Happy Saturday. Your inbox is probably already stuffed with AI newsletters written by AI, which is either very meta or a cry for help. Either way, you're reading the one written for people who run actual businesses, not data centers. Let's get into it.

Google Rewrote the Enterprise AI Playbook This Week. Merck Signed the Check.

Google Cloud Next 2026 was the biggest enterprise AI event of the year, and the headline deal tells you everything: Merck, the pharma giant with 75,000 employees and $65B in annual revenue, committed up to $1 billion over a decade to deploy Google's new Gemini Enterprise platform across its entire business, from drug R&D to manufacturing to corporate back-office. Google Cloud engineers will work on-site alongside Merck teams. This is not a pilot. This is a company betting its operating model on AI agents.

The Merck deal is the clearest signal yet that the 'should we do AI?' debate is over in industries like healthcare and pharma. Novo Nordisk just struck a similar deal with OpenAI. Eli Lilly signed a $2.75B AI drug-discovery agreement with Insilico Medicine. The pattern: non-tech companies are not dabbling, they are naming a primary AI vendor and going deep. The board question is no longer 'are we exploring AI?' It is 'who is our AI anchor partner, and why?'

• Google also launched its Gemini Enterprise Agent Platform, a no-code tool that lets any employee build custom AI 'agents' that work across your apps, basically an AI assistant that can actually do tasks, not just chat.
• Deloitte announced a dedicated agentic AI practice built on Google's Gemini, with plans to expand from 25,000 to 100,000 licensed professionals, meaning your consulting firm will have AI co-workers before your company does.
• Google's own internal stat: 75% of all new code at Google is now AI-generated, up from 50% just last fall, a data point that should alarm any CEO who thinks software development is safely human-only.

Looking ahead, Google I/O on May 19 will bring more consumer-facing announcements, but the enterprise AI arms race is already running at full sprint.

OpenAI Drops GPT-5.5. Your Renewal Timing Just Got Very Interesting.

OpenAI released GPT-5.5 this week, calling it its 'smartest and most intuitive' model yet, designed for agentic work, meaning AI that can complete multi-step tasks on its own. The company also confirmed it is in talks with Wall Street banks about a Q4 2026 IPO targeting a valuation close to $1 trillion. Right now, enterprise AI pricing is a black box, and that opacity benefits OpenAI, not you.

When the S-1 lands, you will finally see real margin data, customer concentration numbers, and benchmarks to negotiate against. If your OpenAI or ChatGPT Enterprise contract is up for renewal before the IPO filing, you may want to push that into Q1 2027. Enterprise buyers with contracts due before the filing window opens are negotiating blind against a company that has all the information.

• GPT-5.5 is available now to Plus, Pro, Business, and Enterprise users, and OpenAI says it outperforms Google's Gemini 3.1 Pro and Anthropic's Claude Opus 4.5 on key benchmarks, though of course OpenAI did the scoring.
• OpenAI's enterprise business now makes up more than 40% of its revenue and is on track to reach parity with its consumer business by end of 2026, so the company is laser-focused on your budget.
• Anthropic, not sitting still, is reportedly fielding investor offers that would value it at $800 billion, more than double its $350 billion valuation from just February, a sign that the AI arms race is nowhere near its peak.

Looking ahead, OpenAI's 'super app' vision combining ChatGPT, Codex, and an AI browser into one enterprise platform could replace several of your current software subscriptions, which is either exciting or terrifying depending on which vendors you like.

The EU AI Act's Biggest Deadline Is 99 Days Away. Most Companies Are Not Ready.

August 2, 2026 is when the EU AI Act's transparency and high-risk AI rules kick in for most organizations. If your company operates in Europe, uses AI that touches EU residents, or deploys AI tools in HR, lending, or critical operations, this applies to you even if you are headquartered in Ohio. The regulation mirrors GDPR in reach, and the fines mirror GDPR in pain: up to €35 million or 7% of worldwide annual revenue for the worst violations.

The good news is there is a proposed 'Digital Omnibus' delay that could push some high-risk deadlines to late 2027. The bad news is that delay is not final, negotiations are still ongoing, and compliance lawyers say to treat August 2026 as the real date. On the US side, the patchwork of state AI laws keeps growing, and cyber insurance carriers are now adding AI-specific riders that can deny coverage if you lack documented AI governance. That last one is the one to put on your CFO's desk today.

• High-risk AI includes resume screeners, employee performance tools, credit scoring, and biometric systems, so if you use any off-the-shelf HR or lending software with AI features, you are likely in scope.
• A 42-state attorney general coalition has signaled coordinated enforcement against AI violations in the US, meaning state-level risk is real even without a federal law.
• The practical first step: ask every AI vendor for their written position on model-level safeguards and audit logs, the same way you asked for a SOC 2 report five years ago.

Looking ahead, the EU's compliance deadline is close enough that if your legal team has not started an AI system inventory, this weekend is a good time to send that email.

92,000 Tech Layoffs in 2026. Half Blame AI. The Truth Is Complicated.

Over 92,000 tech workers have been laid off so far in 2026, with Meta and Microsoft alone announcing more than 20,000 cuts this week. Almost half of the earlier cuts, roughly 37,600 positions, were officially attributed to AI and automation. Meta is cutting 10% of its workforce while simultaneously planning to spend more than $135 billion on AI this year. The pattern is stark: companies are replacing headcount budgets with compute budgets.

