feat: dynamic hype cycle + market signal engine + eBay/CapEx panels

6-source composite Market Signal Score (0-100) per transceiver technology.
New GET /api/hype-cycle/market-signals blends: Norton-Bass hype_score,
hyperscaler CapEx YoY (MSFT +68.8%, GOOG +107%, META +46.8%), price
observation activity ratio 30d vs prior 30d, AI cluster transceiver demand,
eBay secondary market sell-through velocity, internal fast-mover trend.
All 6 queries run in parallel via Promise.all().

Recommendation engine maps hype phase × capex boom × speed class →
Buy/Hold/Watch labels with tooltips. Dashboard Hype Cycle table now shows
Market Signal ● LIVE column + Recommendation column. Hyperscaler CapEx
panel + eBay panel added to hype tab. Procurement: new eBay Market section.
Sourcing Hype Cycle replaced hardcoded seed with live price observation data.

CHANGELOG_PENDING.md

@@ -1,6 +1,7 @@
 # TIP Changelog
 
 Format: `{"d":"YYYY-MM-DD","t":"TYPE","m":"Description"}`
+{"d":"2026-05-14","t":"FEAT","m":"Dynamic Hype Cycle + Market Signal Engine: Hype Cycle tab is now fully data-driven. New GET /api/hype-cycle/market-signals endpoint blends 6 real data sources into a composite Market Signal Score (0–100) per technology: (1) hype_score from Norton-Bass model (30% weight), (2) hyperscaler CapEx YoY avg (Microsoft +68.8%, Alphabet +107.4%, Meta +46.8%), (3) price observation activity ratio 30d vs prior 30d, (4) AI cluster estimated transceiver demand (90d window), (5) eBay secondary market sell-through velocity, (6) internal fast-mover demand trend. Score thresholds: ≥70 green, ≥50 yellow, ≥30 orange, <30 gray. Recommendation engine: buildRecommendation(phase, signalScore, capexYoyAvg, speedGbps) maps hype phase × capex boom × speed class → Buy/Hold/Watch label with color + detail tooltip. Dashboard: Hype Cycle table shows Market Signal ● LIVE column (score + progress bar) + Recommendation column (emoji label, tooltip with reasoning). Market Context cards row above table shows Top Signal, CapEx Boom %, Fast Movers signal, eBay Velocity. New Hyperscaler CapEx panel (SEC filing data) + eBay Secondary Market panel at bottom of hype tab. Procurement: new 🛒 eBay Market sub-section with per-form-factor sell-through grid. All 6 queries run in parallel via Promise.all()."}
 {"d":"2026-05-14","t":"FEAT","m":"Procurement tab: 2 new sections with real data. (1) 📦 Internal Demand — Flexoptix internal SKU velocity from flexoptix_internal_demand table (8,585 SKUs: 70 fast-movers 53k units/12M, 239 regular, 979 slow, 7,297 dead stock). Summary cards with trend %%. Filter by velocity class. API: GET /api/procurement/internal-demand?velocity_class=&limit=&sort=. (2) 🤖 AI Clusters — live AI datacenter announcements from ai_cluster_announcements table (396 in last 30 days). Shows estimated transceiver demand per build, MW scale, company, location, source link. Filter for entries with transceiver estimates. Stats: total announcements, MW, distinct companies, total estimated transceivers. API: GET /api/procurement/ai-clusters?days=&limit=. Replaced misleading DEMO DATA banners on Signals + ABC sections with informational note pointing to Internal Demand data."}
 
 {"d":"2026-05-14","t":"FEAT","m":"Equivalences Explorer: new dashboard tab '🔀 Equivalences' — search 63,362 cross-brand mappings (46 vendors, 7,516 competitor products → 846 Flexoptix alternatives, Ø 93.9% confidence). APIs: GET /api/equivalences (search), /api/equivalences/transceiver/:id (per-product), /api/equivalences/stats, /api/equivalences/top-vendors. Transceiver detail modal now shows equivalences panel (FX alternatives or competitor products) + SVG price history sparklines (30-day, per source vendor) from 392k+ price observations."}

packages/api/src/llm/client.ts

@@ -22,25 +22,112 @@ const CLAUDE_BRIDGE_URL = process.env.CLAUDE_BRIDGE_URL || "http://localhost:325
 // ── Runtime-switchable provider state ──────────────────────────────────────
 // Reads from /opt/tip/blog-llm-settings.json if present (written by /api/blog/llm/switch).
 // Falls back to process.env, then to defaults. No restart required for switches.
+//
+// AUTO-DISCOVERY: At startup and on a periodic refresh, the active fo-blog-v* model
+// is validated against Ollama's actual model list. If the configured model no longer
+// exists (e.g. Magatama trained a new version and Ollama removed older tags), the
+// highest available fo-blog-v* version is picked automatically — no manual env or
+// settings-file update needed after each training cycle.
 
