The Problem with Microsoft
- Trevor Alexander Nestor
- Jul 7
- 17 min read
Updated: Aug 5
So, I was recently employed at Microsoft in a senior level position (April 2024 - June 2025), and what has happened to me should be alarming to everybody - especially those working in tech. Well, let's just say that what happened to me at Microsoft is a signal for what is to come for others, and it doesn't look good. Firstly, everybody has heard about the 9,000 job cuts Microsoft has done just this month, and the tone deaf suggestion by their Xbox executive to turn to AI tools to deal with the fallout of trying to survive in the United States while unemployed (AI can provide 'emotional clarity and confidence' Xbox executive producer tells staff after Microsoft lays off 9,000 employees). This is certainly unfortunate and devastating, but the issues at Microsoft go much deeper than that. Let me explain.
The workforce reduction at Microsoft hit some groups especially hard. Gaming studios under Xbox, for instance, faced game cancellations, entire projects shut down, and hundreds of roles eliminated. These cuts of thousands of workers that have been occurring in waves these past few months are just the tip of the iceberg. While some might argue that reductions in redundant staff might be a good thing to maintain workforce productivity with the evolution of technology, silently, Microsoft has been pressuring for the termination of employees, and that includes also on wrongful or illegal grounds.
Underneath the large-scale layoffs, they are also implementing a strategy of systematically gaslighting, lying to, and scapegoating their employees and engineers across the company. How do I know this? Well, this has happened to me, and I have been gathering stories from many others that this has been occurring to recently. The reasons for this are likely obvious - corporate executives wanted to reduce headcount due to changing tax laws, alleged performance boosts provided by AI, and outsourcing and offshoring of critical support staff to other countries (largely India).
As its new fiscal year began on July 1, Microsoft confirmed an additional 9,000 would be laid off companywide, including 830 cuts in Washington state. Combined with earlier waves in May and June, and silently coercing employees to leave with severance and terminations, Microsoft has now laid off large swathes of workers since spring, including more than 3,100 in Washington, home to its Redmond headquarters.
Although Microsoft claims cuts were meant to "flatten management," publicly available data shows that only 17% of impacted roles in Redmond were classified as managerial - meaning Microsoft executives have been lying. In April, immediately preceding the layoffs, Microsoft filed a federal lobbying disclosure, reporting $2.35 million in lobbying directed at the federal Department of Labor and the Department of Homeland Security – responsible for overseeing visa certifications and employment eligibility. "High-skilled immigration reforms" is a policy goal associated with expanding access to the H-1B visa program.
Microsoft has consistently claimed that the U.S. does not issue enough H-1B visas to meet labor market demands. On its lobbying blog, the company states: "The number of H-1B visas remains very limited. Congress sets a limit on the number of H-1B visas that can be issued each year. Demand far exceeds supply." While it is accurate that U.S. law caps new H-1B visas at 85,000 per year, Microsoft's hiring data offers additional insights. From 2021 to 2024, Microsoft submitted H-1B requests at a rate of 5.17 for every one net new job it created in the U.S. So, we should begin to evaluate if there really are "labor shortages" in the United States, or if Microsoft really would just like to sidestep paying workers a fair wage - especially as cost of living has become increasingly unobtainable, particularly around the Microsoft HQ (I spoke to one coworker that, in spite of working as a Microsoft engineer for nearly 10 years, has only just begun to be able to afford a 3 bedroom home - bizarre from an engineer at the highest market cap corporation on Earth in history).
Microsoft's claims of a lack of available skilled workers appear less about unmet labor needs and more about the company's inability to replace its existing workforce quickly as much as they would like due to the legal limits imposed by the H-1B visa cap - and my case is a prime example.
