The team in Redmond certainly didn’t plan for 2026 to be “the year of Microslop”. But that’s what has been trending on social media in the past few days, with meme images and posts circulating among the tech crowd in particular.

All because of one blog post that Satya Nadella made on his new, personal blog site. Which I personally didn’t consider to be that bad. Yet there were immediately numerous headlines derived from the “We need to get beyond the arguments of slop vs sophistication” part of the blog post. It hit the collective nerve.

A very similar trending AI topic has been doing the rounds recently, after someone on Hacker News realized a thing that happened a year ago already. Yes, the website for Office products now does state “The Microsoft 365 Copilot app (formerly Office) lets you create, share, and collaborate all in one place with your favorite apps now including Copilot.” But Office as a brand has been deprecated way before this. Plus, I already created a better office.com last year with just a few minutes spent on Lovable, inspired by this lunacy.

All of this makes it seem like Microsoft isn’t all that popular these days. I believe this is not the result of any single misstep but rather a series of decisions that Microsoft made over the previous year. As a result, it is now starting 2026 at a weaker position than the year before, based on my personal assessment of the situation.

In this newsletter issue, I’ll take a 360 look at the different audiences with whom Microsoft needs to do a better job engaging with:

Let’s start from the segment this newsletter typically deals with, the business apps, then expand the discussion into other stakeholders.

Purchasing Microsoft 365 Copilot licenses has been the path of least resistance for most enterprise customers for infusing AI into their everyday work. Or at least on paper that has been the idea. The reality may resemble this parody more than most people in the IT business like to admit:

What happens in organizations like this after the initial Copilot licensing deal is done? Well, too often, the answer is: nothing. Which naturally is a commercial risk for getting customers to renew their subscription — let alone increase their paid seat count. To ensure that “the graph goes up and to the right”, there’s now a mad rush to ensure that the bought AI tools are also adopted into daily use.

Recently, The Information reported how Microsoft is now offering customers money to use M365 Copilot. Well, not directly, but rather via paying tens or hundreds of thousands of dollars for training when customers agree to buy a large number of M365 Copilot seats:

While not an unusual practice among enterprise software vendors, the idea how these AI assistants require similar adoption support efforts as complex business systems somewhat erodes the marketing message for Copilot. It also makes you question the validity of the claim in Microsoft’s response to The Information: “We are seeing strong adoption of Microsoft 365 Copilot.” Sure, Jan…

It sure looks like their CEO is seeing something else. Because it was also reported by The Information that Satya Nadella has internally expressed his dissatisfaction with how Copilot works in practice when trying to do basic things:

The leaked email discussion was presumably about Google Gemini vs. the consumer edition of Copilot. Still, as a paying business user of Microsoft 365 Copilot, I continue to experience the “Doh!” moments with my AI assistant on a regular basis. I’d definitely agree with Satya on the “not smart” part. While Copilot is improving, it isn’t catching up to what LLM-powered services in general are able to do.

I don’t do work tasks with Google products, yet the broad praise for Gemini’s latest abilities has made me consider trying the paid version. NotebookLM is on a completely different level than both Copilot Notebooks as well as Copilot in PowerPoint when it comes to creating structured, visual summaries from selected data sources.

Microsoft’s head start with OpenAI partnership allowed the company to push their Bing Chat and later Copilot products out into the market faster than Google could react. (Remember the “we want to make Google dance” flex from Satya in 2023?) Today, some are claiming OpenAI has lost its lead. Microsoft, on the other hand, has been unable to build an AI experience that would have prompted customers to choose Copilot rather than ChatGPT, despite having full access to OpenAI IPR thanks to their partnership.

At this point, we really should be further down the path of AI transformation that wondering whether Copilot can read our mailboxes or not. Yes, it has obviously not been as simple as slapping a GPT model on top of MS Graph and then calling it a day. I can respect the technical challenges that exist in taking traditional business information systems and trying to marry them with the new alien tech like LLMs.

But, you know: you did this to yourself, Microsoft. You sold the idea of apps being over and Copilot being the only UI we need. You burned your ships, remember? Your scientists were so preoccupied with whether or not they could…

Time will tell if initiatives like Agent 365 will manage to restore the faith of business users. To convince them that the toolkit available from the MS cloud aligns with real world needs to deploy, secure and monitor AI agents that customers will adopt from various different vendors. It could work. It could also launch the per-agent licensing model for a post-SaaS world.

Until then, I guess in 2026 the new normal will be to see Satya sending customers his Copilot prompts of the week in direct email marketing campaigns:

2025 was a tough year for Microsoft in the customer segment that’s traditionally been tough to begin with: consumers. It should be no surprise that the decision to forcefully upgrade M365 Personal & Family plans to the Copilot version caused outrage among users. It sent a message that “we’re gonna be doing this, whether you like it or not, because we know better”.