 const SETTINGS_FILE = join(process.env.TIP_ROOT || "/opt/tip", "blog-llm-settings.json");
+const STATIC_FALLBACK_MODEL = "fo-blog-v10";
+const DISCOVERY_REFRESH_MS = Number.parseInt(process.env.BLOG_LLM_DISCOVERY_REFRESH_MS || "", 10) || 10 * 60_000;
 
 interface LlmSettings { provider: string; ollamaModel: string }
 
-function loadSettings(): LlmSettings {
+function loadSettingsRaw(): LlmSettings {
   try {
     if (existsSync(SETTINGS_FILE)) {
-      const raw = JSON.parse(readFileSync(SETTINGS_FILE, "utf8")) as LlmSettings;
+      const raw = JSON.parse(readFileSync(SETTINGS_FILE, "utf8")) as Partial<LlmSettings>;
-      return { provider: raw.provider || "ollama", ollamaModel: raw.ollamaModel || "fo-blog-v7" };
+      return {
+        provider: raw.provider || process.env.BLOG_LLM_PROVIDER || "ollama",
+        ollamaModel: raw.ollamaModel || process.env.OLLAMA_LLM_MODEL || STATIC_FALLBACK_MODEL,
+      };
     }
   } catch { /* ignore corrupt file */ }
   return {
     provider: process.env.BLOG_LLM_PROVIDER || "ollama",
-    ollamaModel: process.env.OLLAMA_LLM_MODEL || "fo-blog-v7",
+    ollamaModel: process.env.OLLAMA_LLM_MODEL || STATIC_FALLBACK_MODEL,
   };
 }
 
-let _settings = loadSettings();
+/**
+ * Sort fo-blog-v{N}[-r{M}] tags newest-first.
+ *
+ * Magatama convention (confirmed by /api/llm/status?lane=fo_blogllm):
+ *   - `fo-blog-vN`       ← base tag, the active production model after adoption
+ *   - `fo-blog-vN-rM`    ← revision metadata, intermediate adapter save
+ *
+ * So within the same major N, the BASE tag wins over -rM revisions.
+ *
+ * Order: higher N > lower N; within same N, base ("no -r") > any -rM revision.
+ */
+function compareFoBlogVersionsDesc(a: string, b: string): number {
+  const re = /^fo-blog-v(\d+)(?:-r(\d+))?$/;
+  const ma = re.exec(a);
+  const mb = re.exec(b);
+  if (!ma || !mb) return a.localeCompare(b);
+  const va = Number.parseInt(ma[1], 10);
+  const vb = Number.parseInt(mb[1], 10);
+  if (va !== vb) return vb - va;
+  // Same major version: base tag (no -r suffix) wins over any -rM revision
+  const aIsBase = ma[2] === undefined;
+  const bIsBase = mb[2] === undefined;
+  if (aIsBase !== bIsBase) return aIsBase ? -1 : 1;
+  // Both have -rM: higher M is newer
+  const ra = ma[2] ? Number.parseInt(ma[2], 10) : 0;
+  const rb = mb[2] ? Number.parseInt(mb[2], 10) : 0;
+  return rb - ra;
+}
+
+interface OllamaTag { name: string }
+interface OllamaTagsResponse { models: OllamaTag[] }
+
+/** Probe Ollama for available fo-blog-v* models. Returns [] on any error (non-fatal). */
+async function fetchOllamaFoBlogTags(): Promise<string[]> {
+  try {
+    const resp = await fetch(`${OLLAMA_URL}/api/tags`, { signal: AbortSignal.timeout(5000) });
+    if (!resp.ok) return [];
+    const data = await resp.json() as OllamaTagsResponse;
+    return (data.models || [])
+      .map(m => m.name.replace(/:latest$/, ""))
+      .filter(n => /^fo-blog-v\d+(?:-r\d+)?$/.test(n));
+  } catch {
+    return [];
+  }
+}
+
+/**
+ * Reconcile configured model against Ollama reality.
+ *
+ * Priority:
+ *   1. Configured model (env or settings file) — if Ollama actually serves it
+ *   2. Highest fo-blog-v* version Ollama actually serves — auto-discovered
+ *   3. Static fallback STATIC_FALLBACK_MODEL — last resort
+ *
+ * Non-blocking: any Ollama failure leaves _settings untouched.
+ */
+async function reconcileWithOllama(): Promise<void> {
+  const configured = _settings.ollamaModel;
+  if (!configured.startsWith("fo-blog-v")) return; // only manage fo-blog-* lane
+  const available = await fetchOllamaFoBlogTags();
+  if (available.length === 0) return;
+  if (available.includes(configured)) return; // configured model still exists
+
+  const sorted = [...available].sort(compareFoBlogVersionsDesc);
+  const winner = sorted[0];
+  if (!winner || winner === configured) return;
+
+  console.log(`[LLM] auto-discovery: configured "${configured}" not in Ollama; switching to latest available "${winner}" (candidates: ${sorted.join(", ")})`);
+  _settings = { ..._settings, ollamaModel: winner };
+  try { writeFileSync(SETTINGS_FILE, JSON.stringify(_settings, null, 2), "utf8"); } catch { /* non-fatal */ }
+}
+
+let _settings = loadSettingsRaw();
+
+// Fire-and-forget initial reconciliation. Subsequent refresh runs every DISCOVERY_REFRESH_MS.
+void reconcileWithOllama();
+setInterval(() => { void reconcileWithOllama(); }, DISCOVERY_REFRESH_MS).unref();
 
