Humans In The Loop -- Monday, April 28, 2026
INTRO
Good morning. While the Dow wobbles and oil flirts with $107 a barrel, the AI industry had a week so busy it needs its own Strait of Hormuz just to let all the news through. Anthropic is eating OpenAI's lunch at the enterprise buffet, regulators are finally showing up with a fork, and somewhere in your office right now an employee is pasting your financial projections into a free AI tool that IT has never heard of. Let's get into it.
IN TODAY'S NEWSLETTER
- Anthropic overtakes OpenAI in enterprise revenue — what it means for your vendor contracts
- The shadow AI time bomb: 3,400 apps your IT team cannot see
- 100,000 tech jobs gone in 2026 — and your board is going to ask about yours
- Regulation roundup: New York, California, and the EU are all moving at once
TOP STORY
Anthropic Is Eating OpenAI's Lunch — And Your Enterprise Contract Is the Plate !
The AI vendor race just flipped. Anthropic, maker of the Claude AI assistant, has overtaken OpenAI in enterprise revenue, hitting a $30 billion annual run rate versus OpenAI's $24–25 billion, and now holds 40% of enterprise AI spending compared to OpenAI's 27%, down from 50% just two years ago. Among companies that newly purchased AI services in March 2026, 65% chose Anthropic over OpenAI.
For a CEO who is not a coder, here is what this means: the safe, brand-name default (OpenAI/ChatGPT) is no longer automatically the right enterprise choice. Anthropic's Claude has a 1-million-token context window — think of it as the ability to read your entire legal contract library in one go — and its safety-first reputation is winning deals in finance, healthcare, and legal, exactly the industries with the highest compliance stakes. If your IT team signed an OpenAI enterprise deal in 2024 and has not revisited it, now is the time.
- Anthropic now has over 1,000 enterprise customers spending $1M+ per year, a number that doubled in just two months.
- Eight of the Fortune 10 are paying Anthropic customers — meaning your largest competitors almost certainly are too.
- OpenAI is pivoting hard toward enterprise with its 'Frontier' platform, serving Oracle, State Farm, and Uber, so the race is far from over.
Looking ahead... Anthropic is eyeing an IPO as early as October 2026, which means its pricing and sales terms could change dramatically once it has public shareholders to satisfy.
--ML
SECURITY
3,400 AI Apps Are Running in Your Company Right Now. You Approved Maybe 10 of Them.
Shadow AI — employees using AI tools that IT never approved — has gone from a quirky HR problem to a full-blown security crisis. Zscaler reports that AI application usage across its customers has exploded to more than 3,400 apps, a quadrupling in just 12 months, with data transfers to those tools topping 18,000 terabytes in 2025. One entertainment company switched on Zscaler's AI traffic monitoring and discovered 4 million AI prompts per week were flowing out of the building.
For non-technical CEOs, the plain-English risk is this: when your employee pastes a client proposal or a patient record into a free AI tool to 'clean it up,' that data may leave your security perimeter permanently. Only 34% of enterprises have AI-specific security controls in place, even as nearly half of cybersecurity professionals rank agentic AI as their top emerging attack vector. A breach at cloud company Vercel this month was traced directly to a compromised third-party AI tool connected to an employee's account.
- Shadow AI is now the most common entry point for enterprise data leakage, according to security firm FireTail.
- The fix is not banning tools — it is publishing a clear approved-AI list and giving employees an enterprise-grade alternative so they stop improvising.
- Cyber insurance carriers are now adding AI Security Riders; companies without documented AI governance may face coverage denials or higher premiums.
Looking ahead... Google and Microsoft both announced shadow AI detection tools this week at Google Cloud Next and RSAC 2026, so your existing cloud vendor may already have a solution you have not turned on.
--ML
WORKFORCE
100,000 Jobs Gone in 2026, and the Layoff Email Is Getting Worse Than the Layoff
Over 100,000 tech workers have been laid off so far in 2026, and nearly half of those cuts are being attributed to AI and automation. Meta announced it would cut 10% of its workforce (about 8,000 jobs), Microsoft trimmed roughly 9,000 roles, and Oracle is in the middle of what could reach 30,000 departures — all while those same companies are collectively planning to spend nearly $700 billion this year building AI infrastructure. Separately, Block CEO Jack Dorsey cut his company nearly in half earlier this year, from 10,000 to fewer than 6,000 employees, calling it the largest AI-attributed layoff in corporate history.
