Tesla just tripled its annual capital expenditure to $25 billion in 2026, up from $8.5 billion last year. The money is going into AI infrastructure, chip design, semiconductor research, battery technology, and Optimus robot production. The Fremont factory, which used to make Model S and Model X, is being converted to manufacture Optimus at scale. A dedicated Optimus plant is also being built near Austin. To put the number in context, Musk noted Amazon is spending $200 billion in capex this year and Google $175 billion. The scale gap is real, but the direction is the same. ...

I came across this infographic and spent more time than I expected just reading through it. It’s a good snapshot of how much the anatomy of a server rack has shifted in the last few years. A rack used to be mostly about compute and storage. CPUs on top, drives somewhere in the middle, some networking at the top, and air blowing through the whole thing. The job of the infrastructure was to stay out of the way of the workload. Now the workload is the infrastructure. GPUs are the centre of gravity, and everything else, power distribution, cooling, interconnects, cable management, is designed around keeping them fed and cold enough to run flat out. ...

Robotics labs have spent years recording humans doing physical tasks to teach their systems when and how to grab, walk, or stack boxes. Meta just brought that playbook to software and computer use, except the demo subjects are its own employees. The program is called Model Capability Initiative. It captures mouse movements, clicks, and keystrokes across hundreds of apps and websites including Google, LinkedIn, GitHub, and Slack. No opt-out. The data goes straight into training AI agents that can perform knowledge work autonomously. ...

Moonshot AI just open-sourced K2.6, and the benchmarks are hard to ignore. It beats or matches GPT-5.4, Opus 4.6, and Gemini 3.1 Pro on Humanity’s Last Exam with tools and SWE-Bench Pro, which are two of the more credible tests for reasoning and coding. It can run for 12 hours straight across 4,000 tool calls. One internal agent apparently ran autonomously for five days. And it can spin up 300 parallel sub-agents at the same time. ...

Hyperliquid is one of those stories that forces you to reconsider most of what you think you know about how companies get built. Jeff Yan, Harvard graduate and gold medallist at the International Physics Olympiad, turned down a $1 billion funding offer, never took a dollar of venture capital, and built a crypto trading exchange that generated over $900 million in profit with 11 employees. At three years old, it has a market cap of $10 billion. The numbers are hard to believe. ...

For 14 years, Google has been the undisputed king of digital advertising. That ends in 2026. Meta’s global net ad revenue is projected to hit $243.5B this year, edging past Google’s $239.5B, and the gap is only expected to grow. What makes this interesting is not just the numbers, it is what they represent. Google built its empire on search, on the idea that intent is the most valuable signal an advertiser can buy. And for a long time, that was true. ...

I’m as bullish on AI as anyone. But there’s a pattern playing out in companies right now that I find genuinely frustrating, and this article puts it well. Goodhart’s Law says that when a measure becomes a target, it ceases to be a good measure. And right now, AI usage is the target. Token counts, Anthropic bills, n8n workflows shared in Slack, skills written in markdown, dashboards tracking adoption. In fact, dashboards all over the place. ...

Andon Labs put an AI called Luna in charge of a real retail store in San Francisco. Not a simulation, not a sandbox. A real shop, real money, real decisions. Luna hired human staff, selected inventory, set prices, and ran marketing outreach, all on her own, for three years. What I find genuinely impressive is not that it worked perfectly, it didn’t, but that it worked at all at this level. Luna was doing things that require judgment: reading job applicants in brief interviews, deciding which products fit the store’s identity, reaching out to suppliers. She picked books on AI risk and handmade art prints for the shelves. She hired on the spot about half the people she met. ...

Stanford HAI just published its 2026 AI Index and it is one of those reports that forces you to sit down and rethink what you thought you knew. AI has reached more than half the world’s population faster than the PC or the internet ever did, but the public mood around it is dark. Trust is at record lows, entry level workers are getting squeezed out, and the gap between what insiders think and what everyone else feels has never been wider. ...

One of the things that still makes robots look robotic is the way they move. Jerky, mechanical, imprecise. A big part of that comes down to how they are built, servo motors crammed into joints, converting rotation into movement in a way that biological muscle simply does not. Researchers at MIT and Politecnico di Bari may have just found a better way. They developed what they call electrofluidic fiber muscles, tiny actuators about as thick as a toothpick that contract when electricity is applied, no motors, no external pumps, no noise. The whole thing works by injecting charge into a sealed dielectric fluid, which creates ions that move the fluid and generate pressure. The result is a fiber that behaves remarkably like real muscle. ...