This Week in AI & Automation

Week of February 24, 2026

Jack Dorsey cut 40% of Block's workforce this week and said every company will do the same within a year. OpenAI finalized the largest private funding round in history — $110 billion at a $730 billion valuation. Perplexity launched a digital worker that coordinates 19 AI models. And Deloitte shipped an enterprise AI transformation platform. The theme this week isn't capability. It's consequence.

The Big Story

Block Fires 4,000 People. Dorsey Says AI Made Him Do It.

On Wednesday, Jack Dorsey announced Block is cutting more than 4,000 employees — roughly 40% of its global workforce — taking headcount from over 10,000 to under 6,000. The reason, stated explicitly: AI.

In a 600-word post on X, Dorsey attributed the decision to a step-change in AI capabilities he noticed in December 2025, when models became "an order of magnitude more capable." He framed the cuts not as cost reduction but as structural transformation, stating the business is strong and gross profit continues to grow.

The severance is generous: 20 weeks salary plus one week per year of tenure, equity vested through end of May, six months of healthcare, corporate devices, and $5,000 in transition support. But the message underneath the severance was pointed. Dorsey wrote: "I think most companies are late. Within the next year, I believe the majority of companies will reach the same conclusion and make similar structural changes."

For enterprise leaders, this is the first Fortune 500-adjacent company to publicly frame AI as the explicit reason for cutting nearly half its workforce. Previous rounds of AI-driven layoffs — at Google, Meta, Amazon — were framed as "efficiency" or "restructuring." Dorsey skipped the euphemism. That honesty is either brave or reckless, but it sets a precedent every board will now reference.

Source: CNN Business, CNBC, TechCrunch

Our Take: Dorsey's prediction that most companies will follow within a year is aggressive but directionally right. The pattern we see in enterprise AI implementations is consistent: AI doesn't eliminate jobs one-for-one — it collapses entire workflow layers. The companies that thrive will be those who restructure their teams around AI capabilities rather than bolting AI onto existing headcount. Block just did it publicly. Others will do it quietly.

Notable Developments

OpenAI Closes $110B — The Largest Private Round in History

OpenAI finalized its funding round on February 27: $110 billion at a $730 billion pre-money valuation ($840 billion post-money). The round breaks down to $50 billion from Amazon, $30 billion from Nvidia, and $30 billion from SoftBank.

Amazon's commitment starts with $15 billion upfront and $35 billion contingent on milestones. OpenAI is also expanding its AWS partnership by $100 billion in compute services on top of the existing $38 billion commitment.

Last week we covered this round in progress. The final numbers came in higher than expected. At $840 billion post-money, OpenAI is now valued more than all but five public companies globally.

Source: Bloomberg, TechCrunch

Our Take: The Amazon-Nvidia-SoftBank trifecta signals something beyond capital. Amazon gets deeper model integration across AWS. Nvidia locks in a customer that will consume enormous GPU fleets. SoftBank gets exposure to what it considers the defining technology platform. For enterprises evaluating AI vendors, the takeaway is that OpenAI's infrastructure commitments guarantee continued aggressive investment in reliability and enterprise features — but also deeper cloud lock-in.

Perplexity Launches "Computer" — A 19-Model Digital Worker

Perplexity launched "Computer" on February 25: a general-purpose digital worker that coordinates 19 AI models on the backend. The system accepts a high-level objective, decomposes it into subtasks, and routes each to the best-suited model.

The architecture: Claude Opus 4.6 handles orchestration and coding. Gemini powers deep research. Nano Banana generates images. Veo 3.1 produces video. Grok handles lightweight tasks. GPT-5.2 manages long-context recall. The system can execute complex workflows autonomously for hours or months without requiring re-prompting.

The price: $200/month on Perplexity's Max tier. Unlike OpenAI's OpenClaw, Computer runs entirely in the cloud rather than requiring local hardware integration.

Source: VentureBeat, Fortune, TechCrunch

Our Take: The multi-model orchestration approach is the right architecture for production work. No single model excels at everything — Claude is strongest at reasoning, Gemini at research, specialized models at generation. The question is whether a $200/month SaaS can match purpose-built enterprise workflows. For most companies, the answer is: it depends on how custom your AI implementation needs are.

Quick Hits

• Deloitte launches Enterprise AI Navigator: An end-to-end enterprise AI transformation platform that delivers tailored roadmaps for moving AI from cost to value. The big four consulting firms are now all shipping AI products, not just strategy decks.
• Tabnine ships Enterprise Context Engine: General availability for what Tabnine calls the missing layer for enterprise AI code generation — organizational context that grounds AI suggestions in your actual codebase.
• UiPath targets healthcare with agentic AI: New solutions for medical records summarization, claim denial prevention, and prior authorization launched at ViVE 2026. Healthcare admin overhead is the perfect agentic AI use case.
• 500 Burger Kings deploy OpenAI-powered headsets: Called "Patty," the headsets flag low inventory, remove sold-out items from digital menus, and alert managers to facility issues — in real time across 500 locations.

Numbers of the Week

What We're Watching

The "AI layoff" narrative is about to accelerate. Dorsey's public attribution gives cover to every CEO who's been considering similar cuts. Expect a wave of restructuring announcements in Q2 framed around AI capability rather than macroeconomic conditions. The companies that handle this well will invest in retraining and role transformation. The ones that handle it poorly will just cut.

Multi-model orchestration is becoming the default architecture. Perplexity's Computer, OpenAI's OpenClaw, and Anthropic's Cowork all point to the same conclusion: the future of enterprise AI is not a single model but a coordinated system of specialized models. This has major implications for AI vendor selection — you need a platform strategy, not a model strategy.

Consulting firms are becoming AI product companies. Deloitte's Navigator follows McKinsey's Lilli and Accenture's myWizard. The big four aren't just advising on AI anymore — they're shipping production software. This blurs the line between advisor and vendor in ways that enterprise procurement teams need to think carefully about.

The Bottom Line

This week drew the sharpest line yet between AI capability and AI consequence. The same technology that attracted $110 billion in private capital also justified eliminating 4,000 jobs. Perplexity showed that coordinating 19 models is now a $200/month service. Deloitte proved even the consultants are shipping AI products now.

The question for enterprise leaders isn't abstract anymore. It's not "should we adopt AI?" or "which model is best?" It's the question Jack Dorsey answered on Wednesday: what does your organization look like when AI handles the work that 40% of your workforce does today? Every company needs an answer. The ones that wait for someone else to go first just watched Block go first.

Get This Week in AI & Automation delivered every Saturday.

Previous Editions

• This Week in AI & Automation: The $100 Billion Week (Feb 21, 2026)
• This Week in AI & Automation: The Age of Agent Teams (Feb 14, 2026)
• This Week in AI & Automation: The $470 Billion Question (Feb 1, 2026)

Have a story we should cover? Contact us.