AI safety, security, infrastructure, and enterprise adoption all collided this week. The latest AI to ROI News & Analysis covers: 🤖 The AI pacing debate — and Dario Amodei’s call for slower frontier development 🥷 Anthropic’s threat report on increasingly sophisticated AI misuse 🚨 OpenAI’s latest security + misalignment incidents 🤝 Salesforce + NVIDIA’s domain-specific CRM model, Koa 🏗️ Oracle’s massive AI infrastructure bet 💼 Anthropic merging Cowork into Claude 💰 OpenAI + Anthropic targeting financial services 🏢 Cohere + Aleph Alpha completing their merger 📜 Microsoft’s new AI code of conduct + kill switches The bigger question: As AI becomes more autonomous, who controls it, who pays for it, and who is accountable for the outcome? Read the full analysis: ai2roi.substack.com/p/ai-to-…

Every AI investment conversation should start with one question: What KPI are we trying to improve? Not which model. Not how many tokens. Not how quickly we can launch a pilot. Ketan Karkhanis, CEO of @thoughtspot, joins Ray Rike on the latest #AItoROI podcast to explain why the path from AI aspiration to ROI starts with the business outcome. They cover: → Why “AI saves you 2 hours a day” isn’t a real value case → Why pilots are where value goes to die → Functional KPIs as the bridge to revenue + margin → AI cost + gross margin challenges → Pricing CFOs can actually forecast → Customer outcomes + C-suite ownership The goal isn’t to prove AI works. It’s to prove what it changes. Listen here, or on your favorite podcast platform: rss.art19.com/episodes/8d917…

85% of enterprises are working on agentic AI. Only 5% have put an agent into production that delivers meaningful ROI. The gap isn’t just model capability. It’s everything around the model. The 6 challenges: Reliability Data + legacy integration Governance + agent identity Cost visibility + ROI Platform fragmentation Workforce trust + org redesign And they compound. Bad data → unreliable agents → higher costs → governance risk → fragmented platforms → lower workforce trust. The enterprises making agentic AI work are treating it as an operating model + infrastructure problem. Workforce Buy-In + Improved Productivity = High ROI Agentic AI Read the full article now: ai2roi.substack.com/p/the-si…

Your AI product may be growing fast while quietly eroding your gross margin. AI-Adjusted Gross Margin makes the economics visible. Track: • Inference • Model hosting • GPU serving • Vector DBs • RAG + orchestration • AI observability A $500K/month product with $175K total COGS = 65% margin. Cut AI COGS by $25K → 70% margin without adding revenue. The metric that connects AI consumption to the financial model. Read the full breakdown on Substack: ai2roi.substack.com/p/ai-to-…

Every AI conversation should be an ROI conversation. Not: Which model should we deploy? Start with: Which KPI are we trying to improve? Ketan Karkhanis, CEO of @thoughtspot, joins Ray Rike on AI to ROI to discuss: • Why pilots are where value goes to die • Functional KPIs → financial impact • Token anxiety + gross margins • Pricing CFOs can forecast • Making customer outcomes the ultimate AI metric Listen here, or on your favorite podcast platform: aitoroi.ai/podcast/why-every…

Last call for the 2026 B2B GTM benchmarks. Sales, Marketing, RevOps & Finance leaders: this is your final chance to contribute your data and benchmark against your peers. 10–15 minutes. Confidential + anonymized. New this year: AI operating maturity, Sales AI agents & AI-native companies. 👉 TAKE THE SURVEY benchmarkit.app/survey/Gtm20…

Where are the killer AI-native application companies? They’re here. But they’re getting squeezed from every direction. AI model companies are moving up the stack. SaaS incumbents own customers, data & systems of record. Agentic platforms are capturing workflows. Coding agents make build-vs-buy more compelling. And AI-native apps have another problem: weaker unit economics from ongoing inference costs. So what separates the winners? Harvey. Abridge. EvenUp. Fieldguide. Four common traits: → Start with a painful, expensive workflow → Accumulate proprietary context → Expand from a task into a system → Sell measurable economic outcomes The AI-native application opportunity is enormous. But the winners will be the companies that control enough workflow, data, trust, and customer economics to keep the model underneath them from becoming the product above them. Read the latest AI to ROI Big Story ↓ ai2roi.substack.com/p/where-… #AI #AItoROI #EnterpriseAI

