Note: "Fear & Greed" is a common index, but we can translate as "Frykt og grådighet". "slightly bearish" -> "litt bearish" (bearish is used in Norwegian too).
Now, section 1: "Vera, the chip that designs chips"
Original: "Vera is NVIDIA's own server-class CPU, built around 88 custom Olympus cores, LPDDR5X memory, and second-generation Scalable Coherent Fabric. By running EDA tools on Vera, NVIDIA is effectively using its own silicon to design the next round of silicon. The early numbers: up to 1.5x higher performance on Cadence's Jasper formal verification tool and Synopsys' VCS simulator. NVIDIA didn't disclose the baseline, so the 1.5x figure is best read as a targeted gain on specific production-class workflows, not a blanket speedup across every design task."
Norwegian: "Vera er NVIDIAs egen serverklasse-CPU, bygget rundt 88 tilpassede Olympus-kjerner, LPDDR5X-minne og andre generasjon Scalable Coherent Fabric. Ved å kjøre EDA-verktøy på Vera bruker NVIDIA effektivt sin egen silisium til å designe neste runde med silisium. De tidlige tallene: opptil 1,5 ganger høyere ytelse på Cadences Jasper-formelle verifiseringsverktøy og Synopsys' VCS-simulator. NVIDIA oppga ikke grunnlinjen, så 1,5x-tallet bør leses som en målrettet gevinst på spesifikke produksjonsklassearbeidsflyter, ikke en generell hastighetsøkning på alle designtasks."
Note: "Olympus cores" -> "Olympus-kjerner". "Scalable Coherent Fabric" is a proper noun, keep as is. "formal verification tool" -> "formelt verifiseringsverktøy". "VCS simulator" -> "VCS-simulator". "baseline" -> "grunnlinje" or "referanse". "blanket speedup" -> "generell hastighetsøkning". "design task" -> "designtask" or "designoppgave". We'll use "designtasks" as a loanword.
Section 2: "Why the design cycle matters"
Original: "Chip design is compute-hungry. Simulation and verification alone can burn through millions of core-hours per tape-out. If Vera holds up in broader deployment, NVIDIA compresses its own development cycles — which means new GPUs and CPUs hit the market faster. That's where the crypto angle sneaks in. Faster GPU refreshes don't move BTC or ETH on their own, but they feed the AI-hardware narrative that's been lifting AI-linked altcoins. The effect is indirect and slow. But it compounds."
Norwegian: "Brikkedesign er beregningskrevende. Simulering og verifisering alene kan bruke millioner av kjernetimer per tape-out. Hvis Vera holder mål i bredere utrulling, komprimerer NVIDIA sine egne utviklingssykluser – noe
