refactor: simplify product section content and layout

- Condensed GPU Architecture section from detailed bullet points to concise three-pillar overview
- Removed redundant descriptive paragraphs from Capabilities and Design sections
- Streamlined Network Hero messaging and added standalone/integrated positioning
- Created new NetworkCapabilities component with four-column grid layout

src/pages/gpu/GpuArchitecture.tsx

@@ -1,28 +1,18 @@
 import { Container } from '../../components/Container'
-import { Eyebrow, SectionHeader, P } from '../../components/Texts'
+import { Eyebrow, SectionHeader } from '../../components/Texts'
 
-const architectureSections = [
+const architecture = [
   {
-    title: 'Distributed GPU Mesh',
-    description:
-      'A peer-to-peer fabric connects GPU nodes across the ThreeFold Grid, exposing them through the Mycelium network.',
-    bullets: [
-      'GPU nodes distributed globally for on-demand acceleration.',
-      'Mycelium network provides encrypted peer-to-peer connectivity.',
-      'Smart contract orchestration allocates and governs resources.',
-      'Real-time monitoring tracks utilization and health.',
-    ],
+    title: 'Sovereign Compute Nodes',
+    description: 'GPUs hosted on hardware you trust.',
   },
   {
-    title: 'Performance Characteristics',
-    description:
-      'Consistency and transparency are built into the fabric so workloads behave predictably anywhere on the planet.',
-    bullets: [
-      'Edge-to-core deployment flexibility for every workload profile.',
-      'Adaptive intelligence optimizes allocation automatically.',
-      'Deterministic performance guarantees availability when booked.',
-      'Transparent cost tracking for every GPU cycle consumed.',
-    ],
+    title: 'Encrypted Mesh Networking',
+    description: 'Secure, private connectivity to workloads.',
+  },
+  {
+    title: 'Reservation & Verification Layer',
+    description: 'Guarantees GPU access and consistency.',
   },
 ]
 
@@ -31,41 +21,24 @@ export function GpuArchitecture() {
     <section id="gpu-architecture" className="bg-white py-24 sm:py-32">
       <Container>
         <div className="mx-auto max-w-3xl text-center">
-          <Eyebrow>
-            Technical Architecture
-          </Eyebrow>
+          <Eyebrow>ARCHITECTURE</Eyebrow>
           <SectionHeader as="h2" className="mt-6 text-gray-900">
-            A peer-to-peer mesh purpose-built for acceleration.
+            HOW IT WORKS
           </SectionHeader>
-          <P className="mt-6 text-gray-600">
-            Mycelium GPU treats every node as part of a sovereign mesh. Encrypted
-            networking, smart contract orchestration, and transparent monitoring
-            ensure your workloads receive precisely the power they request.
-          </P>
         </div>
-        <div className="mt-16 grid gap-8 lg:grid-cols-2">
-          {architectureSections.map((section) => (
+
+        <div className="mx-auto mt-16 max-w-4xl space-y-6 lg:space-y-0 lg:grid lg:grid-cols-3 lg:gap-8">
+          {architecture.map((item) => (
             <div
-              key={section.title}
-              className="flex h-full flex-col rounded-3xl border border-slate-200 bg-white p-8 shadow-sm transition hover:-translate-y-1 hover:border-cyan-300 hover:shadow-lg"
+              key={item.title}
+              className="rounded-3xl border border-slate-200 bg-white p-8 shadow-sm transition hover:-translate-y-1 hover:border-cyan-300 hover:shadow-lg"
             >
               <h3 className="text-xl font-semibold text-gray-900">
-                {section.title}
+                {item.title}
               </h3>
-              <p className="mt-4 text-sm leading-relaxed text-gray-600">
-                {section.description}
+              <p className="mt-3 text-sm leading-relaxed text-gray-600">
+                {item.description}
               </p>
-              <ul className="mt-6 space-y-3 text-sm text-gray-600">
-                {section.bullets.map((bullet) => (
-                  <li
-                    key={bullet}
-                    className="flex items-start gap-3 rounded-2xl border border-cyan-100 bg-cyan-50/60 p-3 leading-relaxed"
-                  >
-                    <span className="mt-1 inline-block size-2 rounded-full bg-cyan-500" />
-                    <span>{bullet}</span>
-                  </li>
-                ))}
-              </ul>
             </div>
           ))}
         </div>

src/pages/gpu/GpuCapabilities.tsx

@@ -37,10 +37,6 @@ export function GpuCapabilities() {
         <div className="mx-auto max-w-3xl text-center">
           <Eyebrow>CAPABILITIES</Eyebrow>
           <H3 className="mt-4 text-gray-900">What You Can Run on Mycelium Cloud</H3>
-          <P className="mt-6 text-gray-600">
-            GPU workloads can run anywhere on the mesh — on cloud capacity, your own machines,
-            or edge nodes — with full sovereignty.
-          </P>
         </div>
 
         <div className="mx-auto mt-16 max-w-5xl">

src/pages/gpu/GpuDesign.tsx

@@ -34,10 +34,7 @@ export function GpuDesign() {
         <div className="mx-auto max-w-3xl sm:text-center">
           <Eyebrow>CORE VALUE</Eyebrow>
           <H3 className="mt-4 text-gray-900">GPU Power You Actually Control</H3>
-          <P className="mt-6 text-gray-600">
-            Mycelium GPU gives you access to real hardware, not gated queues or
-            proprietary runtimes.
-          </P>
+
         </div>
 
         {/* Key Benefits */}