Tech companies have coordinated mass firings while continuing to import H1B to stagnate wages and lower workplace expectations and accountability where visa holders carry anxiety about revocations of their visas, leveraging the anxiety and culture of concealment they grow in the workplaces to implement widely disliked policies (RTO), thus coercing remaining employees to work more for fear of losing their job. This atmosphere then extends to US citizens in the workplace and drives resentments between culture groups and has contributed to the declining workplace culture surrounding big tech giants, and is reminiscent of strategies in previous decades:
Let's start with my story, without breaching any privacy concerns (aka, not including company emails, teams DMs, names of employees unless they are posting publicly, etc). The rhetoric around AI's alleged ability to replace workers is ubiquitous, but the issue that I came across was that Microsoft's AI tools were not trained on evolving security processes, and would produce garbled nonsense when attempting to use it to assist in any way, other than to summarize meetings (in teams, one is able to turn on transcription, so that you can ask questions of the AI rather than interrupt the meeting).
Documentation was wrong, outdated, misleading, or just missing, meaning that tribal knowledge kept by employees that were laid off just went missing, and frequently could not be independently intuited by just looking at code repos (such as when documents neglected to call out permissions that were required for certain tasks, or with Satya's recent security pushes the AI tools are unaware of the procedures - just as two examples). I might even speculate that information is kept deliberately opaque so that engineers or employees have to get the information that is missing from word of mouth - that way if those up the hierarchy like you, they will freely share, and if not, you can be siloed and then axed citing performance issues when lack of the critical information inevitably causes significant delays.
In my case, the issues I've collected are numerous, but some areas that are particularly nefarious are wrongful termination, lack of ADA accommodations (ADHD), and whistleblower reporting of violations of security policy.
Documentation for my onboarding alone was so bad that after months of gaslighting me about it, eventually there was enough complaints that the entire org had to do a hackathon just to fix them. This behavior of gaslighting until things would reach boiling points was a repeated pattern. The entire time I would hear from my manager to "self-learn" or "self-unblock" (which would be a violation of security policy when I did not have permissions to do so) - claiming that "nobody else" was encountering the issues that I was. This was definitely gaslighting - in discussions with several teammates, virtually all of them reported similar issues that I was reporting - especially around evolving security processes that cannot be independently intuited and required outside interventions (Prioritizing security above all else - The Official Microsoft Blog). Just getting the local build environment functional took weeks due to misleading documentation. The difference is, some employees are given timely assistance to resolve issues that require interventions, and others are not - and it does not matter how nicely, formally, or professionally you ask.
All while I was continually gaslighted about my alleged "performance," staff levels above me would frequently ask me for assistance - in fact, while others reported they had a mentor their entire first year for getting newer employees up to speed with proprietary Microsoft processes, this was denied to me - all while tasking would commonly be missing any descriptions at all.
This behavior raises major ethical questions around accessibility and ADA accommodations (which my doctor did explicitly request - including mentorship and instructions) - particularly with those who have conditions like ADHD - and we should also caution that this will only grow worse as AI is deployed in classrooms. In my case, when I would ask questions (such as why certain tables do not exist or the time frame that results are kept for their database which tracks incidents) that would be reasonable for any new employee to ask, especially when documentation was repeatedly found to be wrong, I was treated in a passive aggressive manner and months later I would hear about it, creating a culture of fear and concealment where I repeatedly found multiple employees would all be facing similar issues but nobody wanted to talk about them or work them out to save face.
In fact, in one example when asking basic yes/no question to a coworker which could have saved hours of time, I noted she could see my DM, and deliberately decided to ignore it. When I brought this to the attention of one org manager, I was told that this sort of behavior was okay just simply because she did not feel like responding - all while I was shamed heavily for not responding to an email right away when Outlook frequently would not refresh my inbox such that I would not get mail until several hours after it had been sent to me (regardless of restarting the computer or application), and my immediate manager (who had a mentor the entire time I was employed) failed to respond to my emails for days and would often not appear to weekly syncs without explanation.
On one occasion I was repeatedly told by my manager that as a senior engineer, it was my sole responsibility to debug my own builds - in spite of the fact that I encountered hundreds of build errors that prevented integration of my feature into the repository. After reporting this to the manager, he repeatedly denied this had anything to do with anything outside of my control, and that allegedly "nobody else" was encountering the issues I was seeing. Weeks later and after having had this looked at by other coworkers in roles above me, they admitted that these errors were due to code refactorization in the build systems and several coworkers indeed did report the same issues. Until and up to my wrongful termination my manager continued to deny reality and scapegoat me for these types of delays.
