​​​​​​AI Under Pressure

Thank you to our Sponsor: Morphic

For a while, the AI story looked simple. Companies were racing to build bigger models, buy more chips, sign more cloud deals, and announce larger data center plans. The dominant assumption was that scale itself would create the future. If a company spent aggressively enough, it would secure a winning position before rivals could catch up.

That phase is ending.

AI is now entering a harsher, more divided moment. The spending is still enormous, but the mood around it has changed. The market is no longer rewarding AI ambition on faith alone. It wants evidence. It wants revenue. It wants margins. It wants proof that all of this infrastructure, hiring, and product integration will produce durable business value instead of just headlines.

That is the new split inside AI.

On one side, the spending machine is still accelerating. Alphabet, Amazon, Meta, and Microsoft are expected to invest about $650 billion in AI infrastructure in 2026, up from about $410 billion in 2025. Bridgewater described the boom as entering a more dangerous phase because the capital demands are growing so large and the return profile is still uncertain.

On the other side, investors are becoming more selective and more nervous. Reuters reported in late January that markets were willing to tolerate huge AI spending only when it clearly translated into strong growth. When it did not, punishment was swift. Microsoft fell sharply after investors questioned whether its growth justified the pace of AI investment, while Meta rose because its AI efforts were more visibly improving ad performance and revenue.

That shift tells you something important. AI is no longer being judged only as a technology race. It is being judged as a business model test.

The market is asking a different question now

The old question was: who is moving fastest in AI?

The new question is: who is actually making money from it?

That sounds obvious, but it changes everything. It means companies can no longer rely on broad statements about transformation, productivity, or future potential. They have to show where the return is appearing and how durable that return might be.

Right now, the answers are uneven.

Some companies are benefiting directly from the buildout. The infrastructure layer still looks strong because demand is immediate and tangible. Chips, cloud capacity, data center construction, power, networking, and model hosting all sit close to the spending stream itself. These businesses are getting paid because the buildout has to happen before the broader AI economy can happen. Some investors have shifted toward infrastructure plays during the broader AI selloff for exactly that reason.

But many other parts of the market are being forced into a more uncomfortable position. If AI makes software creation cheaper, customer support more automated, coding faster, and data analysis more accessible, then some existing software and services companies may not gain from AI at all. They may be compressed by it.

In early February U.S. software and data services stocks were slammed by disruption fears, with the S&P 500 software and services index on track to lose about $1 trillion in market value since late January. The selloff reflected growing concern that rapidly improving AI tools could undermine parts of the software sector rather than simply enhance them.

That is the deeper split:

• AI is creating winners through infrastructure demand
• AI is creating anxiety in software and services
• AI is forcing a faster distinction between hype and measurable value

Why investors are getting uneasy

Investor anxiety is not coming from disbelief that AI matters. It is coming from the opposite. Markets now believe AI matters so much that ordinary excuses are no longer enough.

When a company says it is spending tens of billions on AI, investors want to know:

• Is this spending defensive or offensive?
• Does it protect an existing business or create a new one?
• Is the company capturing demand or merely keeping up with rivals?
• Will AI improve pricing power, retention, productivity, or growth in a visible way?
• How long will shareholders have to wait before the return becomes real?

Those are not abstract questions. They go directly to valuation.

A company can survive a period of heavy investment if investors believe it is building an asset with long term strategic value. But if the spending starts to look like an arms race with unclear payback, confidence can erode quickly. The new market mood is especially unforgiving when AI spending rises while growth disappoints.

This is why the AI narrative feels more brittle in 2026 than it did earlier. Expectations are higher. Capital budgets are larger. And tolerance for vague promises is lower.

Thank you to our Sponsor: EezyCollab

The labor story is becoming part of the return story

Another reason the split feels sharper is that AI is no longer just about products and infrastructure. It is starting to reshape workforces, and that creates a new layer of pressure.

Block became a major example of this change. The company said it would cut more than 4,000 jobs, nearly half its workforce, as part of a broader effort to embed AI across operations. Shares jumped after the announcement, showing how strongly markets can respond when AI is framed as an efficiency story.

But the worker response has complicated that narrative. Reporting from The Guardian showed current and former employees arguing that AI tools still cannot fully perform the judgment, strategy, oversight, and regulated decision making that many of their jobs require. Workers said the tools could help with tasks, but not replace the full value of experienced employees.

That matters because it exposes a growing disconnect in the AI economy.