 /** Switch the active LLM provider at runtime. Persists to settings file. */
 export function setLlmProvider(provider: string, ollamaModel?: string): void {
@@ -52,6 +139,15 @@ export function setLlmProvider(provider: string, ollamaModel?: string): void {
 /** Returns the currently active provider config. */
 export function getLlmProvider(): LlmSettings { return { ..._settings }; }
 
+/**
+ * Force an immediate auto-discovery reconciliation against Ollama.
+ * Returns the active settings after reconcile.
+ */
+export async function refreshLlmAutoDiscovery(): Promise<LlmSettings> {
+  await reconcileWithOllama();
+  return { ..._settings };
+}
+
 // Convenience getters used below (re-read on every call for zero-latency switch)
 function provider(): string { return _settings.provider; }
 function llmModel(): string { return _settings.ollamaModel; }

packages/api/src/routes/hype-cycle.ts

@@ -165,6 +165,242 @@ hypeCycleRouter.get("/regional/:tech", (req: Request, res: Response) => {
   });
 });
 
+// ── MARKET SIGNAL COMPUTATION ───────────────────────────────────────────────
+/** Technology → form-factor/speed mapping for cross-table signal aggregation */
+const TECH_SIGNAL_MAP = [
+  { label: "10G-SFP+",     speedGbps: 10,   formFactors: ["SFP+"],            speedLabel: "10G"  },
+  { label: "100G-QSFP28",  speedGbps: 100,  formFactors: ["QSFP28"],          speedLabel: "100G" },
+  { label: "400G-QSFP-DD", speedGbps: 400,  formFactors: ["QSFP-DD"],         speedLabel: "400G" },
+  { label: "800G-OSFP",    speedGbps: 800,  formFactors: ["OSFP"],            speedLabel: "800G" },
+  { label: "1.6T-OSFP",   speedGbps: 1600, formFactors: ["OSFP"],            speedLabel: "1.6T" },
+  { label: "400G-ZR",      speedGbps: 400,  formFactors: ["SFP-DD", "CFP2"],  speedLabel: "400G" },
+] as const;
+
+type PhaseKey =
+  | "plateau_productivity" | "slope_enlightenment" | "trough_disillusionment"
+  | "peak_inflated_expectations" | "innovation_trigger";
+
+function buildRecommendation(
+  phase: PhaseKey,
+  signalScore: number,
+  capexYoyAvg: number,
+  speedGbps: number,
+): { label: string; color: string; detail: string } {
+  const fast = speedGbps >= 400;
+  const capexBoom = capexYoyAvg > 50;
+
+  switch (phase) {
+    case "plateau_productivity":
+      if (fast && capexBoom)
+        return { label: "🚀 Buy — AI Wave", color: "#16a34a", detail: "Commodity pricing + AI infrastructure demand surge. Stock up now." };
+      if (signalScore >= 85)
+        return { label: "✅ Hold — Stable", color: "#2563eb", detail: "Mature commodity market, stable long-term demand." };
+      return { label: "📦 Hold", color: "#64748b", detail: "Commodity market. Price-driven. Order on demand." };
+    case "slope_enlightenment":
+      if (capexBoom)
+        return { label: "🟢 Buy Now", color: "#16a34a", detail: "Growing mainstream adoption + hyperscaler capex boom. Window closing." };
+      return { label: "🟡 Buy", color: "#ca8a04", detail: "Adoption curve steepening. Monitor pricing before large orders." };
+    case "trough_disillusionment":
+      if (signalScore > 50)
+        return { label: "🔍 Buy Opportunity", color: "#7c3aed", detail: "Hype trough but demand signals emerging. Strategic buying window." };
+      return { label: "⏳ Watch", color: "#94a3b8", detail: "Wait for demand confirmation before stocking." };
+    case "peak_inflated_expectations":
+      if (fast && capexBoom)
+        return { label: "⚡ Caution / Buy", color: "#f97316", detail: "Hype peak but real hyperscaler demand. Buy selectively, not speculatively." };
+      return { label: "⚠ Caution", color: "#ef4444", detail: "Peak hype. Verify real end-customer demand before building inventory." };
+    case "innovation_trigger":
+      return { label: "👁 Watch", color: "#94a3b8", detail: "Early stage — too early for volume commitments. Monitor for traction." };
+    default:
+      return { label: "📊 Monitor", color: "#64748b", detail: "Monitor signals." };
+  }
+}
+
+// GET /api/hype-cycle/market-signals — Multi-source demand intelligence
+hypeCycleRouter.get("/market-signals", async (_req: Request, res: Response) => {
