In partnership with
🌟 Editor's Note: Recapping the AI landscape from 02/24/26 - 03/01/26.
🎇✅ Welcoming Thoughts
Welcome to the 33rd edition of NoahonAI.
What’s included: company moves, a weekly winner, AI industry impacts, practical use cases, and more.
This may be the last “NoahonAI” newsletter.
🚨 I’m changing my company name to KPAI Solutions. Keep on eye out for KPAI Newsletter next Tuesday or the following.
Not much will change but the name.
I’m going to be revisiting some old Practical Use Cases that may have been forgotten along the way but remain relevant. See below.
Lots going on with Anthropic and the DoW this week.
Here was the statement from Anthropic CEO Dario Amodei.
Just an interview highlight this week, but an interesting one nonetheless.
Claude reached #1 on the App store.
Meta may go without AI customers but they will never go without compute.
Anthropic dropped some free courses on learning their tools.
A shocker in 2nd place this week.
Claude’s been down a bit today, I’ve had to resort back to GPT and Gemini.
Let’s get started—plenty to cover this week.
👑 This Week’s Winner: OpenAI // ChatGPT
OpenAI is Back On Top. Their week included what could be their final (private) fundraise, continued growth signals, and a contract that may land them in some hot water publicly, but will be a strong positioning move over time. Here’s the Recap:
$110B Mega Round: OpenAI announced a $110B fundraise at a reported $730B pre-money valuation. Amazon is said to be committing $50B (with part structured around future conditions), alongside large checks from NVIDIA and SoftBank. As I said above, this should be their last raise until IPO.
900M Weekly Users: ChatGPT crossed roughly 900M weekly users and 50M paying subscribers. OpenAI continues to push into all sectors with a strong focus on consumer and enterprise. 1B weekly users on the horizon.
OpenAI Over Anthropic? OpenAI won the Department of War bid after Anthropic backed out due to disagreements on usage. The deal is said to be similar except discretion falls on the DoW not OpenAI. Lots of negative publicity here among some consumers but it will die down.
It doesn’t stop there: OpenAI published a new threat report detailing China-linked influence activity and banned related accounts tied to covert operations. Also, on the monetization side, Criteo joined OpenAI’s early ChatGPT advertising pilot, signaling the ads program is continuing to come into place.
From Top to Bottom: Open AI, Google Gemini, xAI, Meta AI, Anthropic, NVIDIA.
⬇️ The Rest of the Field
Who’s moving, who’s stalling, and who’s climbing: Ordered by production this week.
🔵 Meta // Meta AI
Expanded AMD buildout: Meta and AMD announced a major long-term AI infrastructure deal, targeting up to 6 gigawatts of capacity starting in late 2026. Analysts say it could be worth up to ~$100B. Diversified compute resources outside of NVIDIA.
Renting Google TPUs: Meta reportedly signed a multi-billion dollar deal to rent Google’s AI chips for training. It has also reportedly explored buying TPUs outright. This is about renting compute capacity for short term flexibility where AMD and NVIDIA are owned long-term buildouts.
Meta AI Shopping Test: Meta is testing a shopping feature in its AI assistant that shows product results and quick-buy options, signaling a bigger push into AI-driven commerce. Can see this being a big AI angle for Meta, especially given FB marketplace usage.
⚪️ NVIDIA
Excellent Earnings: NVIDIA reported $68B in Q4 revenue (up 73% year over year) and $216B for the full year, with data center sales jumping 75%. AI demand is still driving huge growth. No surprise - priced in.
Investing in Chip Supply Chain: NVIDIA put $2B each into Lumentum and Coherent to secure the laser and optics parts needed for next-gen AI data centers, helping ensure it can keep scaling.
Moving Into 6G Networks: NVIDIA is teaming up with telecom companies to help build the next generation of wireless networks (6G) using AI-driven infrastructure. This is interesting. NVIDIA is going to have the prints on a LOT of infrastructure when its all said and done.
🟠 Anthropic // Claude
Pentagon Dispute + Federal Ban: Anthropic refused to allow use of Claude for domestic surveillance or autonomous weapons, President Trump ordered federal agencies to phase out Anthropic, and DSec Pete Hegseth designated the company a “supply-chain risk,” restricting military contractors from working with it. Anthropic says it will challenge the move. This has been boiling up for a few weeks now. I’d probably side with Anthropic on this one.
