💡 AMD Beat Earnings - Then Crashed 17%

February 9, 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.

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

Featured Event 📆

Hurry Up and 10x: The Path to Real AI Productivity

AI promised 10x productivity.

Your deployment metrics tell a different story.

[

→ Register Here (only a few spots left)

Hear what engineering leaders from OpenAI and ACI Worldwide are personally optimizing across the SDLC to close the gap between velocity claims and reality.

What You'll Learn

Uplevel data shows the specific bottleneck patterns across teams using AI tools.
What foundational practices matter most for AI effectiveness
Engineering practices built for human-paced development

Can't make it? Register anyway - they’ll send you the recording.

→ Hurry Up and 10x

Quick Hits 🎯

🔴 AMD crushes Q4 earnings, then bleeds out 17% on "disappointing" guidance that still beat consensus

🐛 Anthropic's new Claude Opus 4.6 casually discovered 500 zero-day vulnerabilities without being asked

💾 AI data centers are hoarding memory chips - your next laptop upgrade just got more expensive

🤖 OpenAI launches enterprise agent platform hours before Anthropic drops Opus 4.6 (coincidence?)

💰 Fundamental raises $255M to crack big data problems that traditional LLMs can't touch

🧠 Cerebras closes $1B round for wafer-scale AI chips as infrastructure arms race accelerates

📉 Software stocks crater as Claude triggers "SaaSmageddon" - Thomson Reuters down 15%, Legalzoom down 20%

🎁 + 3 other stories shaping the AI spending debate

The Big Picture 🖼️

💡 AMD's Earnings Call Just Exposed How Brutal AI Expectations Have Become.

AMD reported record Q4 revenue of $10.3 billion - up 34% year-over-year - with data center sales hitting $5.4 billion. They beat every analyst estimate. Then their stock got obliterated, dropping 17% in a single day. Why? Their Q1 2026 guidance of $9.8 billion (±$300M) technically beat consensus but fell short of the whisper numbers some analysts were betting on. The message from Wall Street is clear: in the AI chip race, merely beating expectations isn't enough anymore. You need to shatter them. AMD's mistake wasn't underperforming - it was showing up to an arms race with a slightly smaller gun.

💡 Anthropic's Latest Model Found 500 Zero-Days. It Wasn't Looking for Them.

Claude Opus 4.6 dropped Thursday, and buried in the launch announcement was a detail that should terrify every CISO: during pre-release testing, the model autonomously discovered over 500 previously unknown high-severity security vulnerabilities in open-source libraries. No specialized prompts. No security expertise loaded in. Just Claude, given access to debugging tools, doing what it does. The model found buffer overflows in OpenSC, memory corruption bugs in CGIF, and crash-inducing flaws in GhostScript - all while nominally being evaluated for other capabilities. The implications flip both ways: defenders just got a superweapon, but so did attackers once they figure out how to jailbreak these capabilities. Anthropic's adding friction layers to prevent abuse, but the genie doesn't go back in the bottle.

💡 The Memory Chip Shortage Isn't a Blip. It's Your New Reality.

AI data centers are consuming memory chips faster than manufacturers can produce them, and the supply-demand gap isn't closing anytime soon. Chipmakers like Micron have pivoted production lines toward high-margin AI memory, leaving consumer and enterprise segments scrambling for scraps. Dell's CFO already warned in November that PC prices will rise. The bottleneck isn't fab capacity - it's that by the end of 2026, existing facilities will be maxed out. Micron's new Idaho plant won't come online until 2027 at the earliest. Translation: if you're spec'ing hardware for 2026, budget for sticker shock. The AI boom doesn't just create winners and losers in software - it's restructuring the entire hardware supply chain underneath us.

💡 Fundamental Just Raised $255M to Solve the Problem LLMs Can't.

Traditional large language models choke on structured datasets with billions of rows - the kind of data that actually runs enterprises. Fundamental's Nexus foundation model doesn't use transformer architecture at all. It's deterministic, purpose-built for big data, and designed to analyze financial records, logistics data, and supply chain systems that make GPT-style models hallucinate. The $255M Series A from GV and General Catalyst signals something critical: we're past the "LLMs for everything" era and into specialized model architectures for specific problem domains. If your 2026 strategy still assumes transformers are the hammer for every nail, you're already behind. The infrastructure layer is fragmenting, and the winners will be whoever picks the right architecture for the right workload.

💡 Software Stocks Just Had Their "ChatGPT Moment" - In Reverse.

When Anthropic released industry-specific plugins for Claude Cowork last Friday, the software industry lost nearly $1 trillion in market cap over two days. Thomson Reuters dropped 15%. Legalzoom fell 20%. FactSet Research cratered 10%. The Nasdaq had its worst two-day tumble since April. Why? Because Claude Opus 4.6 can now perform financial analysis, legal research, and compliance work that traditionally required specialized SaaS subscriptions. Goldman Sachs announced they're building Claude-based agents for accounting and compliance - functions that currently employ thousands and rely on third-party software vendors. The panic isn't irrational. This isn't about AI augmenting knowledge work - it's about collapsing entire software categories into general-purpose models. Investors are recalculating the terminal value of any company whose moat was "we have proprietary data and a domain-specific UI."