+  try {
+    const { pool } = await import("../db/client");
+
+    // ── Parallel data fetch ──────────────────────────────────────────────────
+    const [hypeRows, priceRows, capexRows, aiRows, ebayRows, internalRows] = await Promise.all([
+      // Latest hype cycle per technology
+      pool.query(`
+        SELECT DISTINCT ON (technology)
+          technology, hype_phase, hype_score, asp_current_usd,
+          r_squared, computed_at, current_share, years_to_next_phase
+        FROM hype_cycle_analysis
+        ORDER BY technology, computed_at DESC
+      `),
+
+      // Price observation activity: last 30d vs prior 30d (scraping frequency = demand proxy)
+      pool.query(`
+        SELECT
+          t.speed_gbps,
+          t.form_factor,
+          COUNT(*) FILTER (WHERE po.time >= NOW() - INTERVAL '30 days') AS obs_30d,
+          COUNT(*) FILTER (WHERE po.time >= NOW() - INTERVAL '60 days'
+                            AND po.time <  NOW() - INTERVAL '30 days')  AS obs_prior_30d
+        FROM price_observations po
+        JOIN transceivers t ON t.id = po.transceiver_id
+        WHERE po.time >= NOW() - INTERVAL '60 days'
+        GROUP BY t.speed_gbps, t.form_factor
+      `),
+
+      // Hyperscaler capex — most recent per company
+      pool.query(`
+        SELECT DISTINCT ON (company)
+          company, period_label, capex_usd_millions,
+          dc_capex_est_millions, yoy_growth_pct
+        FROM hyperscaler_capex
+        WHERE yoy_growth_pct IS NOT NULL
+        ORDER BY company, period_end DESC
+      `),
+
+      // AI cluster demand last 90 days
+      pool.query(`
+        SELECT
+          COALESCE(SUM(estimated_transceivers), 0) AS total_tx,
+          COUNT(*) AS cluster_count,
+          COUNT(*) FILTER (WHERE estimated_transceivers > 0) AS with_estimates
+        FROM ai_cluster_announcements
+        WHERE announced_date >= NOW() - INTERVAL '90 days'
+      `),
+
+      // eBay marketplace velocity
+      pool.query(`
+        SELECT DISTINCT ON (form_factor, speed_label)
+          marketplace, keyword, form_factor, speed_label,
+          sold_count_30d, active_listings, avg_sold_price
+        FROM marketplace_velocity
+        WHERE sold_count_30d IS NOT NULL
+        ORDER BY form_factor, speed_label, scraped_at DESC
+      `),
+
+      // Internal demand: fast-mover trend
+      pool.query(`
+        SELECT
+          COUNT(*) FILTER (WHERE velocity_class = 'fast_mover' AND demand_trend_pct > 0) AS fast_mover_pos,
+          COUNT(*) FILTER (WHERE velocity_class = 'fast_mover' AND demand_trend_pct < 0) AS fast_mover_neg,
+          AVG(demand_trend_pct) FILTER (WHERE velocity_class = 'fast_mover') AS avg_fast_trend,
+          AVG(demand_trend_pct) FILTER (WHERE velocity_class = 'regular')    AS avg_regular_trend
+        FROM flexoptix_internal_demand
+        WHERE is_internal = true
+      `),
+    ]);
+
+    // ── Pre-compute global signals ───────────────────────────────────────────
+    const capexYoyValues = capexRows.rows
+      .map((r) => parseFloat(r.yoy_growth_pct))
+      .filter((v) => !isNaN(v));
+    const capexYoyAvg = capexYoyValues.length
+      ? capexYoyValues.reduce((a, b) => a + b, 0) / capexYoyValues.length
+      : 0;
+
+    const aiData = aiRows.rows[0];
+    const totalAiTx = parseInt(aiData?.total_tx ?? "0") || 0;
+
+    const internalData = internalRows.rows[0];
+    const avgFastTrend = parseFloat(internalData?.avg_fast_trend ?? "0") || 0;
+
+    // Build price activity map: speedGbps+formFactor → ratio (30d/prior30d)
+    const priceMap = new Map<string, number>();
+    for (const row of priceRows.rows) {
+      const obs30 = parseInt(row.obs_30d) || 0;
+      const obsPrior = parseInt(row.obs_prior_30d) || 1;
+      const key = `${row.speed_gbps}__${row.form_factor ?? ""}`;
+      priceMap.set(key, obs30 / obsPrior);
+    }
+
+    // Build eBay map: speedLabel → sold_count_30d
+    const ebayMap = new Map<string, number>();
+    for (const row of ebayRows.rows) {
+      const speed = (row.speed_label ?? "").replace(/\s+/g, "");
+      ebayMap.set(speed, parseInt(row.sold_count_30d) || 0);
+    }
+
