💡 The $26.5B IPO That's Really About Who Controls the Chip Supply

July 16, 2026

Get Codeinated ☕

Join 40,000 others and get Codeinated in 5 minutes. The free weekly email that wakes up your tech knowledge. Five minutes. Every week. No drowsiness.

In partnership with

☕ Morning! 💡 Your Weekly 5-Minutes of Caffeine and Tech Clarity

Featured Tool 📊

FREE OFFER: The CISO's Guide to Social Engineering Susceptibilities

AI-crafted phishing emails are 4.5x more likely to get clicked

Hi Friend of NINJIO,

AI has cut the time to produce a convincing phishing email from 16 hours to about five minutes. Those emails now arrive personalized to each target, and deepfaked follow-up calls have turned phishing into a multi-stage operation.

The observable signals individuals were trained to catch are largely absent from AI-generated attacks today. Many phishing simulation programs were built for a different version of this threat.

NINJIO's new guide for cybersecurity leaders covers what phishing simulation programs need to look like now, including:

How simulation frequency affects behavioral retention and why a quarterly cadence falls short
How to map simulations to emotional triggers and build individual susceptibility profiles
How to extend simulation coverage to vishing and SMS-based attacks
Which metrics beyond click rate show whether behavioral change is happening

Download Now

Matt Lindley

Chief Innovation & Info. Security Officer

matt.lindley@ninjio.com

91% of NINJIO Clients report a stronger cybersecurity posture after 8 months with the program.

Interested in improving your cybersecurity program?

Book a Demo

You're receiving this because you've previously interacted with NINJIO.

To change your preferences choose a link below.

NINJIO, LLC, 880 Hampshire Rd., Suite B, Westlake Village, CA 91361, United States, (805) 864-1999

Manage preferences

Quick Hits 🎯

🍏 SK Hynix just raised $26.5B in the biggest foreign IPO in US history and Washington is already asking for new fabs on American soil
🧠 Anthropic pulled back the curtain on how Claude actually thinks, right as OpenAI pushes toward a everything-app strategy
🧭 Hugging Face's CEO says open models aren't a side project anymore. They're the main event for enterprise AI
🗣 Companies are quietly ditching rented AI infrastructure for models they own and control
🧩 Open source pricing keeps dropping. Anthropic's business isn't shrinking. Here's why both things are true
🚀 DeepSeek cut prices 75% this week. The bill for running AI at scale barely moved
💀 A new attack vector called "slopsquatting" is exploiting AI coding tools - and most teams have no defense for it

🎁 + 2 other stories you might find useful

Our Partner 🎉

Your employees are connecting AI to everything. Now what?

ChatGPT and Claude don't just answer questions anymore. Employees are connecting them directly to Notion, Linear, Jira, and the rest of your stack. The AI can read, write, and take actions on company data. Most IT and security teams have no visibility into any of it.

Harmonic Security Connectors changes that. It sits inline with every AI-to-app connection, so you see each call, control what data moves, and block destructive actions before they happen. Employees notice nothing different.

See what's actually running across your business in a live demo.

See what's connected

The Big Picture 🖼️

💡 The Chip Supply Chain Just Became a Geopolitical Bargaining Chip.

SK Hynix didn't raise $26.5B just to fund growth. It raised the largest foreign IPO in US history, and now it's being pushed to build fabs on American soil. That's not a financing story, it's a positioning story.

Governments have figured out that whoever controls memory and chip production controls the pace of the entire AI industry. If a single factory goes down, every model that depends on that hardware stalls with it. Regionalizing supply is now a policy priority, not just a business decision.

For anyone building or buying AI infrastructure, this is worth watching closely, the vendors who control regional supply are about to have leverage they've never had before. Read what's actually driving the biggest foreign IPO in US history.

💡 Transparency Is Becoming Anthropic's Competitive Wedge.

While OpenAI keeps building toward a bundled, everything-in-one-app experience, Anthropic went the other direction, publishing research into how Claude actually reasons internally.

That's not just a research flex. In regulated industries, being able to show your work matters as much as raw capability. A model you can audit is a model legal, healthcare, and finance teams can actually deploy. Safety isn't a constraint here, it's turning into the feature that unlocks entire markets OpenAI's approach can't touch yet.

The takeaway: as platforms race to bundle more into fewer apps, the ability to explain what's happening under the hood might be worth more than another few points on a benchmark. See what Anthropic's research revealed and how it stacks up against OpenAI's super app push.

💡 Open Source Stopped Being the Backup Plan.

Hugging Face's Clem Delangue is making the case that open AI platforms aren't a scrappy alternative to proprietary models anymore, they're becoming the foundation enterprises actually build on.

If that holds, the competitive advantage shifts. It's no longer about who owns the biggest private model. It's about who participates best in a shared ecosystem of tools, weights, and infrastructure. Speed and interoperability start to matter more than raw model size.

For engineers and technical leaders, this changes what "moat" even means going forward. Here's why open source AI matters more than ever, according to the person building the platform for it.

💡 Enterprises Are Done Renting Their AI.

Hugging Face's CEO is describing a shift happening quietly across large companies: building internal pipelines on models and datasets they control, instead of paying a monthly bill to rent someone else's.

That's a bigger deal than it sounds. It means the real competition isn't just "whose model is smarter," it's who can govern, reproduce, and securely deploy AI at scale. Owning the pipeline becomes the moat, not owning the model.

If you're evaluating AI vendors right now, this is the question worth asking: are you renting capability, or building it? Read why companies are done renting their AI.

💡 Open Source Isn't Killing Anthropic - Yet.