Before you call an emergency board meeting: experts including OpenAI CEO Sam Altman warn that some of this is 'AI washing' — blaming AI for layoffs companies were going to do anyway. But the structural shift is real. Job postings requiring AI skills are up 67% year-over-year, while traditional software engineering roles have fallen 23%. The message for non-tech CEOs is that the talent market is bifurcating fast: generalist roles are shrinking and AI-fluent roles are commanding premiums.
- IBM is actually expanding entry-level hiring, saying human oversight of AI remains essential — a useful counter-narrative for your own recruiting conversations.
- Oracle's layoff was announced via a terse 6 AM email signed 'Oracle Leadership' — research shows companies that handle cuts poorly see 34% higher voluntary attrition among the employees who stayed.
- Salesforce cut 4,000 customer support roles last year; CEO Marc Benioff's rationale was simple: 'I need less heads.'
Looking ahead... Cognizant's Chief AI Officer says real AI-driven productivity gains — and the layoffs that follow them — are still six to twelve months away for most companies, so the wave you are seeing now may be the preview, not the main event.
--ML
REGULATION
Fifty State Laws, One Federal Wish List, and a Brussels Deadline Creeping Closer
The regulatory picture got more complicated this week, not less. New York Governor Hochul signed amendments to the RAISE Act in late March, shifting it toward a transparency-and-reporting framework that mirrors California's law — meaning AI vendors must now disclose training data, deployment protocols, and safety incidents. At the federal level, the Trump White House released a National Policy Framework recommending Congress create a single national standard and override state AI laws that 'impose undue burdens,' but that is a recommendation, not a law, and states are still fully enforceable today.
Across the Atlantic, EU AI Act compliance deadlines are officially in play for companies operating in Europe. High-risk AI rules covering employment, healthcare, and financial services were set to require compliance by August 2026, though the European Commission is negotiating a possible extension to late 2027. Smart companies are not waiting. Meanwhile, Indiana, Utah, and Washington all passed new laws this year specifically prohibiting health insurers from using AI as the sole basis for denying claims — a direct signal to any CFO using AI in benefits administration.
- California's training-data transparency law is already in effect as of January 2026, and enforcement guidance from the attorney general remains thin — meaning you need your own legal team to define compliance.
- A 42-state attorney general coalition is actively hunting AI violations, with enforcement actions against deployers up significantly in 2025.
- The EU's Code of Practice for labeling AI-generated content is expected by June 2026 — relevant for any marketing team using generative AI in customer-facing materials.
Looking ahead... The DOJ's new AI Litigation Task Force has explicit authority to challenge state AI laws it considers unconstitutional, so the federal-vs-state battle could produce rapid, unpredictable changes to your compliance obligations in the back half of 2026.
--ML
VENDOR NEWS
Google Rebuilt Its Entire AI Platform. Here Is What That Actually Means for Your Business.
At Google Cloud Next this week, Google announced the Gemini Enterprise Agent Platform — a complete overhaul of its enterprise AI infrastructure that replaces the old Vertex AI product. In plain English: Google is now treating AI agents the same way it treats cloud servers — as managed, auditable, enterprise-grade workloads with identity controls, security policies, and monitoring built in. Deloitte simultaneously announced it is deploying Gemini Enterprise to 100,000 of its own professionals and has built a library of over 1,000 pre-built industry-specific AI agents for client deployments.
For a CEO who buys software, not builds it, the key insight is that the 'agentic AI' wave — AI that acts, not just answers — is now a real product, not a demo. EY is using it to process 1.4 trillion lines of audit data annually. Salesforce's Agentforce claims to have cut case resolution times by 84% at Reddit. OpenAI this week also published details on its enterprise 'Frontier' platform, which lets companies like State Farm and Uber deploy AI agents that work across all their internal systems, not just one app.
- Adobe rebranded its marketing platform as 'CX Enterprise,' introducing persistent AI agents called 'Coworkers' that run continuously toward business goals without waiting to be prompted.
- OpenAI released GPT-5.5 this week, positioning it as a step toward a unified AI 'super app' that combines research, coding, and browser tools in one interface.
- Infor's new research found that 49% of organizations are still in early-stage AI deployment — meaning most of your competitors have not figured this out yet either, and the window to move first is still open.
Looking ahead... OpenAI's enterprise business now makes up more than 40% of its revenue and is on track to match its consumer business by year-end, which means the product roadmap is going to start looking a lot more like what your ops team needs and a lot less like what teenagers use for homework.
