Flying cars are officially here 🤯

AI is becoming the biggest infrastructure build-out in U.S. history. AI spending is projected to reach 3.63% of GDP. Bigger than railroads, highways, electrification, canals, and telecom.

Success isn’t the absence of problems. It’s identifying them early and addressing them head-on. Problems are inevitable. How quickly you identify and respond to them defines the outcome.

Marketing is sales at scale. Sales is one-to-one. Marketing is one-to-many. Both have the same job: Create trust. Communicate value. Move people to action.

Tom Brady’s career changed when he stopped blaming everyone else. Stop worrying about what you’re not getting. Own the opportunities you do get. Show up prepared. Bring the energy. Treat practice like the game. Confidence comes from preparation + ownership. “What can I do better?” 💪🏼

You don’t need to be extraordinary every day. You just need to keep showing up. Consistency > Motivation 💪🏼

AI data centers are usually discussed in terms of what they consume. Power. Water. Land. But that’s only part of the picture. They’re also creating jobs, expanding local tax bases, funding schools and infrastructure, and supporting one of the largest investment cycles of the AI era. PwC estimates AI could add $4.4T to U.S. GDP by 2030 and support millions of jobs. And in communities already seeing major data center investment, the impact is tangible. More tax revenue. More infrastructure. More economic activity. We should absolutely measure what AI infrastructure consumes. We should also measure what it creates.

"You want to make mistakes. You just want to make them small, iterate and fix them quickly." - Michael Dell

Shadow AI is real. 78% of AI users are already bringing their own tools to work. If you don’t think your team is using AI, your head is in the sand. Leaders need to empower their people AND keep company data secure. Enable it. Secure it. Don’t ignore it.

Exactly. This is the shift I’ve been writing about. As AI moves toward reasoning and agents, context becomes increasingly important, and that puts enormous pressure on memory bandwidth, capacity, storage, and data movement. Data wants to be close to compute. The AI boom is becoming a memory boom. I went deeper on this here: xitter.cf/jeffmcleod/status/2077…

“Humans will be a rounding error on the internet.” That sounds absurd until you look at the trajectory. Cloudflare believes machine-generated traffic could exceed human traffic by 1,000x within five years. The internet is being rebuilt in real time: • AI talking to AI • Agents buying from agents • Software creating software • Humans moving up the stack The next era of the internet will be agent-first. 🚀

AI is driving a $1 trillion memory supercycle. That was the clearest takeaway from #FMS2026 this week. For years, the focus was on compute. Now the bottleneck is shifting. Training needs compute. Inference needs memory. Agents need context. Context needs storage. Across presentations from analysts, neocloud providers, and memory and storage companies, one message came through repeatedly: This isn’t another memory cycle. It’s a structural shift. Hyperscalers are locking up supply years in advance. HBM demand is exploding. Long-term supply agreements are replacing spot purchases. Memory vendors are gaining significant pricing power. The AI race is becoming a race to move, store, and retrieve information as quickly as possible.

Meta says its AI hacked another company during cybersecurity testing. OpenAI did it. Anthropic did it. New frontier benchmark: Can your model pass SWE-bench and breach a corporate network?

This slide from SK hynix at #FMS2026 is staggering. 🤯 AI compute performance: 300x+ GPU memory capacity: 12x+ Memory demand for AI workloads: 50,000x+ Reasoning models, inference, and AI agents are driving an explosion in context, data movement, and memory requirements. The AI boom is becoming a memory boom. And we’re still in the early innings.