Here's the tension: open-weight models keep getting cheaper and better, and yet Anthropic's business keeps growing. Both of those things are true at the same time, and understanding why matters.

The pattern looks like this.

Teams prototype fast on open baselines, then pay for safety-hardened, fine-tuned, enterprise-grade deployment when it's time to ship something that actually matters. Open source fuels experimentation. Frontier labs capture the revenue when reliability and compliance are non-negotiable.

The lesson for anyone building AI products: price competition on the open side doesn't automatically threaten the paid side, as long as you're solving a different problem. Here's the framing behind why open source isn't hurting Anthropic yet.

💡 Cheaper Tokens Don't Mean Cheaper AI.

Get Codeinated ☕

Join 40,000 others and get Codeinated in 5 minutes. The free weekly email that wakes up your tech knowledge. Five minutes. Every week. No drowsiness.

Subscribe Free

DeepSeek cut prices 75% this week. If you assumed that would blow up the economics of running AI at scale, the actual numbers say otherwise, usage keeps growing faster than prices are falling.

The real differentiation isn't who has the cheapest tokens anymore. It's who can orchestrate models efficiently, routing the right task to the right model, cutting latency, and getting more accuracy per dollar spent. Cheap tokens without smart orchestration just means you burn through them faster.

For technical teams, the leverage has moved from "which model is cheapest" to "how well do you run what you've got." Here's the 100x problem that price cuts alone can't solve.

💡 AI Coding Tools Opened a New Security Hole.

Forget typosquatting, there's a new attack called slopsquatting, and it's specifically built to exploit AI coding assistants that hallucinate package names.

Here's how it works in practice: an AI tool suggests a dependency that doesn't exist. Someone registers that exact package name with malicious code inside it. The next AI-assisted developer who trusts the suggestion pulls it straight into production.

If your team is using AI coding tools without dependency verification in place, this is worth fixing this week, not next quarter. Read about the software supply chain threat AI coding tools accidentally created.

💡 Gaming's Platform Wars Are a Preview of What's Coming Everywhere Else.

Xbox's struggles point to something bigger than one console losing ground, it's a signal that platform-agnostic, cloud-first experiences are winning against anything tied to a single box.

The same logic applies well beyond gaming. Users increasingly expect to move between devices without losing their progress, their data, or their personalization. Companies still betting on a single hardware ecosystem are competing against companies betting on flexibility everywhere.

If you're making platform decisions for your own product, this is the lesson worth borrowing: bet on reach across devices, not ownership of one. See what's actually happening inside Xbox's platform strategy.

💡 The Optimistic Case Is Quietly Building Underneath All of This.

Between modular data center design, manufacturing signals, and energy-efficient hardware, there's a real case forming for AI infrastructure that's cheaper to run, easier to govern, and faster to adapt than what came before.

None of these threads the chip IPO, the safety research, the open source shift, the price wars, the security gap, the platform wars, are separate stories. They're all pieces of the same shift: AI is becoming less about who has the flashiest model and more about who builds the most trustworthy, well-governed system around it.

The takeaway for the week: infrastructure and governance decisions made now are the ones that compound. Catch the full read in this week's Weekly Dose of Optimism.

Trending Tools 📈

oven-sh/bun (+166 ⭐ per day, 🔷 Rust) Link

An all-in-one JavaScript runtime, bundler, test runner, and package manager
Helps teams: cuts build and test cycle time by replacing three or four separate tools with one

astral-sh/ruff (+41 ⭐ per day, 🔷 Rust) Link

An extremely fast Python linter and formatter
Helps teams: catches code quality issues in CI without slowing builds down

Infisical/infisical (+60 ⭐ per day, 🔷 TypeScript) Link

Open-source secrets, certificate, and access management platform
Helps teams: centralizes credential management and closes off a common supply chain risk

tailscale/tailscale (+49 ⭐ per day, 🔷 Go) Link

Zero-config private networking built on WireGuard
Helps teams: gives distributed teams secure access without the usual VPN setup pain

caddyserver/caddy (+34 ⭐ per day, 🔷 Go) Link

A web server with automatic HTTPS out of the box
Helps teams: removes the manual certificate work from hosting and reverse proxy setups

docker/compose (+38 ⭐ per day, 🔷 Go) Link

Define and run multi-container applications from a single YAML file
Helps teams: makes local dev environments and CI setups reproducible across the team

donnemartin/system-design-primer (+154 ⭐ per day, 🔷 Python) Link

A comprehensive, free guide to designing systems at scale, with built-in flashcards
Helps teams: gives engineers at every level a shared vocabulary for architecture conversations

microsoft/PowerToys (+214 ⭐ per day, 🔷 C) Link

A set of utilities that extend what Windows can do out of the box
Helps teams: small productivity wins that add up fast for anyone on Windows

Tech Trend of The Week 📊

🔍 "SK Hynix IPO" is spiking in US search interest this week

The largest foreign IPO in US history doesn't happen quietly, and it didn't this time either. Search interest jumped as soon as the $26.5B raise hit the wires and stuck around once reports surfaced that Washington wants new fabs built on US soil, not just capital raised abroad.

The signal: chip supply is no longer a background story people trust to play out on its own. It's become something people are actively trying to understand in real time - which usually means it's about to show up in a lot more boardroom conversations, not just tech headlines.

Our Partner 🎉

Your competitor's growth lead already saw the spend spike.

While your team is still in standup, the other growth lead already got the alert. Viktor is an AI employee that lives in Slack. It watches your Meta and TikTok spend overnight, flags the underperformer by 7am, and drafts the new brief before your first meeting.

Get Started for Free