💡 The AI Infrastructure Arms Race Just Went Nuclear.

Cerebras locked in a $1 billion round to scale wafer-scale AI chips. ElevenLabs raised $500M to expand beyond voice into multimodal agents. Bedrock Robotics closed $270M to deploy autonomous construction fleets. Positron raised $230M for AI inference chips at a $1 billion valuation. This isn't venture capital - this is industrial policy executed through private markets. Every major AI player is making multi-billion-dollar bets 2-4 years into the future, banking on revenues that don't exist yet. The spending is so aggressive that even bulls are asking if this is a bubble. But here's the thing about infrastructure races: they create real capabilities even if the business models don't pencil out. The question isn't whether the spending is sustainable - it's who gets left holding worthless assets when the music stops.

💡 OpenAI and Anthropic Launched Competing Products on the Same Day. That's Not an Accident.

Thursday morning: OpenAI releases a new enterprise platform for building and managing AI agents. Thursday afternoon: Anthropic drops Claude Opus 4.6 with "agent teams" that coordinate across multiple autonomous AI workers. Both companies emphasized the same pitch - moving beyond single-task automation toward multi-agent systems that function like human teams. The timing screams competitive urgency, but the real story is convergence. Every major AI lab is now racing toward the same architectural vision: AI that doesn't just answer questions or write code, but orchestrates complex, multi-step workflows without human intervention. Whoever nails reliable multi-agent coordination first doesn't just win a product category - they obsolete entire job functions. The enterprise software landscape in 2027 will look nothing like 2025.

💡 Anthropic's Super Bowl Ad Isn't About Claude. It's About Business Models.

While you're watching the Super Bowl tomorrow, Anthropic will air $16M+ worth of commercials with one message: "Ads are coming to AI. But not to Claude." The tagline is a direct shot at OpenAI, which announced last month it'll start testing ads in ChatGPT. The subtext matters more than the ads themselves. OpenAI signed $1.4 trillion in infrastructure deals in 2025 and needs new revenue streams to service that spending. Anthropic, backed by Amazon and Google, can afford to position itself as the "premium, ad-free" option. This isn't just branding - it's a bet that enterprise customers will pay more for AI that doesn't monetize their conversations. The lesson? In mature markets, business model becomes moat. The company that figures out sustainable unit economics wins, even if their model is technically second-best.

Uplevels premier event (few spots left) →

Trending Tools 📈

openclaw/openclaw

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

(+27.6k ⭐ this week, 🔷 TypeScript) → Link

Personal AI assistant that runs on any OS, any platform - the open-source answer to proprietary agents
Why it's hot: Community rallying around open AI agents after seeing what Claude Cowork can do

KeygraphHQ/shannon
(+5.9k ⭐ this week, 🔷 TypeScript) → Link

Fully autonomous AI hacker that finds real exploits in web apps - 96.15% success rate on XBOW Benchmark
Why it matters: Security testing is about to get democratized (and terrifying)

obra/superpowers
(+46.1k ⭐ this week, 🔷 Shell) → Link

Agentic skills framework for software development that actually works
Teams using it: Replacing boilerplate workflows with agent-driven automation

HKUDS/nanobot
(+10.9k ⭐ this week, 🔷 Python) → Link

Ultra-lightweight AI agent framework - the "Claude-lite" for resource-constrained environments
Adoption driver: Not everyone needs (or can afford) Opus-level capabilities

aquasecurity/trivy
(Steady growth, 🔷 Go) → Link

Find vulnerabilities, misconfigurations, secrets in containers, Kubernetes, code, clouds
Why now: Claude Opus 4.6 finding 500 zero-days reminded everyone their code is probably full of bugs too

d2lang/d2
(Rising interest, 🔷 Go)

Create architecture diagrams with simple scripting language - no drag-and-drop required
Use case: Documentation workflows speeding up as AI generates both code and diagrams

Tech Trend of The Week 📊

🔍 "AMD earnings" search interest spiked 100%+ this week

On Tuesday, February 3rd, AMD reported their Q4 2025 earnings after market close.

Their Q1 2026 guidance of ~$9.8B, while beating consensus estimates, fell short of the most bullish analyst expectations.

The result?

Their stock crashed 17% on Wednesday, February 4th - the worst single-day decline since 2017.

Search interest exploded as investors, engineers, and industry watchers tried to understand what went wrong when the numbers looked so right.

The signal: We've entered a new phase of the AI chip wars where meeting expectations equals failure. Nvidia set the bar with monster beats quarter after quarter.

AMD's "solid performance" got punished because it wasn't exceptional.

For anyone building on AI infrastructure, this recalibration matters - the companies supplying your compute are now being valued on trajectory, not results.

💡 AMD Beat Earnings - Then Crashed 17%