Are 40% Staff Cuts the New AI Normal?

February 27, 2026 · Episode Links & Takeaways

Author

AI Daily Brief
February 27, 2026

HEADLINES

Nano Banana 2

Google dropped Nano Banana 2, which brings the reasoning and text-rendering capabilities of Nano Banana Pro to the Flash tier — meaning it's roughly half the cost and dramatically faster. It's now the default image model across all subscription tiers. VentureBeat framed this as a land grab, noting Qwen Image 2.0 is cheaper still, but arguing the real play is production-ready infrastructure: "The next wave of enterprise AI image adoption will be driven not by the models that produce the most beautiful images, but by the ones that produce good-enough images fast enough and cheaply enough to deploy at scale." Google is increasingly flexing the integration of all their systems — CEO Sundar Pichai demoed "Window Seat," which combines Nano Banana 2 with live weather data and geolocation to generate views from any window in the world.

Claude Signups Triple

Anthropic is having one of their strongest growth periods on record. The Information reports daily signups for Claude have tripled since November, paid subscribers have more than doubled since October, and free users are up 60% over the past month. Anthropic says Claude Code and Claude Cowork are driving the surge. One of the fascinating phenomena right now is that technical complexity doesn't seem to be as big a barrier to adoption as it used to be — when it comes to AI work tools, people are willing to go the extra mile if they can really get benefit out of it.

IBM Tanks on Anthropic COBOL Blog

IBM lost 13% on Monday — their largest single-day drawdown since March 2020 — and the trigger wasn't even a new Anthropic feature. It was a blog post about using Claude to modernize legacy COBOL systems. IBM does far more than maintain COBOL, and Anthropic first demoed this capability three months ago. Last June, Morgan Stanley profiled using OpenAI models to save 280,000 developer hours reviewing 9 million lines of COBOL. This is a very clear example that market participants aren't reacting to new developments — they're catching up on more than a year of AI advancements and seriously thinking through the implications for the first time.

Meta Scraps Advanced AI Chip

Meta has pulled the plug on their most advanced custom AI chip after hitting design roadblocks. The MTIA program has now seen two training chip iterations cancelled — the bet now seems to be on a simpler inference chip. This comes alongside massive chip-buying deals with NVIDIA, AMD (valued at over $100B), and a new multibillion-dollar deal to rent Google TPUs as a training cluster. What it feels like is that the calculus around paying the NVIDIA tax has changed — custom silicon projects just aren't as valuable as getting GPUs on racks at any cost.

Microsoft Gets OpenClaw-ified

Microsoft launched Copilot Tasks — an agent with its own virtual computer and browser designed for offloading mundane tasks. They described it as a "to-do list that does itself." It's a limited research preview for now, but if you wanted any clearer sign that everyone is getting OpenClaw-ified, look no further.

MAIN STORY

Are 40% Staff Cuts the New AI Normal?

Jack Dorsey announced that 4,000 employees at Block would be laid off — a 40% reduction in headcount, making this the largest tech layoff on a percentage basis in recent years and one of the only ones to explicitly cite AI as the reason. Block's stock soared 25% on the news. But whether AI is really at the core of the decision is one of the most interesting questions in tech right now. We are in a recalibration moment — everyone from AI practitioners to Wall Street investors to white collar workers are grappling with the tools having crossed a critical threshold over just the last few months. That process is going to be chaotic.

Dorsey's Memo: "Something Has Changed"
He doesn't actually use the term AI in the memo — which feels very intentional. He's not talking about a single tool but the entire system that surrounds getting work done.
Dorsey wrote: "we're not making this decision because we're in trouble. our business is strong… but something has changed." Block incubated an internal AI agent called Goose last year, originally for coding but since deployed across non-technical teams. On the earnings call, Dorsey validated the December inflection point: "Something happened in December of last year, where the models got an order of magnitude more capable and more intelligent, and it's really shown a path forward in terms of us being able to apply it to nearly every single thing that we do."

The Overhiring Skeptics
To date, we don't actually have a really clear example of a company massively slashing headcount due to AI efficiency gains and having that actually be the case.
Block tripled headcount from 3,900 to 12,500 between 2019 and 2022. By comparison, Robinhood operates with 2,500 employees at a $70B market cap; Coinbase has 4,500 at $50B. Bond investor Will Slaughter called it "unwinding an insane COVID overhiring binge" and said AI efficiency is "a much more appealing cover story" than managerial incompetence. Dorsey actually came back to address this, acknowledging the overhiring but arguing Block is now targeting $2M+ gross profit per person — 4x pre-COVID efficiency. The Amazon, Duolingo, and Klarna precedents all showed CEOs framing layoffs around AI but later walking back the connection.

The Market Reaction
Even a 25% jump wasn't enough to put Block back on firm footing — the stock is down 80% from its all-time high and still isn't back to its opening price for this year.
The 25% overnight surge was extraordinary even by the standards of layoff-driven stock pumps, but Block is still down year to date. . The concern: other CEOs will see that reaction and want to replicate it.

The Narrative Is the Narrative
It doesn't matter whether Block can really find 40% productivity gains from AI or if this is just the reflux from the tech hiring binge. The only thing that matters is that the narrative is now believable — and Jack Dorsey has fired the starting gun on the race to realize AI productivity.
When companies cut 40% of staff or an Anthropic plugin announcement wipes $40 billion off a company's market cap, those things are happening not because people have a clear sense of where we are — they're happening because we are unmoored and have no sense of where we are. We are in the midst of a dramatic repricing of everything. That said, the more dramatic nature of this move will help people slough off their complacency. Efficiency cuts are not the end game for AI and work — this is a period, potentially a very painful one, that we have to get through to get to the other side where the real opportunity lies.

ALSO REFERENCED