+    // ── Per-technology signals ───────────────────────────────────────────────
+    const technologies = hypeRows.rows.map((r) => {
+      const tech = TECH_SIGNAL_MAP.find((t) => t.label === r.technology);
+      const phase = (r.hype_phase ?? "innovation_trigger") as PhaseKey;
+      const hypeScore = parseInt(r.hype_score) || 0;
+
+      // Price activity ratio: average across matching speed+formFactor combos
+      let priceRatios: number[] = [];
+      if (tech) {
+        for (const ff of tech.formFactors) {
+          const key = `${tech.speedGbps}__${ff}`;
+          const ratio = priceMap.get(key);
+          if (ratio !== undefined) priceRatios.push(ratio);
+        }
+      }
+      const priceActivityRatio = priceRatios.length
+        ? priceRatios.reduce((a, b) => a + b, 0) / priceRatios.length
+        : 1;
+
+      // eBay velocity for this speed
+      const ebayVelocity = tech ? (ebayMap.get(tech.speedLabel) ?? 0) : 0;
+
+      // AI cluster: allocate demand proportionally for high-speed techs
+      const aiBoostTx = tech && tech.speedGbps >= 400 ? totalAiTx : 0;
+
+      // Compute composite score (0–100)
+      let score = hypeScore * 0.3;
+      const capexBoostPts = capexYoyAvg > 100 ? 18 : capexYoyAvg > 50 ? 12 : capexYoyAvg > 20 ? 5 : 0;
+      const priceBoostPts = priceActivityRatio > 1.3 ? 10 : priceActivityRatio > 1.0 ? 5 : -3;
+      const aiBoostPts    = aiBoostTx > 100000 ? 14 : aiBoostTx > 50000 ? 9 : aiBoostTx > 10000 ? 4 : 0;
+      const ebayBoostPts  = ebayVelocity > 200 ? 8 : ebayVelocity > 100 ? 5 : ebayVelocity > 50 ? 2 : 0;
+      const intlBoostPts  = avgFastTrend > 10 ? 6 : avgFastTrend > 0 ? 3 : avgFastTrend < -20 ? -5 : 0;
+
+      score += capexBoostPts + priceBoostPts + aiBoostPts + ebayBoostPts + intlBoostPts;
+      const marketSignalScore = Math.max(0, Math.min(100, Math.round(score)));
+
+      const rec = buildRecommendation(phase, marketSignalScore, capexYoyAvg, tech?.speedGbps ?? 0);
+
+      // Signal drivers list for tooltip
+      const drivers: string[] = [];
+      if (capexBoostPts > 0) drivers.push(`Hyperscaler CapEx +${capexYoyAvg.toFixed(0)}% YoY avg`);
+      if (priceActivityRatio > 1.1) drivers.push(`Price obs +${((priceActivityRatio - 1) * 100).toFixed(0)}% MoM activity`);
+      if (aiBoostTx > 0) drivers.push(`~${(aiBoostTx / 1000).toFixed(0)}k transceivers in AI cluster builds`);
+      if (ebayVelocity > 0) drivers.push(`${ebayVelocity} units sold on secondary market (30d)`);
+      if (intlBoostPts > 0) drivers.push(`Internal fast-movers trending ${avgFastTrend > 0 ? "+" : ""}${avgFastTrend.toFixed(1)}%`);
+
+      return {
+        technology: r.technology,
+        phase,
+        hypeScore,
+        aspCurrentUsd: r.asp_current_usd,
+        marketSignalScore,
+        recommendation: rec,
+        drivers,
+        speedGbps: tech?.speedGbps,
+        priceActivityRatio: Math.round(priceActivityRatio * 100) / 100,
+        ebayVelocity,
+        computedAt: r.computed_at,
+      };
+    });
+
+    // ── Global context ───────────────────────────────────────────────────────
+    const globalContext = {
+      hyperscalerCapex: capexRows.rows.map((r) => ({
+        company: r.company,
+        periodLabel: r.period_label,
+        capexMillions: parseFloat(r.capex_usd_millions),
+        dcCapexMillions: parseFloat(r.dc_capex_est_millions),
+        yoyGrowthPct: parseFloat(r.yoy_growth_pct),
+      })),
+      capexYoyAvg: Math.round(capexYoyAvg),
+      capexBoom: capexYoyAvg > 50,
+      totalAiClusterTx90d: totalAiTx,
+      aiClusterCount90d: parseInt(aiData?.cluster_count ?? "0") || 0,
+      internalFastMoverTrend: Math.round(avgFastTrend * 10) / 10,
+    };
+
+    res.json({ success: true, technologies, globalContext, computed_at: new Date().toISOString() });
+  } catch (err) {
+    console.error("Market signals error:", err);
+    res.status(500).json({ success: false, error: "Failed to compute market signals" });
+  }
+});
+
 // GET /api/hype-cycle/analysis — Bass-fitted results from DB (hype_cycle_analysis table)
 hypeCycleRouter.get("/analysis", async (_req: Request, res: Response) => {
   try {

