AI is scaling. But the economics are getting more complicated. This week in AI: 🧠 Meta enters the frontier-model race 💸 CoreWeave’s debt rattles the AI trade 🤖 Google brings agents into enterprise workflows 🔥 TSMC & ASML signal continued chip demand 📉 Falling token prices give enterprise buyers leverage 🛡️ New developments highlight growing AI security risks The takeaway? Model performance is only part of the equation. Infrastructure economics, security, vendor flexibility, and measurable ROI will shape what comes next. 📖 Read the full AI to ROI News & Analysis: ai2roi.substack.com/p/ai-to-… #AI #EnterpriseAI #AItoROI

AI pacing may buy time. But is it enough? As AI-powered cyberattacks accelerate and autonomous agents expand the attack surface, Ray Rike and Peter Buchanan argue that the conversation needs to go further. Their proposal: treat AI as critical U.S. infrastructure. That means stronger independent testing, incident reporting, security standards, and public-private coordination. The question isn’t just how fast AI should advance. It’s how we protect the systems that increasingly depend on it. 📖 Read the full story on Substack: ai2roi.substack.com/p/ai-pac… #AI #Cybersecurity #AItoROI

AI pilots don’t deliver ROI. Operational change does. On the latest #AItoROI podcast, former Morgan Stanley Head of AI Jeff McMillan joins Ray Rike to discuss what separates AI programs that deliver measurable returns from those stuck in pilot mode. 5 takeaways: • Benchmark before you build • Redesign processes before automating • Redeploy capacity to realize productivity gains • Let Finance validate ROI while business owners own outcomes • Pick 3 high-impact use cases and dedicate your best people to delivery More AI isn’t the goal. Better business outcomes are. 🎧 Listen here, or on your favorite podcast platform: rss.art19.com/episodes/a0b7a… #AItoROI

AI adoption is one thing. Measuring its impact on GTM is another. Scale Venture Partners + Benchmarkit are benchmarking how B2B software companies are using AI across Sales, Marketing & Sales Development. We’re measuring: 🤖 Workflow adoption 🏢 Org design 💰 AI spend & ROI 📊 GTM performance ⚙️ Infrastructure & tech stack Take the survey in <10 minutes and get the 2026 benchmark report, peer comparisons, and early access to findings. Responses are confidential and anonymized. 👉 scalevp.benchmarkit.app/surv…

OpenAI and Anthropic both launched models AND cut prices on the same day. The rest of the week: 💸 GPT-6 Sol/Luna: 50% cheaper than GPT-5.6 🧠 Opus 5.5: first Opus price cut ever. 40% lower cost per task. ⚡ TypeSafe Jev: decision model at 1/100th Claude's cost 🤖 Meta Muse: #1 in both app stores. Zero-day bug. Amazon blocked it. 📊 AI = 8.1% of enterprise software spend (was 1.4% a year ago) 🛡️ Anthropic hires Accenture as safety evaluator for $1B 💻 Google launches the Googlebook at $899 It was a week. 🔗 ai2roi.substack.com/p/ai-to-… #AI #OpenAI #Anthropic #Meta #AItoROI

AI “pacing” may be the newest buzzword in the frontier AI debate. But slowing down model releases isn’t the whole problem. AI is already embedded in critical systems across finance, healthcare, cybersecurity, defense, government, and enterprise operations. Meanwhile, AI-powered attacks are becoming faster and more automated. Our latest AI to ROI Big Story argues that AI should be treated as critical infrastructure — with oversight that extends beyond the labs building frontier models. Pacing may buy time. The bigger question: what do we do with that time? 👉 Read the full story on Substack: ai2roi.substack.com/p/ai-pac…

How does your GTM organization compare with peers? The 2026 B2B GTM Benchmarking Study from Norwest + Benchmarkit is measuring: • GTM scale & investment • Marketing efficiency • Sales productivity • Strategy & operations • AI adoption & ROI New for 2026: AI operating maturity — from experimentation to agent-led execution — plus benchmarks for Sales AI agents and AI-native vs. traditional SaaS companies. 10–15 min. Confidential + anonymized. 👉 Take the survey: benchmarkit.app/survey/Gtm20… Benchmarks publish Nov. 2026.

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