AI, in this case, can be seen as a tool of surveillance, gaslighting, and plausible deniability - by overselling the claims of AI, it is possible then to defer responsibility for its failures. A mid-2025 analysis by Attila Hajdu surveyed multiple agentic AI evaluations and found that, on average, autonomous agents fail at their assigned tasks about 62% of the time (https://hajduattila.com/ai/2025/05/12/why-62-of-ai-agents-fail-and-why-its-our-fault/). Researchers at Carnegie Mellon University ran a simulated software-company workflow benchmark and observed failure rates between 70% and 90% on typical multi-step business tasks (https://www.deep-analysis.net/is-the-agentic-ai-glass-half-empty-or-half-full/). Salesforce’s CRMArena-Pro study reported only a 35% success rate for multi-turn customer-relationship-management scenarios - implying about a 65% failure rate (https://ppc.land/salesforce-study-reveals-enterprise-ai-agents-fail-65-of-multiturn-tasks/). Troublingly, AI is trained on documentation, and without documentation, internal AI tools lack context to navigate evolving processes or even for ingesting massive repos - often with thousands of files - each with over 10,000 lines of code.
When I reached out to the Microsoft research group regarding some of my findings relating to Majorana physics (the physics implicated in Microsoft's novel approach to quantum computing), I was also chastised for reaching out to others outside of my immediate team, in spite of the fact that now leading scientists have cited my work. Hardly a supportive environment of creative innovation... The message was clear - you are going to be siloed in a box like a performing monkey - do not reach out for help, do not seek the perspectives of or bounce ideas off of others who have worked at Microsoft for 5 or more years, just keep your mouth shut and use the (nonfunctional) AI, (missing) documentation, and (nonexistent) support, so that we can then scapegoat you for the resultant delays and claim it is your performance.
Surveying 31,000 full-time knowledge workers across 31 countries between February and March 2025, Microsoft’s 2025 Work Trend Index Annual Report revealed a startling new use case for AI: employees are increasingly deploying generative tools as social shields rather than purely productivity aids, and Microsoft is creating this culture of concealment deliberately.
Alarmingly, 17% of respondents said they turn to AI out of fear of human judgment, 16% to avoid the need for explanation or hand-holding from colleagues, 15% to take full credit for work, and 8% to sidestep the everyday friction of collaborating with co-workers - highlighting a rising trend of anxiety and social avoidance in today’s workplaces which exactly mirrors my experience The 2025 Annual Work Trend Index: The Frontier Firm is born - The Official Microsoft Blog.
Employees, in other words, are to be like brains in a vat - used to train their AI - failing to recognize that value within organizations scales up exponentially with collaboration, which will then be reflected in what the AIs glean. Reed’s Law states that network utility grows as 2^N, where N is the number of participants, so the number of possible subgroups (and thus the collaborative value) explodes exponentially. Harvard Business Review further observes that when organizations shift from treating collaboration as a value to actively teaching it as a skill, they unlock disproportionately large performance gains across teams.
I was given a Performance Improvement Plan (PIP) due to delays on my feature work, which were due to delays coming from IT issues outside of my control. Immediately after reporting these issues to HR, and that my manager had explicitly instructed me to work around security and overlook when I reported unauthorized access to sensitive files I shouldn't have had access to (that I was only aware were sensitive due to recognizing code names in them that signaled that which I remembered from my previous affiliations amd career history) I was retaliated against. Under suggestion by other coworkers I filed a Report-IT-Now case with HR to no avail.
Microsoft has been outsourcing their IT support to Wipro, which I found would repeatedly lie, put me on endless holds, and hang up the phone on me. More troublingly, Microsoft no longer has their own IT department on campus - and while they are offshoring critical IT support to contractors in other countries, they have also been victim of high profile cyber attacks that have targeted US companies and government resources on Azure cloud by attackers in other countries. This problem extends not only to employees at the company - but also to customers who depend on Microsoft support for cloud services and other Microsoft products.