packages/api/src/routes/procurement.ts

@@ -424,3 +424,79 @@ procurementRouter.get("/internal-demand", async (req: Request, res: Response) =>
     res.status(500).json({ error: "Internal server error" });
   }
 });
+
+// ─────────────────────────────────────────────────────────────────────────────
+// GET /api/procurement/hyperscaler-capex
+// Hyperscaler quarterly CapEx from SEC filings — demand context for transceivers
+// ─────────────────────────────────────────────────────────────────────────────
+procurementRouter.get("/hyperscaler-capex", async (_req: Request, res: Response) => {
+  try {
+    const result = await pool.query(`
+      SELECT
+        company, period_label, period_end,
+        capex_usd_millions, dc_capex_est_millions,
+        yoy_growth_pct, filing_type, source_url
+      FROM hyperscaler_capex
+      ORDER BY period_end DESC, capex_usd_millions DESC
+    `);
+
+    const summaryResult = await pool.query(`
+      SELECT
+        company,
+        MAX(period_end) AS latest_period_end,
+        MAX(period_label) AS latest_period,
+        (ARRAY_AGG(capex_usd_millions ORDER BY period_end DESC))[1] AS latest_capex,
+        (ARRAY_AGG(dc_capex_est_millions ORDER BY period_end DESC))[1] AS latest_dc_capex,
+        (ARRAY_AGG(yoy_growth_pct ORDER BY period_end DESC))[1] AS latest_yoy_growth,
+        AVG(yoy_growth_pct) FILTER (WHERE period_end >= NOW() - INTERVAL '365 days') AS avg_yoy_12m
+      FROM hyperscaler_capex
+      GROUP BY company
+      ORDER BY latest_capex DESC NULLS LAST
+    `);
+
+    res.json({
+      data: result.rows,
+      summary: summaryResult.rows,
+    });
+  } catch (err) {
+    console.error("Hyperscaler capex error:", err);
+    res.status(500).json({ error: "Internal server error" });
+  }
+});
+
+// ─────────────────────────────────────────────────────────────────────────────
+// GET /api/procurement/marketplace-velocity
+// Secondary market (eBay) sell-through as demand signal
+// ─────────────────────────────────────────────────────────────────────────────
+procurementRouter.get("/marketplace-velocity", async (_req: Request, res: Response) => {
+  try {
+    const result = await pool.query(`
+      SELECT DISTINCT ON (form_factor, speed_label)
+        marketplace, keyword, form_factor, speed_label,
+        sold_count_30d, active_listings, avg_sold_price,
+        min_price, max_price, currency, scraped_at
+      FROM marketplace_velocity
+      ORDER BY form_factor, speed_label, scraped_at DESC
+    `);
+
+    const hotResult = await pool.query(`
+      SELECT DISTINCT ON (form_factor, speed_label)
+        marketplace, form_factor, speed_label,
+        sold_count_30d, active_listings, avg_sold_price
+      FROM marketplace_velocity
+      WHERE sold_count_30d > 0
+      ORDER BY form_factor, speed_label, scraped_at DESC
+    `);
+
+    res.json({
+      data: result.rows,
+      hot: hotResult.rows.sort(
+        (a: { sold_count_30d: string }, b: { sold_count_30d: string }) =>
+          parseInt(b.sold_count_30d) - parseInt(a.sold_count_30d)
+      ),
+    });
+  } catch (err) {
+    console.error("Marketplace velocity error:", err);
+    res.status(500).json({ error: "Internal server error" });
+  }
+});

packages/dashboard/hot-topics.js

@@ -31,13 +31,28 @@
     }
     blogPipelineRunning = true;
 
