AI in Business: a handful print money, the rest debate ROI May 5th 2026

Curated by Alexis

The cloud hyperscalers just handed in their report cards, and AI is the only subject that matters. Google Cloud hit $20 billion in quarterly revenue — a 63% jump year-over-year — while Azure grew 40% and Microsoft's annualized AI revenue reached $37 billion, up 123%. The private markets matched that energy: Sierra, Bret Taylor's enterprise AI agent company, closed a $950 million Series E at a $15.8 billion valuation on Monday. And running beneath all the euphoria, a landmark PwC study confirmed what skeptics have long suspected — three-quarters of AI's economic gains are flowing to just one-fifth of companies, and the gap is actively widening.

Watch & Listen First

Strategy Summit 2026: Who's Going to Succeed with AI? — Harvard Business Review's April episode with MIT principal research scientist Andrew McAfee on why most organizations are measuring the wrong things and what separates leaders from the rest. Essential listening before your next board-level AI budget conversation.

From AI Strategy to AI Results: How Enterprises Finally Break Through — A practitioner-level breakdown of why AI pilots stall at production and what organizational changes actually close the gap. Specifically useful if you're navigating the pilot-to-scale transition right now.

Key Takeaways

The hyperscalers are the cleanest AI revenue story right now. Google Cloud's 63% growth (backlog: $460B) and Azure's 40% growth with $627B in committed obligations are the clearest market signal that enterprise AI spending is durable, not discretionary.
Enterprise AI agents are the next platform category. Sierra's $15.8B valuation — built on $150M ARR in just eight quarters, with Fortune 50 customers — is a direct claim that AI agents will consume the $400B customer service market whole.
74% of AI's economic value is going to 20% of companies. PwC's global study shows the differentiator is using AI for revenue growth and industry convergence, not operational efficiency — a mandate most AI programs are still getting wrong.
Legal AI has become a serious enterprise battleground. Legora ($5.6B, backed by Nvidia) and Harvey ($11B, backed by Sequoia) are now running celebrity ad campaigns competing for the same accounts — this vertical moved from niche to mainstream faster than anyone predicted.
Palantir's U.S. commercial revenue grew 104% YoY. The company most associated with government contracts is now the earnings benchmark that makes enterprise CFOs reconsider their own AI timelines.

The Big Story

Sierra's $950M Round Is a Direct Bet That AI Agents Eat Customer Service · May 4, 2026 · TechCrunch

→ Bret Taylor's Sierra closed a $950 million Series E at a $15.8 billion post-money valuation — up 58% from its $10 billion mark last fall — led by Tiger Global and Google's GV, with Benchmark and Sequoia also participating. The figure that actually matters isn't the valuation: it's the $150 million in ARR Sierra crossed in just eight quarters, with customers including Prudential, Cigna, Blue Cross Blue Shield, Rocket Mortgage, and one in three of the world's largest banks. Taylor's core thesis is that $400 billion is spent globally on customer service annually, and AI agents will absorb the majority of it — a claim his revenue trajectory is beginning to validate empirically. The competitive implication is blunt: Salesforce, ServiceNow, and every legacy CRM vendor now face a company that already has half the Fortune 50 running production deployments. Sierra's ARR sprint from $100M in November to $150M by February represents the kind of velocity that historically resets category valuations and forces incumbents to make defensive acquisitions — watch for M&A interest from the usual suspects before year-end.

Also This Week

Google Cloud Tops $20B Quarterly Revenue for the First Time as AI Demand Strains Capacity · April 29, 2026 · TechCrunch
→ Google Cloud's 63% growth — outpacing Azure (40%) and AWS (28%) — was explicitly described as capacity-constrained, meaning demand is running ahead of what Google can provision; for enterprise buyers in multi-year cloud negotiations, that supply-side ceiling is a risk worth pricing into contract terms today.

Palantir Reports 85% Revenue Growth, U.S. Commercial Business Surges 104% Year-Over-Year · May 4, 2026 · CNBC
→ Palantir's $1.63B quarter and full-year guidance raised to 71% growth — against a $7.27B consensus — signals that operational AI embedded in defense, logistics, and finance workflows is the first segment where AI generates measurable, recurring returns at scale rather than productivity anecdotes.

Legal AI Startup Legora Hits $5.6B Valuation with Nvidia and Atlassian Backing · April 30, 2026 · TechCrunch
→ Nvidia's NVentures investment in Legora — described as its first legal AI bet — confirms that the chip giant is systematically building equity stakes in every enterprise vertical where its compute runs, a quiet platform strategy that law firm GCs should interpret as a long-term infrastructure dependency, not a neutral technology endorsement.

Microsoft's Annualized AI Revenue Reaches $37B, Up 123%, as Azure Growth Hits 40% · April 29, 2026 · Microsoft
→ With commercial remaining performance obligations up 99% to $627 billion, Microsoft has effectively pre-sold its AI roadmap for years ahead — any enterprise still in EA negotiations should treat that backlog figure as a signal about Microsoft's leverage position at renewal time.

Anthropic Revenue Reportedly Surpasses OpenAI, October IPO Targeted at ~$380B · May 2026 · Trading Key
→ If the annualized revenue figure holds — Anthropic at $30B+ versus OpenAI's $25B — it reframes the competitive narrative entirely and would make Anthropic's public debut one of the most consequential pricing events for AI valuations in 2026.

From the Lab

Artificial Intelligence, Productivity, and the Workforce: Evidence from Corporate Executives · NBER / Federal Reserve Bank of Atlanta
→ A survey of nearly 750 CFOs found AI-driven labor productivity gains are real and concentrated in high-skill services and finance — but the study documents a critical "productivity paradox": perceived gains are significantly larger than measured gains, suggesting companies are claiming AI wins they cannot yet prove in the revenue line. Aggregate employment is expected to fall less than 0.4% in 2026, but large firms expect to shed headcount while small firms expect modest growth, a divergence that will shape labor market dynamics unevenly across industries. For operators: the gap between perception and measurement is precisely where board scrutiny should focus before the next AI investment request lands.

PwC 2026 Global AI Performance Study: 20% of Companies Capture 74% of AI's Economic Value · PwC
→ Across thousands of companies globally, PwC found that top AI performers generate 7.2x more AI-driven value than the average competitor, are 1.9x more likely to run autonomous AI operations, and are increasing decisions made without human intervention at 2.8x the rate of peers — the single strongest differentiating factor is using AI to capture growth opportunities, not to cut costs, which inverts how most enterprise AI programs are currently framed and funded.

Worth Reading

Sierra raises $950M as the race to own enterprise AI gets serious — TechCrunch's full breakdown of Sierra's growth metrics and what they mean for Salesforce and ServiceNow's competitive position.
Google Cloud surpasses $20B, but says growth was capacity-constrained — The supply-constraint angle is underreported in the earnings coverage and matters significantly for anyone mid-negotiation on cloud infrastructure.
Legal AI startup Legora hits $5.6B valuation and its battle with Harvey just got hotter — The Jude Law vs. Gabriel Macht celebrity ad war is a useful leading indicator of how fast a B2B AI vertical can move from niche to category-defining.

The earnings data is in: AI is generating real returns for hyperscalers, well-positioned operators, and the companies that moved earliest — every executive still waiting for certainty is now paying a compounding opportunity cost.

Sources:

Get more from AI Weekly

More signal, less noise — pick your channels.