It is important to not view this problem under the lens of racism towards those in other countries, and it could be good to increase opportunities abroad - however, the problem becomes when this is used to undermine basic functional working conditions that are needed for employees to perform. When I was reporting possible security breaches, I was told to overlook them. In fact, my PIP explicitly instructed me to "work around" the SAW security - which beyond being a violation of the security policy and company ethics, was not technically feasible.
After reporting that the PIP was not actionable, all HR did was attempt to re-interpret or re-word the PIP, which did not change the underlying issue - in order to progress on my feature, the tooling that was required was on the SAW, and there was no way to "work around" that - coworkers and managers attested to that. In other words, I was set up to fail.
There was a whole month long delay for example just to get a replacement Yubikey due to Yukibey shortages alone. My Secure Access Workstation (SAW) which contained secure tooling I needed to access failed with a spontaneous hardware fault, then the same happened to my replacement, and then my loaner as well. Managers were not aware of KPIs I was driving leading up to the PIP, and even recommended me to transfer to another org - pointing to a hasty decision.
Before my PIP I was given a severance agreement and told that if I did not sign the agreement, I would be forced onto the 45 day PIP. I did not sign the agreement, and then, only 4 days in to my 45 day PIP, I was terminated. For the entire 4 day duration of my PIP, I did not have access to any basic functional Microsoft assets to do any tasking. None of my company assets worked (hardware faults), the IT line was not functional, all on-site IT had been outsourced to Wipro, and org management would (like they did since the beginning) gaslight me heavily if I reached out to them or coworkers.
What was also strange, is that in my PIP, I was simultaneously told to not depend on others, but also to depend on others more to ensure issues were resolved quicker (see any issues with that with what you have read so far?). So, I was told to either take a severance or be forcibly put on a 45 day long PIP, then I was not provided the basic functional Microsoft work environment to succeed, and then wrongfully terminated only 4 days in without an opportunity to even begin to perform, and denied the severance.
After my wrongful termination (yes, this would technically still be illegal, even with "right-to-work" laws), when requesting copies of my PIP (which fortunately many other employees saw beforehand and I was able to get snapshots of), HR first lied, claiming they lost the PIP. Then after pressuring them, they finally gave me the PIP - but it was blocked due to Microsoft security - then they had attempted changing it to try to cover up parts that I had initially criticized as inactionable. HR also did not even begin investigating my reports until after I was terminated, pointing to serious lack of foresight on their part.
In previous years, there have been notable examples of Microsoft’s troubling employment practices which include the 2012 case Michael Mercieca v. Tracy Rummel & Microsoft Corp., which resulted in an $11.6 million jury award after finding wrongful discharge; following the May–June 2023 mass layoffs, Sanford Heisler Sharp McKnight opened an investigation that could yield substantial lost‐wage and punitive damages; in Thompson v. Microsoft (5th Cir. 2021), the Fifth Circuit affirmed summary judgment for Microsoft even after it removed an ASD-diagnosed employee from his EA role, leaving him with no compensation or accommodations.
In March 2022, whistleblower Yasser Elabd alleged the company paid over $200 million per year in bribes and claimed he was fired in retaliation; more recently, in July 2024 Microsoft agreed to pay $14 million to settle a California probe into retaliation and discrimination against employees taking medical or family-care leave; and under the company’s “ManageRewards” performance-management system, cash awards are tied to slider scores - employees rated 120 or higher receive significant bonuses, while those scoring below 80 see minimal to no cash - compounded by abrupt terminations often without severance, leaving affected workers with zero compensation.
So, besides AI and outsourcing, why is Microsoft doing this, in spite of record profits? While Microsoft owns empty offices all over Redmond and Bellevue - why are they claiming that they don't have the resources for their employees? There is some recent legislation to consider which incentivizes these recent sort of layoffs.