+    // Fetch the current active model name so we never show a stale hardcoded version.
+    var initialModelLabel = (window._activeFoBlogModel) || 'FO_BlogLLM';
+    fetch(API + '/api/blog/llm/status', { headers: authHeaders() })
+      .then(function(r) { return r.json(); })
+      .then(function(data) {
+        var m = data && data.llm && data.llm.model;
+        if (m) {
+          window._activeFoBlogModel = m;
+          var s = document.getElementById('bp-step');
+          if (s && s.textContent.indexOf('Connecting to FO_BlogLLM') === 0) {
+            s.textContent = 'Connecting to FO_BlogLLM (' + m + ')';
+          }
+        }
+      }).catch(function() {});
+
     var pipelineEl = document.getElementById('blog-pipeline-status');
     if (pipelineEl) {
       pipelineEl.innerHTML =
         '<div style="background:linear-gradient(135deg,#1a1a1a,#2a2a2a);color:white;padding:2rem;border-radius:12px;text-align:center;margin-bottom:1rem">' +
         '<div style="font-size:1.4rem;font-weight:700;margin-bottom:1rem">Generating Blog with AI...</div>' +
         '<div id="bp-status" style="font-size:1rem;color:#FF8100;margin-bottom:0.5rem">Starting 10-step Flexoptix Style pipeline...</div>' +
-        '<div id="bp-step" style="font-size:0.85rem;color:#aaa">Connecting to FO_BlogLLM (fo-blog-v7)</div>' +
+        '<div id="bp-step" style="font-size:0.85rem;color:#aaa">Connecting to ' + initialModelLabel + '</div>' +
         '<div style="margin-top:1.5rem;background:#333;border-radius:8px;height:8px;overflow:hidden">' +
         '<div id="bp-bar" style="width:2%;height:100%;background:#FF8100;transition:width 0.5s ease"></div></div>' +
         '<div id="bp-pct" style="font-size:0.8rem;color:#666;margin-top:0.5rem">0%</div>' +
@@ -137,7 +152,7 @@
             if (bar) bar.style.width = prog.pct + '%';
             if (pct) pct.textContent = prog.pct + '%';
             if (status) { status.style.color = '#FF8100'; status.textContent = prog.label || ('Step ' + prog.step + '/10'); }
-            if (step) step.textContent = 'Step ' + prog.step + '/10 · fo-blog-v7 via adapter bridge';
+            if (step) step.textContent = 'Step ' + prog.step + '/10 · ' + (window._activeFoBlogModel || 'fo-blog-v10') + ' via adapter bridge';
           } else {
             _stallCount++;
             // After 5 consecutive non-running polls (~40s), show stall warning

packages/dashboard/index.html

@@ -916,6 +916,11 @@
       </div>
       <div id="hype-svg-container"></div>
     </div>
+    <!-- Market Context cards — loaded dynamically from market-signals API -->
+    <div id="hype-market-context" style="display:grid;gap:0.75rem;grid-template-columns:repeat(auto-fill,minmax(210px,1fr));margin-bottom:1.25rem">
+      <div class="loading pulse" style="grid-column:1/-1;padding:0.75rem">Loading market signals…</div>
+    </div>
+
     <div class="card mt">
       <div class="table-wrap">
         <table>
@@ -923,9 +928,8 @@
             <th>Technology<span class="sort-arrow"></span></th>
             <th class="tip" data-tip="Current phase in the technology adoption lifecycle.">Phase<span class="sort-arrow"></span></th>
             <th class="tip" data-tip="Position on the hype curve (0-100%).">Position<span class="sort-arrow"></span></th>
-            <th class="tip" data-tip="Cumulative market adoption based on Norton-Bass diffusion model. 0-100% of total addressable market. MODELL — keine echten Marktdaten.">Adoption <span style="font-size:0.58rem;color:#6366f1;font-weight:600">[M]</span><span class="sort-arrow"></span></th>
+            <th class="tip" data-tip="Composite Market Signal Score (0–100). Blends hype score + hyperscaler capex trend + price observation activity + AI cluster demand + eBay secondary market + internal demand velocity. Higher = stronger real-world demand signal.">Market Signal <span style="font-size:0.58rem;color:#16a34a;font-weight:600">● LIVE</span><span class="sort-arrow"></span></th>
-            <th class="tip" data-tip="Estimated year of peak hype — Norton-Bass model estimate, not real forecast data.">Peak <span style="font-size:0.58rem;color:#6366f1;font-weight:600">[M]</span><span class="sort-arrow"></span></th>
+            <th class="tip" data-tip="Data-driven buy recommendation based on hype phase + multi-source market signals.">Recommendation<span class="sort-arrow"></span></th>
-            <th class="tip" data-tip="Years until mainstream — Norton-Bass model estimate.">To Plateau <span style="font-size:0.58rem;color:#6366f1;font-weight:600">[M]</span><span class="sort-arrow"></span></th>
             <th class="tip" data-tip="Current OEM ASP in USD — from Mouser/market data.">OEM ASP<span class="sort-arrow"></span></th>
             <th class="tip" data-tip="Bass model goodness-of-fit (R²). Higher = more reliable forecast.">R²<span class="sort-arrow"></span></th>
           </tr></thead>
@@ -934,7 +938,7 @@
       </div>
     </div>
     <div style="margin-top:0.5rem;font-size:0.7rem;color:var(--text-dim);display:flex;gap:1.5rem;flex-wrap:wrap">
-      <span><span style="color:#6366f1;font-weight:700">[M]</span> = MODELL — Norton-Bass mathematische Schätzung, keine echten Marktdaten</span>
+      <span><span style="color:#16a34a;font-weight:700">● LIVE</span> = Real market data: hyperscaler capex + price activity + AI cluster demand + eBay velocity + internal demand</span>
       <span>OEM ASP = real (Mouser/Marktdaten)</span>
     </div>
     <div class="card mt" style="border-left:3px solid var(--cyan)">
@@ -974,6 +978,31 @@
         <div class="loading pulse" style="padding:2.5rem;text-align:center">Loading sourcing data…</div>
       </div>
     </div>
+
+    <!-- Hyperscaler CapEx panel -->
+    <div class="card mt" style="border-left:3px solid #2563eb">
+      <div style="display:flex;justify-content:space-between;align-items:center;margin-bottom:1rem;flex-wrap:wrap;gap:0.5rem">
+        <div>
+          <div style="font-size:1rem;font-weight:700;color:var(--text-bright)">Hyperscaler CapEx <span style="font-size:0.72rem;font-weight:700;background:#2563eb22;color:#2563eb;border:1px solid #2563eb44;border-radius:3px;padding:1px 6px">SEC Filings</span></div>
+          <div style="font-size:0.72rem;color:var(--text-dim);margin-top:2px">Quarterly infrastructure spend — primary demand driver for high-speed transceivers</div>
+        </div>
+        <div id="capex-avg-badge" style="font-size:0.82rem;font-weight:700;padding:4px 12px;background:#16a34a11;border:1px solid #16a34a33;border-radius:6px;color:#16a34a"></div>
+      </div>
+      <div id="capex-table-container">
+        <div class="loading pulse">Loading capex data…</div>
+      </div>
+    </div>
+
+    <!-- eBay Marketplace panel -->
+    <div class="card mt" style="border-left:3px solid #f97316">
+      <div style="margin-bottom:0.75rem">
+        <div style="font-size:1rem;font-weight:700;color:var(--text-bright)">eBay Secondary Market <span style="font-size:0.72rem;font-weight:700;background:#f9731622;color:#f97316;border:1px solid #f9731644;border-radius:3px;padding:1px 6px">Demand Signal</span></div>
+        <div style="font-size:0.72rem;color:var(--text-dim);margin-top:2px">Secondary market sell-through rates — real end-buyer demand by form factor</div>
+      </div>
+      <div id="ebay-velocity-container">
+        <div class="loading pulse">Loading marketplace data…</div>
+      </div>
+    </div>
   </div>
 
