Because they are like us. Don't want to take dependencies on others for something so key to their economy.
MARKET TRENDS
-
xAI Faces Development Challenges Ahead
By
–
He's trying. xAI has a lot of things to fix. It'll be messy for a while.
-
Folding-Wing eVTOL Technology Revolutionizes Vertical Flight
By
–
Folding-Wing eVTOL: Speed Meets Vertical Freedom
— Ronald van Loon (@Ronald_vanLoon) 15 avril 2026
via @ZappyZappy7
#Aviation #Technology #Innovation #EmergingTech #TechForGood pic.twitter.com/KrYUOlDs7MFolding-Wing eVTOL: Speed Meets Vertical Freedom
via @ZappyZappy7 #Aviation #Technology #Innovation #EmergingTech #TechForGood -
Lists Beat Algorithms: Building Curated Tech Industry Collections
By
–
This is the role of lists. Engage in lists and that resets the algo a bit. What you share and write about is the most important. I built lists of the entire tech industry. They are way better than the algo: https://
x.com/scobleizer/lis
ts
… -
AI Startups Struggle with Differentiation in Crowded Market
By
–
I have some ideas. 🙂 The real problem is that there are so many startups doing very similar things. Hard to differentiate between all of them, especially ones heavily focused on AI. It's hard for anyone to tell the difference between even Anthropic and OpenAI until you use
-
Bleeding Edge Products and Long-Term Innovation Impact
By
–
Looking at the comments, people don't understand the role of bleeding edge products. The first Macintosh could have been seen as a disaster too. It even got Steve Jobs fired since it didn't sell well. But it set up decades of new products that followed that DID sell well. The
-
SambaNova Intel Launch Agentic AI Inference Blueprint H2 2026
By
–
ICYMI: SambaNova + @Intel just laid out a new blueprint for agentic AI inference — disaggregated by design. RDUs for decode, Xeon 6 for orchestration + tool calling. Built for real enterprise workloads. H2 2026. @sallywf @eetimes
-

Nvidia’s AI Chip Dominance: Supply Chains, Competition, and Geopolitics
By
–
— NVIDIA AI (@NVIDIAAI) 15 avril 2026
The Jensen Huang episode. 0:00:00 – Is Nvidia’s biggest moat its grip on scarce supply chains?
0:16:25 – Will TPUs break Nvidia’s hold on AI compute?
0:41:06 – Why doesn’t Nvidia become a hyperscaler?
0:57:36 – Should we be selling AI chips to China?
1:35:06 – Why doesn’t Nvidia -
Jensen Huang Defends Position Calmly at GTC Conference
By
–
One reason I love Jensen is how he defends his position. One of the calmest CEOs around, I watched him answer questions from the world's press at GTC for two hours.
— Robert Scoble (@Scobleizer) 15 avril 2026
Listen to about 1 hour and 10 minutes into this to see him get ferocious about taking on Dwarkesh and stick up… https://t.co/8qpv4PnVVjOne reason I love Jensen is how he defends his position. One of the calmest CEOs around, I watched him answer questions from the world's press at GTC for two hours. Listen to about 1 hour and 10 minutes into this to see him get ferocious about taking on Dwarkesh and stick up
-

Blackwell GPU: 50× Higher Token Output Than Hopper
By
–
The outcome numbers tell the real story. Looking only at compute cost misses the story. Blackwell may appear ~2× more expensive than Hopper. FLOPS/$ suggests ~2× improvement. But the real outcome is dramatically different for NVIDIA Blackwell: >50× higher token output