--ML
MONEY & M&A
VC Poured $330 Billion Into AI in One Quarter. Most of It Went to Four Companies.
Q1 2026 broke every venture capital record ever recorded. Global VC investment hit $330.9 billion, more than double Q4 2025's total, driven almost entirely by AI megadeals. Four rounds — OpenAI ($122B), Anthropic ($30.6B), xAI ($20B), and Waymo ($16B) — accounted for $188 billion, or 65% of all global venture investment in the quarter. For context, that single quarter of AI funding exceeded all of 2024's global venture capital.
For non-tech business owners, the M&A message is more actionable. Buyers are done paying for AI promises and now want proof — specifically, net revenue retention above 120% and proprietary data that competitors cannot replicate. Vertical AI companies (those built for specific industries like legal, healthcare, or logistics) are commanding premium acquisition multiples right now. If your business has deep industry-specific data and you are considering a sale in the next two years, AI-native positioning could significantly change your valuation story.
- AI accounted for 80% of all global venture funding in Q1 2026 — up from 55% in Q1 2025.
- Legal AI firm Legora raised $550 million at a $5.55 billion valuation; medical AI platform OpenEvidence raised $250 million at a $12 billion valuation — vertical AI is having a moment.
- Gartner forecasts global AI spending will reach $2.52 trillion in 2026, a 44% increase year-over-year, suggesting the market is nowhere near saturation.
Looking ahead... With cross-border M&A up 47% from a year ago and AI driving four of the six biggest deals this quarter, your corporate development team should be asking right now whether there is a vertical AI acquisition that could accelerate your own roadmap before the multiples get any higher.
--ML
STAT: PILOT PURGATORY
49%
Nearly half of organizations globally are still in early-stage AI deployment, stuck in pilots and partial rollouts, according to a new Infor survey of 1,000 business decision-makers. The top three reasons: data security concerns (36%), no internal AI talent (25%), and unclear ROI (23%) — which means the barrier is almost never the technology.
WHAT ELSE IS BREWING
- OpenAI's enterprise business now tops 40% of its total revenue and is on pace to match consumer revenue by year-end — Sam Altman's pitch to your CIO just got a lot more aggressive.
- Indiana, Utah, and Washington passed laws banning health insurers from using AI as the sole basis for denying claims — if you run a benefits program, your legal team needs to read these now.
- C.H. Robinson cut about 1,400 logistics jobs after rolling out AI-driven pricing, scheduling, and shipment tracking tools — the freight industry's AI transformation is no longer theoretical.
- IBM is bucking the trend and tripling entry-level hiring in 2026, arguing that AI creates more need for human oversight, not less — worth citing the next time your board asks whether to freeze headcount.
- Automation Anywhere says AI agents now auto-resolve over 80% of IT support requests, cutting IT service management costs by up to 50% — and deployment takes as little as 8 weeks.
- Ocado, the UK's largest online grocer, cut 1,000 jobs in March after completing a 'very significant phase of investment' in robotics and automation — grocery and retail are not immune.
- Anthropic is fighting the Pentagon in court after the U.S. Department of Defense put Claude on its 'supply chain risk' list — federal contractors using Claude should monitor this closely.
- EY's AI platform now processes 1.4 trillion lines of audit data annually across 160,000 global engagements — if your auditors are not using AI, you may want to ask why you are paying them like they are.
Written by the Humans In The Loop desk. Sources: Crunchbase, KPMG Venture Pulse Q1 2026, Cooley Law State AI Laws Alert (April 24 2026), Zscaler AI Security Report, FireTail Security Boulevard (April 2026), CNBC (April 24 2026), Tom's Hardware layoffs tracker, AIAgentStore.ai, Google Cloud Next press releases, OpenAI Enterprise blog, Infor Enterprise AI Adoption Impact Index, FifthRow April 2026 Playbook, MarketingProfs AI Update (April 24 2026), Yahoo Finance / Motley Fool (April 27 2026), TradingEconomics, Sunday Guardian Live markets (April 27 2026), PANews / TradingKey Anthropic vs OpenAI analysis, MedCityNews clinical supply chain (April 2026), GHX Healthcare Supply Chain Predictions 2026, ICAEW / DRCF Agentic AI Foresight Paper (April 28 2026), Holland & Knight AI Regulation alert, GlobalPolicyWatch Q1 2026 tech regulatory update.