Cowork Upgrade: Expanded Cowork with plugin marketplaces, deeper tool integrations (Google Workspace, DocuSign, Salesforce, etc.), stronger admin controls, and cross-app workflows. Nice, would like to see Cowork get less buggy when running, but great add-on. Enterprise focused.
Vercept Acquisition: Anthropic acquired Vercept to help agents operate inside real applications. Vercept’s standalone product is being wound down as the tech folds into Claude. Great move. Includes tech that can remote operate Macs and more. Will strengthen Claude agentic features - think Cowork.
🟣 Google // Gemini
Nano Banana 2: Google launched its new image model in the Gemini app and Workspace, promising faster, higher-quality image generation and editing. Nice. Best image model gets better.
Galaxy S26 Agentic Tasks: Samsung’s S26 integrates Gemini for multi-step, cross-app tasks via a side-button trigger, rolling out first in the U.S. and Korea. This is cool, mobile agentic AI adoption is way behind PC as it stands now. This could be an early step in changing that.
Intrinsic Folded Into Google: Alphabet moved its robotics unit Intrinsic into Google, reinforcing a push toward a standardized “Android for robots” platform. Building the software behind the latest robotics push.
🔴 xAI // Grok
Debt Cleanup Ahead of IPO: The X/xAI debt stack (~$17.5B) is being refinanced, including plans to buy back $3B in bonds early. The move simplifies the cap table and improves credit optics ahead of potential IPO. Makes sense. Who’s gonna IPO first (Anthropic, X, OpenAI).
Trade-Secrets Lawsuit Dismissed: A federal judge dismissed xAI’s trade-secrets case against OpenAI for insufficient allegations, though xAI can amend and refile. Good, seemed a distraction for all involved.
Another Co-Founder Exit: Co-founder Tony Wu resigned, adding to a string of senior departures at xAI. Could be related to recent controversies or could just be looking for a new venture.
🤖 Impact Industries 📢
Robotics // Airbnb Robots
Faraday Future completed its first commercial delivery of embodied AI robots to Golden Hills Investment, a Florida-based luxury vacation rental operator. This deployment represents the first commercial use of such robotics in the U.S. hospitality sector. The company plans to ship 200 units during the 2026 season, targeting premium hotels and dealerships to automate guest services. This move positions the firm as a pioneer in integrating advanced, embodied intelligence within the shared-living and hospitality industries.
Read the Story
Advertising // Agentic Campaigns
Amazon Ads launched Creative Agent, a conversational AI tool designed to manage the entire advertising workflow. Within Creative Studio, the tool handles research, scriptwriting, video animation, and voiceover for various formats. By leveraging retail insights and shopping signals, Creative Agent provides professional-quality campaign development at no additional cost. Initially launching across the UK and major European markets, this platform takes high-end advertising capabilities from initial concept to final delivery for any brand.
Read the Story
💻 Interview Highlight: Dwarkesh with Dario Amodei
Interview Outline: The interview focuses on the final stages of the AI exponential, moving beyond language modeling into reinforcement learning (RL) and "world models". Amodei explains his "Big Blob of Compute" hypothesis—where raw scale outweighs clever algorithms—and predicts a "country of geniuses in a data center" by 2026 or 2027. He also discusses the "diffusion lag" that will delay AI’s global economic impact despite its rapid technical progress.
About the Interviewee: Dario Amodei is the CEO and co-founder of Anthropic, an AI safety and research company. Formerly the VP of Research at OpenAI, he led the teams responsible for GPT-2 and GPT-3 and has been a central figure in defining the scaling laws that govern modern large language models.
Interesting Quote: ”I’m ~90% that within ten years we get a ‘country of geniuses in a data center,’ but for verifiable tasks like coding I think we could be there in one to two years (maybe one to three).”
Condensed Interview Highlight — Dario Amodei on AI Risk, Scale & the Future of Work
Dwarkesh: What exactly is the scaling hypothesis at this point?
Dario Amodei: It is the "Big Blob of Compute Hypothesis." The core idea is that clever algorithms matter far less than raw scale. Only a few things truly drive progress: massive compute, vast data quantity and quality, and an objective function that can scale to the moon.
Dwarkesh: How does viewing biology as "information" change the way we approach medicine?
Dario Amodei: We need to stop treating biology as just matter and start seeing it as a computable information-processing system. AI will be the "starting gun" for curing most diseases because breakthroughs are currently bottlenecked by intelligence. Once we have a "country of geniuses" in a data center, we can radically accelerate drug discovery.
