[Daily Briefing] Nvidia eyes your PC 💻 / TSMC price hike 📈 / Cisco root bug 🔒
Nvidia wants to be inside your next laptop, not just the graphics card
For years Nvidia has been the company that sells the chip that makes your PC good at games and, lately, at AI. Now it wants the other big chip too. Reports this week say Nvidia is building its own line of PC processors for a new wave of "AI agent" laptops and desktops, working with Microsoft, Dell and HP to get them into machines people actually buy.
That drops it straight into a roughly $200 billion market that Intel and AMD have shared between them for decades. Nvidia's angle is the bit it already rules. If the next round of PCs is judged on how well they run AI on the device itself, the firm that makes the best AI silicon starts with a very handy head start.
It is a real shift in the shape of the PC world. For most of computing history the processor was Intel's to lose, with AMD playing the plucky challenger. A third player this well funded walking in with the AI story on its side changes the sums for everyone, the laptop makers included, since they now get to play all three off against each other.
Why it matters: if your next work laptop boasts about "on-device AI", there is a fair chance the brain doing it wears an Nvidia badge. More competition in PC chips usually means better machines at keener prices, which is a rare bit of good news for anyone who actually has to buy the things.
Read more on Future Technology
Quick hits
Cisco is racing to patch a root-level bug that is already being used. CVE-2026-20262 in Catalyst SD-WAN Manager lets an attacker climb all the way to root through a crafted file upload, and it hits every deployment type. If you run it, patching is a this-weekend job. Full story (3 min read)
TSMC is putting its prices up again. The world's most important chipmaker has settled on base-price rises of 5 to 10 percent from 2027, after a quarter where net income jumped 77 percent. When TSMC charges more, the cost tends to ripple out to almost everything with a modern chip inside it. Full story (3 min read)
Intel just landed a milestone it badly needed. It says it is the first to ship high-volume logic chips made with ASML's High-NA EUV machines, alongside better yields on its 18A process. After a bruising couple of years, that is a sign the comeback plan is turning into actual silicon. Full story (2 min read)
Apple's AI is quietly running on Nvidia chips in Google's cloud. Three fierce rivals, one shared stack. Nvidia confirmed its GPUs inside Google Cloud are helping power Apple Intelligence, a tidy reminder that behind the marketing, everyone rents from everyone. Full story (3 min read)
The data-centre gold rush has a fresh billion-dollar name. Switch is reportedly in talks to raise billions at a valuation above $50 billion, with Brookfield and KKR circling, to build ever-larger campuses for AI demand. The picks-and-shovels layer of AI keeps quietly minting money. Full story (2 min read)
Tool of the day
With the whole industry suddenly obsessed with running AI on your own machine, here is a way to try it without paying anyone a monthly fee. Ollama is a free tool that lets you download and run open AI models locally, straight on your own PC or Mac, with no account and nothing leaving your device.
Honest take: it is the gentlest on-ramp to local AI I have found, and keeping your data on your own machine is a genuine privacy win. It is not magic, though. The better models are large, so you will want a decent chunk of memory and ideally a proper GPU, or you will be watching the cursor blink. Start with a small model, see how your machine copes, then work up. More on Future Technology
That is your weekend briefing. One question back to you: would you trust an "AI PC" to run things on the device itself, or do you still want that happening in the cloud where someone else worries about it? Hit reply and tell me, I read every one.
Nath, Future Technology
Forwarded this? Get your own free briefing at futuretechnologyhq.com/newsletter
Some links may be affiliate links.