Executives may present AI as a force multiplier that allows much smaller teams to do more. Investors may reward that framing because it signals margin improvement. But inside organizations, employees often see a more complicated reality:

• AI can accelerate narrow tasks without replacing full roles
• Quality control still requires human review
• Regulated industries cannot simply trust automated systems
• Customer facing AI can still make serious mistakes
• Institutional knowledge is hard to automate

So the labor question becomes part of the investor question. If layoffs are justified in the name of AI, but AI cannot actually absorb the missing work at sufficient quality, then the short-term market reward may create long term operational weakness.

The pressure to prove returns is now operational

This is the most important shift of all. The pressure is no longer only external. It is moving inside the company.

Leaders are under pressure to prove AI returns in quarterly results. Product teams are under pressure to launch features that sound meaningfully AI native. Employees are under pressure to use AI tools more often. Customers are being pushed toward AI agents and support systems whether they asked for them or not.

That creates a dangerous temptation. Companies can start optimizing for the appearance of AI transformation rather than the substance of it.

You can already see the forms that takes:

• counting AI usage rather than business value
• forcing AI into workflows where it adds friction
• overclaiming labor replacement before systems are mature
• using AI language to reframe ordinary cost cutting
• rewarding AI announcements faster than AI outcomes

This is why the next stage of AI competition may not be about who spends the most. It may be about who can prove the cleanest conversion from spending into results.

That proof will likely come from a few areas first:

• revenue growth tied to AI enabled products
• measurable productivity improvements without quality collapse
• customer retention gains from better experiences
• new margins created by automation that actually works
• lower costs that do not create bigger downstream problems

The companies that can show those outcomes will keep investor trust. The ones that cannot may find that the market has become far less patient.

What comes next

AI is not collapsing. The spending is too real, the infrastructure demand is too large, and the strategic importance is too high for that. But the market is entering a sorting period.

The winners will not simply be the loudest companies or the ones with the largest CapEx numbers. They will be the ones that can answer a more basic question with credibility: what, exactly, is this spending producing?

That is the split inside AI now. The money is still flooding in. The excitement is still there. But the easy phase is over.

From here on, AI has to do more than impress.

It has to pay off.

Looking to sponsor our Newsletter and Scoble’s X audience?

By sponsoring our newsletter, your company gains exposure to a curated group of AI-focused subscribers which is an audience already engaged in the latest developments and opportunities within the industry. This creates a cost-effective and impactful way to grow awareness, build trust, and position your brand as a leader in AI.

Sponsorship packages include:

• Dedicated ad placements in the Unaligned newsletter

• Product highlights shared with Scoble’s 500,000+ X followers

• Curated video features and exclusive content opportunities

• Flexible formats for creative brand storytelling

📩 Interested? Contact [email protected], @samlevin on X, +1-415-827-3870

Just Three Things

According to Scoble and Cronin, the top three relevant and recent happenings

​​​​Block Workers Push Back on AI Layoff Claims

Block employees and former staff argue that Jack Dorsey’s claim that AI productivity justified cutting nearly half the company is overstated. While they say AI tools can help with coding, support, and routine tasks, they insist the technology still lacks judgment, strategy, emotional intelligence, and the ability to handle regulated financial work without heavy human oversight. Many believe the layoffs were driven less by real AI replacement and more by investor pressure, while workers describe being pushed to train and use tools that are still too limited to do their jobs fully. The Guardian

Netflix Buys Ben Affleck’s AI Studio

Netflix is acquiring Ben Affleck’s AI filmmaking company, InterPositive, and bringing Affleck on as a senior advisor. The company’s tools help filmmakers use AI to handle technical production tasks like removing stunt wires, reframing shots, adjusting lighting, and improving backgrounds using material they have already filmed. The deal reflects Netflix’s growing interest in AI for filmmaking, while Affleck and Netflix both stress that the technology should support human creativity rather than replace it. NPR

AI Data Centers and the Return of the Man Camp

Temporary worker camps once linked to oil booms are now being used to house laborers building major AI data centers. The summary centers on Target Hospitality, which has secured $132 million in contracts for a large camp in Dickens County, Texas, where more than 1,000 workers could eventually live. The company sees the rapid expansion of U.S. data center construction as a major growth opportunity, but it also faces criticism because it owns an ICE family detention center in Texas that has been accused in court filings of poor food conditions and inadequate care for children with special dietary needs. TechCrunch

Scoble’s Top Five X Posts