Most enterprises can see their AI bill. Far fewer can predict it. @Benchmarkitai + Mavvrik surveyed 396 companies and found: → 98% track AI costs → Only 11% forecast within 10% of budget → 54% miss forecasts by >26% → 47% cite data/platform costs as the top source of budget misses → Only 36% can attribute agentic costs by agent → 25% have halted an AI initiative due to unexpected costs The problem gets harder with agentic AI, where non-deterministic workloads and retry loops make spend harder to predict. The takeaway: visibility first. Economic value second. Then scale. Ray Rike and Peter Buchanan unpack the research in the latest AI to ROI episode. Listen/read the full research ↓ aitoroi.ai/podcast/enterpris… #AI #AItoROI #FinOps #EnterpriseAI

AI adoption is accelerating. AI’s financial impact? Not so much. McKinsey found: → 44% of orgs are scaling AI, up from 38% → 40% of $1B+ companies are scaling AI agents → 32% have skipped buying software because they can build it with agents → 60% expect to increase AI investment → Only 37% report any AI-attributed EBIT impact Meanwhile, OpenAI, Anthropic, Meta, Google, Salesforce, Snowflake & Palo Alto are pushing the market forward at full speed. The takeaway: AI adoption is becoming table stakes. Turning it into measurable ROI is the real challenge. Read this week’s AI to ROI News & Analysis ↓ ai2roi.substack.com/p/ai-to-… #AI #AItoROI #EnterpriseAI

How does your GTM organization stack up against your peers? Norwest + Benchmarkit are building the 2026 B2B GTM Benchmarking Study — covering: → GTM investment & headcount → Marketing efficiency → Sales productivity → Strategy & operations → AI adoption & ROI → AI operating maturity → Sales AI agents → AI-native vs. traditional SaaS The survey takes 10–15 minutes, and responses are fully anonymized. Help shape the benchmarks B2B leaders will use to compare their GTM performance in 2026. Take the survey ↓ benchmarkit.app/survey/Gtm20… #B2B #GTM #AI #SaaS

Stripe is betting $7.5B on the AI routing layer. OpenRouter went from a $1.3B valuation to a $7.5B acquisition in just 85 days. Why? Routing is becoming the control point for enterprise AI spend. 8M developers. 400+ models. 80+ providers. ~$160M annualized revenue. And the real moat may be the routing data. The question for CFOs: Is your AI gateway an honest broker? Listen to the full #AItoROI episode discussion here: rss.art19.com/episodes/bd380…

Stripe just agreed to pay $7.5B for OpenRouter. Why? OpenRouter sits between developers and 400+ AI models, routing workloads based on cost, speed, availability, and performance. In other words: the gateway to AI inference. Stripe sees a familiar playbook: Payments → sit between businesses & banks AI → sit between developers & model providers The bigger bet? AI may not have one winning model. It may need a neutral layer that routes between many of them. But there’s a catch: OpenRouter’s value comes partly from being neutral. If Stripe starts steering customers toward its own commercial interests, that advantage could disappear. Our latest analysis looks at the $7.5B deal, the AI gateway land grab, and who could ultimately control the infrastructure between AI applications and models. Read the full analysis → ai2roi.substack.com/p/stripe… #AI #AIInfrastructure #FinOps

AI adoption is no longer the interesting question. Has AI actually become part of how your GTM organization operates? @NorwestVP + @Benchmarkitai are conducting the 2026 B2B GTM Benchmarking Study to find out. We're benchmarking: 📊 GTM investment 📈 Marketing efficiency 🎯 Sales productivity ⚙️ GTM strategy & ops 🤖 AI adoption + ROI 🧠 NEW: AI operating maturity This year includes a 5-stage AI maturity model, from experimentation to agent-led execution, plus benchmarks for Sales AI agents and deployment stage. Results will be segmented by revenue, ACV, pricing model, and product profile — including AI-native vs. traditional SaaS. ⏱️ Takes 10–15 minutes. 🔒 100% confidential + reported only in aggregate. Help build the benchmark for how B2B GTM teams are actually operating with AI. 👉 Take the survey: benchmarkit.app/survey/Gtm20… #B2B #SaaS #GTM #AI

