In partnership with
In Today’s Issue:
🤑 OpenAI drops a $100 monthly Pro tier for developers
💵 OpenAI projects a massive leap to $100 billion in ad revenue by 2030
📈 MedGemma achieves stunning accuracy jumps in 3D medical imaging and pathology
📉 The AI infrastructure boom hits a wall of severe supply chain delays
✨ And more AI goodness…
Dear Readers,
OpenAI just dropped a $100 monthly Pro tier right into the heart of the AI coding wars, and the timing tells you everything about where this industry is headed. With Anthropic quietly surpassing OpenAI in annualized revenue and locking down its ecosystem, this isn't just a pricing move, it's a declaration of war over the developer wallet.
But the battle for AI supremacy isn't only playing out in subscription tiers. Today we're also looking at why nearly half of all planned U.S. data centers are stalling out while Iran literally threatens to bomb the ones being built in the Gulf, why OpenAI thinks it can build a $100 billion ad business inside a chatbot, how Google's new 4B medical model is reading 3D brain scans without breaking a sweat, and why a restricted cybersecurity AI model has the industry on edge.
Plus, an AI-generated cow from 2014 that will make you appreciate just how far we've come. Let's get into it.
All the best,
🚨 AI Cyber Models Raise Alarms
OpenAI is reportedly preparing a restricted rollout of a powerful new cybersecurity-focused AI model, mirroring Anthropic’s limited release of its Mythos system due to fears of advanced hacking capabilities. The move highlights a turning point where AI tools can autonomously discover vulnerabilities or potentially disrupt critical infrastructure, prompting cautious, invite-only access for trusted organizations.
Axios corrected themselves shortly after: This is not the "Spud" model, but a completely different one.
🚀 OpenAI Targets Massive Ad Expansion
OpenAI is projecting explosive ad revenue growth, from $2.5B in 2026 to a staggering $100B by 2030, driven by rapid user expansion and highly targeted chatbot-based advertising. Early tests already hit $100M ARR in under two months, signaling strong demand and scalability.
The strategy hinges on reaching 2.75 billion weekly users (ambitious goal) and leveraging direct user intent (what people explicitly ask for), but it raises concerns about trust, user experience, and shifting incentives toward advertisers over users.
🩺 Google Med(ical)Gemma 1.5 Gets Smarter
Google’s MedGemma 1.5 4B upgrades the original open medical model into a much broader multimodal system, adding 3D CT/MRI reading, whole-slide pathology, chest X-ray bounding-box localization, multi-timepoint imaging, and stronger lab report/EHR understanding. The headline numbers are impressive: +11% absolute accuracy on 3D MRI classification, +3% on 3D CT, a 47% macro F1 jump in whole-slide pathology, +35% IoU for anatomical localization, +5% on MedQA, and +22% on EHRQA versus MedGemma 1 4B.
Let that sink in for a moment: a relatively efficient 4B open model can handle much richer clinical data types without becoming unusably heavy, giving researchers and developers a more practical base for specialized healthcare AI tools. However, The report is also careful about limits, some legacy benchmarks dipped, and Google explicitly says the model is not ready for deployment without clinical fine-tuning. So the real win here is as a foundation model for building next-generation medical workflows, not a plug-and-play doctor replacement.
But hey, 4B medical model sounds really impressive, doesn’t it?
Andrej Karpathy: There’s a growing gap in how people perceive AI because casual users judge it based on outdated or free models, while professionals using cutting-edge tools like OpenAI Codex and Claude Code see dramatic, rapidly advancing capabilities, especially in technical domains where AI performance and progress are far more significant.
OpenAI’s Chief Scientist on Continual Learning Hype, RL Beyond Code, & Future Alignment Directions
OpenAI's $100 Pro Play
The Takeaway
👉 OpenAI introduced a $100 monthly ChatGPT Pro plan offering 5x Codex usage over Plus, filling a gap between $20 and $200, with a temporary 10x boost running through May 31.
👉 The launch is a direct competitive response to Anthropic, which recently hit $30 billion ARR (surpassing OpenAI) and blocked third-party coding harnesses like OpenClaw from using Claude subscriptions.