Dwarkesh: Why do you think AGI is a few years away rather than a decade?
Dario Amodei: I am 90% certain we reach this level by 2035. It is hard to be more certain because of "irreducible uncertainty." Geopolitical turmoil or the destruction of chip fabs could delay us by ten years, but on a sane timeline, it is crazy to think this won't happen by 2035.
Dwarkesh: What has been the most surprising thing about AI progress recently?
Dario Amodei: The lack of public awareness regarding how close we are to the end of the exponential growth curve. It is wild to me that people are still focused on old, tired political debates when we are standing at the edge of a world-changing technological shift.
Dwarkesh: Why is it so critical that democracies win the AI race?
Dario Amodei: AI could create a self-fulfilling cycle of authoritarian power that is nearly impossible to displace. I want democratic nations to hold the strongest hand so that the rules of the road for the post-AI world are based on liberal values rather than high-tech totalitarianism.
👨💻 Practical Use Case (Issue 06 Revisited): Claude Code
Difficulty: Advanced
Last week I was raving about Claude Code, and now that I’ve used it some more, I’d like to share what I’ve learned. In terms of “Vibe Coding” → coding without actually writing code, Claude Code is far and away the best solution on the open market. Claude Code is not an app or website, rather it’s what's called a CLI: Command Line Interface. A tool that can be downloaded and implemented directly in your PC’s Terminal. Here’s what it looks like:
Claude Code, like most CLIs, is excellent for frontend work (websites, UI/UX) and solid on backend projects like APIs, data pipelines, and server logic. The more complex the build, the more SWE fundamentals you’ll need, but for basic or mid-level projects you don’t need to know much. Since Claude Code is essentially at the forefront of “vibe coding,” I tend to look at what people I respect say about it to shape how I use CC. Here’s what YC founder Paul Graham had to say:
Anyways, even if you have never written a line of code in your life, here’s how you can build your first project with Claude Code (Disclaimer: It does cost $20/month).
Locate and open your terminal app on your PC.
Go to the Claude website and sign up for Claude, choose the pro plan.
Go to the Anthropic website and follow the steps to install node.js (just a download), and paste the line of code shown on the site into your terminal window and hit enter.
Once pasted into your terminal, you should see a popup, and click Enter/Return to indicate “Yes”.
You’re all set! Talk to your terminal like you would ChatGPT. Tell it anything you want like “I want to build a website about Penguins”.
Watch it create to-do lists, ask for approval to continue, and work its magic!
If you like building, or you want to improve on some of your workflows, i’d highly suggest giving Claude Code a try!
Issue 33 Update: Since writing this around 6 months ago, Claude Code has only gotten better. It’s introduced new features like sub-agents writing code in parallel, new project memory features, checkpoints, and more. If you haven’t tried Claude Code yet, there’s never been a better time than now!
💻 Startup Spotlight
Gushwork AI
Gushwork — Mastering "Answer Engine Optimization" (AEO) for SMBs.
The Problem: Traditional SEO is becoming less important over time. Buyers are no longer just scrolling through Google links; they are asking ChatGPT, Perplexity, and Gemini for recommendations. If your business isn't "cited" or mentioned in those AI-generated answers, you instantly lose reach in the new discovery landscape.
The Solution: Gushwork uses a "swarm" of AI marketing agents to handle what they call Answer Engine Optimization (AEO). Their platform automatically generates search-optimized content, builds high-quality backlinks, and monitors lead generation to ensure a company’s website is the one being analyzed and recommended by AI models.
The Backstory: Founded in 2023 by Nayrhit Bhattacharya and Adithya Venkatesh, Gushwork originally focused on general workflow outsourcing before pivoting hard into AI-led marketing. They just closed a $9 million seed round in February 2026 led by Lightspeed and SIG, valuing the company at $33 million after seeing explosive demand from B2B service providers.
My Thoughts: I was actually considering starting a consultancy strictly focused on GEO (Generative Engine Optimization), same thing as what these guys are calling AEO, but I decided to go in my current direction instead. I haven’t used this product yet, I set up a few custom autos to do something similar, but it’s definitely an important aspect of operating a business in the age of AI.
“It’s not likely you’ll lose a job to AI. You’re going to lose the job to somebody who uses AI”
- Jensen Huang | NVIDIA CEO
Dario Amodei and Pete Hegseth did not agree, but what do you think: Should AI Agents be in charge of life and death decisions in a war? Till Next Time,
Noah on AI