AI adoption is scaling faster than companies can govern the bill. Only 11% of companies can forecast AI spend within ±10% in 2026, down from 15% last year. And: 💸 49% modified pricing due to unexpected AI costs 🛑 25% stalled/shelved an AI project 📊 98% track AI costs, but only ~5% can allocate them to the customer, app, or agent level So how do we move from AI Cost Governance → AI Cost Optimization? Ray Rike sits down with Sundeep Goel, CEO of Mavvrik, on the latest AI to ROI Podcast to discuss: • AI cost forecasting • Cost allocation vs. tracking • The breakdown of per-seat pricing • AI waste + optimization • Dynamic model selection • Prompt caching • Measuring AI ROI Cost is measurable. Value is harder. 🎧 Listen to the episode: rss.art19.com/episodes/b6eee… Subscribe and leave us a ⭐⭐⭐⭐⭐ if you're enjoying AI to ROI. #AI #AIROI #FinOps

💰 @stripe just agreed to pay $7.5B for @OpenRouter Why spend that much on an AI gateway? Because the future of enterprise AI may not belong to a single model. It may belong to whoever controls how workloads move between them. OpenRouter gives developers access to 400+ models and 80+ providers through one API, with routing based on cost, speed, availability, and performance. And the competition is already heating up: Ramp, Cursor, Snowflake, Databricks, AWS, Microsoft Azure + dozens of startups are building their own gateways. The bigger prize? The data generated by all that routing. Which models win. For which workloads. At what cost. Under what conditions. That could make the AI gateway the control layer for enterprise AI. Ray Rike and Peter Buchanan break down the deal and what it means for the AI infrastructure market. 👉 Read the full story: ai2roi.substack.com/p/stripe… #AI #FinOps #EnterpriseAI

Kimi K3 drops. Headlines say Chinese AI caught the US labs. The AI trade sells off. Nobody asked: does cheaper = more profitable? Ray Rike & Peter Buchanan ran the actual math. The price war math: • Frontier models: $6-8/M tokens to serve • Anthropic Opus at $25/M = 70% gross margin • Reprice to $4-6 (Meta Muse Spark range) = -65% margin The metric that matters: cost per COMPLETED task • Claude Opus: 90% completion, $2.56/task • Muse Spark: 75% completion, $5.80/task • Token price: 75% cheaper. Delivered cost: 2x more. Ray: Closed-weight labs still win on total cost of ownership. Peter: Hyperscalers are the quiet winners either way. 🎧 rss.art19.com/episodes/3337c… #AI #Anthropic #OpenAI #AIEconomics #AItoROI

AI is moving fast. The economics underneath it are moving even faster. This week in AI to ROI: 💰 NVIDIA + Wall Street: $500B AI financing push 📈 CoreWeave: 2x growth, mounting losses ⚡ NVIDIA: cheaper AI routing for agents 🚀 SpaceXAI: Grok 4.6 + Grok Bot 🔧 AMD + Taalas: inference economics 💧 Anthropic: AI watermarking 🛡️ OpenAI: new cyber model 🌊 Open-weight AI keeps gaining momentum The bigger question: Where does the ROI actually accrue — and who is paying for it? Ray Rike and Peter Buchanan break down the biggest AI signals of the week and what they mean for enterprise leaders. Read the full issue → ai2roi.substack.com/p/ai-to-… #AI #AIROI #EnterpriseAI

Think your brand gets 20–50% of its traffic from LLMs? One $3B company thought so. The measured answer: 1%. On the latest AI to ROI podcast, Ray Rike and Stephan Bajaio of VibeLogic discuss Web Presence Intelligence, AI marketing measurement, attribution, baselines, control groups, and where the real ROI may be hiding. 🎙️ Listen now: rss.art19.com/episodes/84596…

Can you answer one question: What is your AI ROI? “Saved time” and “increased productivity” aren’t enough anymore. The Definitive Guide to AI Metrics gives you a framework to connect AI investments to business outcomes. 📊 81 metrics 🏢 13 business functions 📈 3 reporting tiers Measure AI to ROI—not just AI activity. Download the Big Book of AI Metrics: benchmarkit.ai/ai-big-book-o… #AI #AIROI #AIMetrics