Under the 2025 Tax Reform Act (signed in late June 2025), businesses regained several deductions that were due to sunset. For starters, domestic R&D outlays incurred after December 30, 2024, can once again be deducted in full in the year spent - just as they were before 2018 - which parallels the severance write-off strategy by letting firms accelerate big expenses without spreading them over multiple years. Section 179 expensing also climbed to a $1.25 million ceiling (with indexation thereafter), giving companies a massive near-term shield against taxable income when they invest in equipment, while severance remains equally “front-loaded” on the books. At the same time, interest-expense rules under IRC §163(j) have been rolled back toward their pre-2022 form: for tax years beginning after December 31, 2024, the cap on deducting net interest (once pegged to EBITDA) shifts back to an EBIT basis, loosening limits on how much borrowing a company can write off.
In practice, that means firms can borrow more aggressively to buy plant and kit - or even fund severance obligations - while still capturing sizable deductions against book income.
Finally, the House-passed “One Big Beautiful Bill” would make permanent the 100 percent bonus-depreciation regime that was always slated to decline after 2022. If enacted, every dollar spent on qualifying machinery, software, or even “qualified improvement property” could be deducted immediately, further stacking capital-write-off incentives alongside personnel-reduction deductions.
Under the Inflation Reduction Act of 2022, any company reporting over a billion dollars in annual profits (Microsoft has been laying off and wrongfully terminating employees in spite of record-breaking profits) now faces a 15 percent alternative minimum tax based on its financial-statement income rather than its taxable income. That shift means that firms can shrink their minimum-tax base by accelerating deductible expenses into the current year. Severance payments and accelerated vesting of stock awards are fully deductible when paid, so a well-timed round of layoffs, especially in the fourth quarter, can meaningfully cut a company’s book-minimum-tax liability.
At the same time, limits on deductible executive compensation have tightened. Section 162(m) used to cap deductions for pay above $1 million only for the CEO, CFO, and a few other top officers; today, thanks to rules rolled out under the American Rescue Plan Act of 2021 and IRS guidance coming into force after 2026, the cap will apply to more senior employees and eliminate deductions for most performance-based awards.
Faced with the prospect of nondeductible compensation, companies are reevaluating which C-suite and senior roles they actually need, and in some cases are restructuring or reducing headcount to stay below the threshold. Layer onto that the evolving limits on interest-expense deductions under Section 163(j) and the revived ability to expense domestic R&D costs immediately under Section 174, and you see why labor costs suddenly look far more flexible. Instead of carrying personnel expenses forward over several years, a lump-sum severance write-off can produce an immediate tax benefit that outstrips the cost of maintaining the same headcount into the next fiscal year. Mass layoffs are also being driven from the top down; including by hedge funds: https://www.tcifund.com/files/corporateengageement/alphabet/15th%20November%202022.pdf
So, what can be done about this? Since going public with my story, I have received comments and emails every day from others both inside the company and who have also been wrongfully terminated, and there seems to be a fear that by speaking about these issues, one might end up on a "blacklist" and will be excommunicated from the industry. This is not true - in the past even before employment for Microsoft, I have publicly disclosed my layoff from Meta, and have publicly criticized Boeing after being terminated for reporting that the power strip they had selected could catch on fire. Fear of retaliation runs high in the industry that contributes to a decline of work culture in tech, particularly for H1B visa holders who could lose their visa status if they are laid off.
Others have suggested that I might compromise a possible case by discussing things publicly - but the public pressure and discussion are precisely what is needed to organize any larger lawsuit or action (within NDA limits - which means no internal chats or specifics can be publicly disclosed) - which has been strengthened by the stories of numerous others across Microsoft and the industry that I've had discussions with. These companies rely on creating a culture of fear, insecurity, and concealment because executives know that it is easier to gaslight or scapegoat anybody once they are isolated - and thus avoid accountability.
The good news is, it is possible to fight and win - in fact I've had discussions with others only a week after providing my story that have had encouraging stories in their cases against Microsoft and these tech giants. While finding representation can be challenging in an industry that often overwhelms local legal services, I've even had encouraging discussions with attorneys about my case - though they have admitted they personally do not have the bandwidth.
When there are Microsofters from around the company currently agreeing with what I've written here, there is definitely a larger problem at the company. What I can say for certain, is in my case, Microsoft technically made a fatal error.