👉 OpenAI hired OpenClaw creator Peter Steinberger in February to lead personal agent strategy, turning a competitor's restriction into a recruitment and product opportunity.
👉 Codex now has over 3 million weekly users with 70% month-over-month growth, making AI coding tools the primary revenue battleground between OpenAI and Anthropic heading into mid-2026.
The AI coding wars just got a new price tag. OpenAI launched a $100 monthly ChatGPT Pro tier, slotting right between its $20 Plus and $200 Pro plans, with one clear target: developers hungry for more Codex usage. The new tier offers 5x the coding capacity of Plus, covering local messages, cloud tasks, and code reviews across models like GPT-5.4 and GPT-5.3-Codex. To sweeten the deal, OpenAI is temporarily doubling those limits through May 31.
But here's what makes this really interesting: the timing. Anthropic just reported $30 billion in annualized revenue, surpassing OpenAI's roughly $24 billion, largely driven by Claude Code adoption. Days before the new tier dropped, Anthropic blocked third-party harnesses like OpenClaw from using Claude subscriptions. OpenAI responded by hiring OpenClaw creator Peter Steinberger and rolling out a plan that directly mirrors Anthropic's Max 5x pricing. With over 3 million weekly Codex users and 70% month-over-month growth, this is OpenAI playing catch-up and offense at the same time.
Therefore, the new $100 plan can be seen more as a strategic release intended for increasing use by professional clients. It's a very attractively priced tier aimed at key customers.
Why it matters: OpenAI's new pricing tier signals that AI coding tools have become the central battleground in the subscription revenue war between the two biggest AI companies. The race to capture professional developers is now being fought not just on model quality, but on access, pricing tiers, and strategic talent acquisitions.
Sources:
🔗 https://venturebeat.com/orchestration/openai-introduces-chatgpt-pro-usd100-tier-with-5x-usage-limits-for-codex
🔗 https://techcrunch.com/2026/04/09/chatgpt-pro-plan-100-month-codex/
Speak naturally. Send without fixing.
Wispr Flow turns your voice into clean, professional text you can send the moment you stop talking. Not rough transcription you have to clean up. Actual polished text - ready for email, Slack, or any app.
Reid Hoffman sends 89% of his messages with zero edits using Flow. Millions of people worldwide have made it part of how they work, including teams at OpenAI, Vercel, and Clay.
Speak the way you think. Go on tangents. Change your mind mid-sentence. Flow strips the filler, fixes the grammar, and gives you text that reads like you spent five minutes writing it.
Works on Mac, Windows, iPhone, and now Android - free and unlimited on Android during launch.
Not a graph today, but instead an image: AI generated cow, 2014. When you consider how far we've come, it almost makes you dizzy.
The AI infrastructure crisis is worsening rapidly
The backbone of the AI revolution is under siege, and the threats are coming from two very different directions. In the U.S., roughly half of all data centers planned for 2026 are facing delays or outright cancellations. The culprit? A brutal supply chain bottleneck. Critical electrical components like transformers, switchgear, and batteries, which make up less than 10% of construction costs, have become impossible to source fast enough. Lead times for high-power transformers have ballooned from around two years before 2020 to as long as five years today. Despite billions in planned spending, only about a third of announced capacity is actually under construction.
Meanwhile, the geopolitical picture is even more alarming. Iran's IRGC has explicitly threatened OpenAI's $30 billion Stargate data center in Abu Dhabi, releasing satellite imagery of the facility. This follows actual drone strikes on AWS data centers in the UAE and Bahrain in March, the first deliberate military attacks on commercial data centers in history. Iran has named 18 major tech companies as potential targets.
The AI buildout isn't just a technology race anymore. It's a question of physical security and global supply chain resilience. And right now it’s not looking.
The global AI infrastructure buildout faces a perfect storm of domestic supply chain failures and international military threats, putting billions in planned compute capacity at risk. Whether it's missing transformers in Texas or missile threats in Abu Dhabi, the physical layer of AI is proving far more fragile than anyone anticipated.
The best marketing ideas come from marketers who live it. That’s what The Marketing Millennials delivers: real insights, fresh takes, and no fluff. Written by Daniel Murray, a marketer who knows what works, this newsletter cuts through the noise so you can stop guessing and start winning. Subscribe and level up your marketing game.