The honest truth, per OpenAI CEO Sam Altman himself, is that there is real 'AI washing' happening: some companies are blaming AI for layoffs they would have made anyway. IBM is actually tripling entry-level hiring this year, arguing human oversight of AI is essential. The real risk for non-tech CEOs is not which side of this debate is right. It is that your best mid-career employees are watching this news and updating their resumes. How you talk about AI internally matters as much as which AI tools you deploy.

• UKG, an HR software company backed by Blackstone, laid off 950 employees this month citing AI, which is a pointed signal when an HR platform is cutting HR jobs due to HR automation.
• Logistics firm C.H. Robinson previously cut about 1,400 jobs after rolling out AI tools for pricing, scheduling, and shipment tracking, a direct real-world example of AI replacing operational roles in a non-tech industry.
• Salesforce laid off 4,000 customer support roles, with CEO Marc Benioff stating AI reduced the need for human agents, a preview of what agentic AI could mean for any company with a large support function.

Looking ahead, Cognizant's chief AI officer says it will take another six to twelve months before most companies start seeing real productivity gains from AI, meaning the pain may arrive before the payoff.

Venture Capital Just Had Its Biggest Quarter in History. Most of It Went to Four Companies.

Investors poured a record $330.9 billion into startups globally in Q1 2026, more than double the prior quarter. Four companies absorbed 65% of it: OpenAI ($122B), Anthropic ($30.6B), xAI ($20B), and Waymo ($16B). That concentration is important context: the AI investment boom is real, but it is narrowly distributed at the very top. For everyone else in the startup ecosystem, early-stage funding is healthy but not historic.

For non-tech CEOs, the more relevant M&A story is the mid-market one. Acquirers are paying premium prices for vertical AI companies, meaning AI tools built specifically for healthcare, legal, logistics, or finance. If your company owns proprietary data in a specific industry and you have not thought about your AI acquisition value, your board should ask the question. Companies without AI moats are having a much harder time raising capital or commanding acquisition premiums.

• Medical AI company OpenEvidence raised $250M at a $12B valuation, with over 700,000 physicians already using it for clinical decisions, proof that sector-specific AI commands serious multiples.
• Legal AI firm Legora raised $550M at a $5.55B valuation, joining Harvey as evidence that every licensed profession is becoming an AI acquisition target.
• Gartner forecasts global AI spending will hit $2.52 trillion in 2026, a 44% increase year over year, meaning your vendors' AI price increases are not arbitrary, they are structural.

Looking ahead, the rule of thumb from acquirers is blunt: if your AI features are not driving measurable customer retention or expansion, there is no premium, just a regular acquisition price.

Half of Business Leaders Think They Can Scale AI. Half of Them Are Wrong.

A fresh survey of 1,000 C-suite and director-level leaders across retail, food and beverage, manufacturing, automotive, and logistics found that 80% believe their organization has the internal capability to manage an AI implementation. But 49% are still stuck in the pilot phase. The gap between confidence and execution is the defining business problem of 2026.

The three barriers that actually stop companies from scaling are data security and compliance concerns (36%), lack of internal AI talent (25%), and unclear ROI (23%). None of those are technology problems. They are governance, hiring, and measurement problems. The companies scaling AI today solved those three things first and bought the software second.

• Deloitte's 2026 State of AI in the Enterprise report found AI tools are now available to workers at 60% of surveyed organizations, a shift from pilots to real enterprise scale across financial services, healthcare, and manufacturing.
• EY's Canvas platform now processes 1.4 trillion lines of audit data annually across 160,000 global engagements, showing what 'agentic AI at scale' actually looks like in a professional services firm.
• The board question to answer this quarter: do you have a named executive accountable for AI governance, and does that person report to the CEO or get buried in IT?

Looking ahead, the fastest path to AI ROI remains the same as it was a year ago: start with one high-volume, low-risk process, measure it ruthlessly, and let the results fund the next step.

THE NUMBER

$300B

The amount investors poured into AI startups in Q1 2026 alone, surpassing the full-year 2025 total in a single quarter. Four companies ate 65% of it. The other 5,996 startups split the rest.

What else is brewing

• Unilever is deploying AI agents across its entire organization to serve 3.7 billion consumers, which means AI is now closer to your shampoo than your IT department.
• Google internally now generates 75% of all new code with AI, up from just 25% a year ago, and engineers approve it rather than write it, a preview of every technical job description in 2027.
• Merck's earlier internal AI deployment cut clinical study report drafting time from two to three weeks down to three to four days, and cut drafting effort from 180 hours to 80 hours, the clearest before-and-after ROI stat in pharma this year.
• UKG, the HR software giant, laid off 950 employees citing AI this week, and yes, we noticed the irony that an HR platform is automating HR roles.
• Adobe launched a new 'CX Enterprise Coworker,' an AI agent for marketing and customer experience teams that works across AWS, Anthropic, Google, Microsoft, and OpenAI platforms, meaning your marketing team's next hire might not be a human.
• The EU's AI Commission made €63.2 million available this week to support AI innovation in health and online safety, a reminder that Europe is funding AI while also regulating it.
• Snap cut 1,000 jobs this week, with CEO Evan Spiegel citing AI-driven efficiencies, meaning even photo-sharing apps are restructuring around automation.

Written by the Humans In The Loop desk. Sources: TechCrunch, CNBC, Google Cloud Blog, Merck.com, FiercePharma, Crunchbase, KPMG Venture Pulse Q1 2026, Tom's Hardware, Layoffs.fyi, Infor Enterprise AI Adoption Impact Index, Deloitte 2026 State of AI in the Enterprise, EU AI Act Service Desk, Kiteworks, EU Digital Strategy, RD World Online, Bank Info Security, Constellation Research.