who will continue to drive the bulk of our sales volume. Koryl has deep experience in novel retail solutions, spanning marketplaces, luxury, specialty, and mass retail after starting his career in the technology space. He most recently took on turning around The RealReal's business and applying customer-centric and data-driven tactics to enact meaningful change to the company's bottom line.

Please welcome both Marc and Koryl to RL, and you'll be hearing more from them soon.

Are you a Meta employee? Got insight to share? Contact the reporter Jyoti Mann via Signal at jyotimann.11 or email at [email protected]. Reach out from a nonwork device.

• Meta's CTO told staffers to "leave or disagree and commit" to working at the company.
• Andrew Bosworth was responding to some employees voicing concerns over recent policy changes.
• The company recently changed its approach to internal Q&A sessions to crack down on leaks.

Meta's chief technology officer, Andrew Bosworth, told some employees to quit and "consider working elsewhere" if they thought all staffers must like the company's policies.

Bosworth's remarks were in response to comments on a post he shared on January 30 in an open group on Meta's internal Workplace forum.

In the group, called "Let's Fix Meta," Bosworth shared an article by The Verge about CEO Mark Zuckerberg's comments to employees in an all-hands meeting that day, on which Business Insider first reported.

Along with the article link, he wrote: "As predicted, the entirety of todays Q&A leaked. It sounds like someone just gave the entire audio feed to a journalist. I saw all the angry/sad reaccs about the change to the format and I share a sense of loss about it, but I think this makes it clear it was the right call."

Commenting on the post in the group, which has nearly 12,000 members, one employee wrote: "1. Company changes policies to specifically target the LGBTQ community, 2. Cuts its own data-backed DEI programs, 3. Leadership goes on a far-right podcast to explain changes instead of addressing employees, 4. Limits free speech internally…and there's surprise?"

In response, Bosworth noted that in his original post, he indicated it wasn't a surprise, saying that apart from specifics, "if your view is 'everyone has to like all the policies we have and if they don't it is appropriate to leak' then I think you should consider working elsewhere."

Another person commented, saying they "agree leaking is not productive or rational," adding: "Emotional and scared employees don't do productive things."

In a separate comment, the same person said: "Blaming leaks for why Mark's policy decisions cannot even be discussed, much less appealed, is the slap in the face. We're all here because when we were hired, we were the best candidate for the job." The person added that employees were being treated poorly.

The exchange continued, with Bosworth replying: "You should quit if you feel that way, I mean it." He also expressed confusion at the suggestion that employees were being mistreated.

He added: "Unless you are referring to the policy changes, in which case Mark spent quite a while talking through them, it just sounds like you don't agree. In that case you can leave or disagree and commit."

Ahead of the company's January all-hands meeting, Meta's vice president of internal communications told employees in a Workplace post that the social media giant was changing its process for selecting questions in Q&A sessions.

"We will skip questions that we expect might be unproductive if they leak or things like People related questions that have already been answered," the executive wrote.

In other comments seen by BI, several Meta employees expressed concern over recent content moderation changes.

One employee criticized what they called leadership's silence on "transphobic/homophobic policies." Another asked where employees could voice criticism if internal discussions were discouraged. A third employee said the situation was making Meta a "more hostile place to work."

Meta has faced continued internal dissent in response to Zuckerberg's January announcement that it would cut "low-performers" and recent policy changes. Some employees recently questioned the company's removal of posts on Workplace, with one employee describing it as a "free speech issue."

Meta didn't immediately respond to a request for comment.

Are you a Meta employee? Got insight to share? Contact the reporter Jyoti Mann via email at [email protected] or via Signal at jyotimann.11. Reach out from a nonwork device.

The edtech giant has begun notifying individuals outside of the US and Canada affected by the breach

© 2024 TechCrunch. All rights reserved. For personal use only.