At least in the business segment, Microsoft needs to be polite towards decision makers and try to convince them of all the benefits and ROI of subscribing to more cloud services in multi-year agreements. With consumers, this common decency is sometimes absent — because there’s no high-profile decision maker to engage with as peers. Instead, it’s like herding cattle and getting them to just click on the things that the tech corporation wants.

The arrogance in communication coming from leaders like Mustafa Suleyman caused pushback not just in social media posts from individual influencers but also from tech media. One contender for headline of the year was PC Gamer’s take on the inability for the AI CEO to understand the reactions from customers:

2025 was quite a delicate moment in the personal computing space as Microsoft ended their support for Windows 10 (the version users generally felt comfortable with) and forced people to upgrade to Windows 11. Especially folks with perfectly working PCs that weren’t supported by Win11 requirements were eager to look for options like Linux — supported by community campaigns such as End of 10.

Those who chose to accept the MS route and went with Windows 11 faced a growing likelihood of bugs and issues. The reports about decline in quality were so widespread that even though I personally didn’t encounter a bad update messing my flow of work, the evidence is hard to ignore at this point. 2025 wasn’t a great year for Windows in terms of its reputation.

UPDATE: Oh the irony! Just when I was finishing up this newsletter issue and did a reboot of my Win11 Dell desktop, the whole thing melted down. Now, I’m typing this from an Ubuntu Linux mini-PC while my main Windows PC is going through surgery (data backup and OS wipe). There almost wasn’t a newsletter issue this week, thanks to… Well, either 3rd-party drivers or Windows 11 quality issues.

On the entertainment front, Xbox seems to be facing an even bigger crisis. The handheld devices branded Xbox Ally don’t seem like much of an Xbox and are more like half-baked Windows devices with an Xbox logo on them. The Xbox movie store was closed. Xbox Game Pass announced the largest price hike ever. Again, not being a gamer myself, it seems like customers who bought products and content from Microsoft in this segment didn’t have much to cheer for in 2025. Some are claiming Xbox is already officially over, thanks to how poorly it has done compared to competition.

In the minds of ordinary folks, the role of Microsoft increasingly appears to be “those tools we use at work”. That’s a tricky spot since most workers aren’t able to influence what tools are used at their office and how exactly. Being forced to use some tech at work is one reason people may want to exercise their autonomy while at home — and thus end up choosing an alternative tech provider for their devices, software, AI, and entertainment.

The individuals that arguably have the single biggest potential impact on Microsoft’s success are the ones with skills to fix the gaps between AI promises and AI reality. That’s a skill in high demand, though, and this is forcing even Satya Nadella to clear his calendar for high priority tasks. Like luring the brightest minds in the AI field to come and join the Redmond team.

The CEO of a near $4TB company making recruitment calls certainly is quite something. It reminds me of the old days when things were the other way around, and customers insisted on getting to talk with Bill Gates. That was the era of centralization and OS dominance, though. As expressed by Tim O’Reilly in his post about AI and the Next Economy, “Microsoft recentralized the industry by monopolizing operating systems and office applications” after the PC had first decentralized it. In such an era, you can choose who to talk to if you hold the central power.

The biggest and most impactful news is the salary level, though. We know things are getting crazy out there when OpenAI is paying employees $1.5 million each. No matter what the median salary really is or how the stock-based compensation eventually plays out, this sends a message to those who are in the potential call list for Satya. They don’t need to settle for pennies.

In the case of Microsoft in particular, the potential candidates out there may well be aware of what some call “The Deal” at Microsoft. Meaning, there’s a consensus on the company not paying the same level of salaries as its peers. In return, the requirements from employees aren’t as strict as they might be at Amazon or Meta.

So, less pay in exchange for less pressure to devote yourself to your job only, you say? Hey, that sounds almost like our Nordic model! Where the work ethics are relatively high, in exchange for the ability for anyone to leave the meeting at 4 PM to pick up their kids from the kindergarten, no questions being asked. Round here, most of us accept both lower pay and higher taxes in exchange for not having to focus 110% of your attention on delivering shareholder value. Because there are other potentially more rewarding values in life — at least if you believe the World Happiness Report results.

Unfortunately, when the markets heat up and VCs are searching for AI investment targets to accept their money, we rarely hear the “won’t someone PLEASE think of the children?!?” plea. Instead, the biggest salaries take over the news headlines and fuel the endless hustle culture posts on LinkedIn. When Microsoft leaders try to form their own superintelligence team that’s bound to be unlike 99% of the existing teams that ship established enterprise software products, the risk is that the hustle culture spills over to places where it harms the very foundations of company culture MS was built on.

“I am altering the deal. Pray I don’t alter it further.” The words of Darth Vader may now also be heard at the Redmond Death Star mother ship. 2025 seemed to be the year when the deal expired. Microsoft being a country club for programmers is no longer a valid description of company culture:

Despite the stock market seeing steady AI-fueled growth, the number of tech layoffs also grew dramatically as the ZIRP era ended. Microsoft joined the other big tech corporations in reducing headcount at a steady pace. Tom Warren from The Verge reported how Microsoft’s constant layoffs risked creating a culture of fear.

Employee morale was also tested at Build 2025 conference where Microsoft decided to give a platform to Elon Musk. At the moment when he was leading the charge of DOGE in dismantling the US government structures, putting especially minorities at risk, this understandably caused a lot of tension internally:

Today, at the start of 2026, Grok and xAI are making headlines thanks to the complete lack of guardrails in image creation. The current scandal about “CSAM on-demand” abilities of Grok is testing the very narrative of politicians on both the right and the left side — who usually can at least agree that sexual images of minors aren’t to be tolerated in the society. This headline from the Financial Times captures the incredible and absurd state of things:

Remember when tech vendors were still talking about Responsible AI? It’s been a while since I’ve seen any bold claims about responsibility, so I had to consult AI to check what it thought about Azure & Grok being best buddies. Could there be a teeny-weeny conflict between Microsoft’s responsible AI principles and what we’ve seen Grok do in practice? Surprisingly, Claude plus Gemini found “a few” issues:

This leads us to the very reason why the actions of one software company matter so much that I spend time analyzing and writing about it. Big Tech is really big. To put it into perspective, the current market cap of Microsoft is equal to the GDP of India. Their annual revenue is almost equal to the GDP of Finland. Microsoft’s net income (profit!🤑) is 75% of how much taxes my home country collects every year.

I vibe coded a quick Microsoft in Global Society scorecard page to put these metrics alongside the rightful concerns that arise from the sheer size of the software giant and what its possible impact is on the world. Today, everyone should be aware of how the urgency of digital sovereignty has risen to the top of EU’s agenda. In short, it means that the software licensing costs and cloud ROI are no longer the sole factors determining whether non-US customers can buy or use Microsoft services.

In 2025 we heard statements from Members of the European Parliament that called for completely removing Microsoft-based services from the public sector in Finland and in the EU. The extortion mechanisms available to the US government were activated with the ICC sanctions and that was the exact wake-up call cloud vendors would have wanted to avoid at all costs. Yet there was nothing they could do to block it, and now everyone has the concrete example to use when illustrating the risk of relying on foreign cloud services:

No matter how huge it is as a corporation, Microsoft cannot separate itself from the government where its HQ is based. I honestly don’t expect them to be a major player in driving the technoligarchy movement that is now putting liberal democracies to test all around the globe. But no matter whether you’re Palantir or Microsoft, you are bound by the physical reality of nation states and their power structures. You could claim to have “do no evil” as your org’s guiding principle, but when the President abducts your CEO in the middle of the night calls you and tells you to “do evil”, that makes precious little difference.

Microsoft surely realizes how bad these recent political events are for its business. In theory, they are probably the global tech vendor with most skills and possibilities to address concerns of us Europeans. They completed their EU Data Boundary in 2025 and have tried to ensure their sovereign cloud offering could allow keeping things as-is, rather than migrating away to something built on non-US tech like the EuroStack presents. And yet everything they say and build can be at any time destroyed by a single post on Truth Social. Welcome to 2026…

If politics are an area where the hands of Microsoft are tied, another crucial societal vector where they could decide to do the right thing is the environment. Instead, they’ve decided to walk back from earlier plans to become carbon neutral, let alone negative. My second most popular LinkedIn post of 2025 was the critique I shared about Microsoft abandoning their emission reduction goals. It was founded on this article about what happened to the carbon moonshot:

The only concrete thing that appears to be holding back the growth in emissions for the hyperscalers today is that there’s not enough energy supply available to power all the data centers they’d want to build. With the current lack of any governmental pressure to restrict emissions (“drill baby drill”), it looks like we can’t expect Big Tech to help us in preventing the climate crisis from shaping the world dramatically in the coming years and decades.

As long as the investors expect to reach a return on investment from scaling generative AI into ever bigger computing factories, it’s unlikely that Microsoft, Google, Amazon, Meta or Apple would stop increasing their energy and resource consumption. On one hand, it wouldn’t take much to stand out from all these players and be just slightly less destructive towards our shared habitat. But then, would there be a quantifiable reward from it that would offset the short-term competitive disadvantage?

Rather than seeing this as a failure from the tech corporations to restrict themselves, it is more about whether society in 2026 can find a way to make them do the right thing. Given the right incentives and sanctions, I think Microsoft and the rest would adapt to the limits of growth. If we’d demand innovation on this front, rather than demanding bigger AI models and unlimited data processing power, the R&D investments would flow into that direction.

If we don’t demand this, we’ll just get more of what we got in 2025.