   </div>
@@ -1568,6 +1597,7 @@
       <button onclick="showProcSection('abc')" id="proc-btn-abc" class="proc-btn">ABC Classes</button>
       <button onclick="showProcSection('demand')" id="proc-btn-demand" class="proc-btn" style="background:rgba(22,163,74,0.08);border-color:rgba(22,163,74,0.3);color:#16a34a">📦 Internal Demand</button>
       <button onclick="showProcSection('ai-clusters')" id="proc-btn-ai-clusters" class="proc-btn" style="background:rgba(124,92,252,0.08);border-color:rgba(124,92,252,0.3);color:#7c5cfc">🤖 AI Clusters</button>
+      <button onclick="showProcSection('marketplace')" id="proc-btn-marketplace" class="proc-btn" style="background:rgba(249,115,22,0.08);border-color:rgba(249,115,22,0.3);color:#f97316">🛒 eBay Market</button>
       <button onclick="showProcSection('market')" id="proc-btn-market" class="proc-btn">Market Intelligence</button>
       <button onclick="showProcSection('lifecycle')" id="proc-btn-lifecycle" class="proc-btn">Lifecycle Events</button>
       <div style="flex:1"></div>
@@ -1658,6 +1688,18 @@
       </div>
     </div>
 
+    <!-- eBay Marketplace section -->
+    <div id="proc-section-marketplace" style="display:none">
+      <div style="display:flex;gap:0.5rem;margin-bottom:1rem;align-items:center;flex-wrap:wrap">
+        <span style="font-size:0.8rem;font-weight:700;color:var(--text)">Secondary Market Demand (eBay)</span>
+        <div style="flex:1"></div>
+        <span style="font-size:0.72rem;color:var(--text-dim)">Higher sell-through = real end-buyer demand, price floor signal</span>
+      </div>
+      <div id="proc-marketplace-grid" style="display:grid;gap:0.75rem;grid-template-columns:repeat(auto-fill,minmax(280px,1fr))">
+        <div class="loading pulse">Loading marketplace data...</div>
+      </div>
+    </div>
+
     <!-- AI Clusters section -->
     <div id="proc-section-ai-clusters" style="display:none">
       <div style="display:flex;gap:1rem;margin-bottom:1.25rem;flex-wrap:wrap" id="ai-cluster-stats">
@@ -3336,17 +3378,42 @@ async function loadHypeCycle() {
     dot.addEventListener('click', function() { openHypeDetail(this.getAttribute('data-tech')); });
   });
 
+  // Load market signals in parallel and enrich the table
+  var marketSignals = {};
+  var globalCtx = {};
+  try {
+    var msData = await api('/api/hype-cycle/market-signals');
+    if (msData.success && msData.technologies) {
+      msData.technologies.forEach(function(ms) { marketSignals[ms.technology] = ms; });
+      globalCtx = msData.globalContext || {};
+      renderHypeMarketContext(globalCtx, msData.technologies);
+    }
+  } catch(e) {
+    var ctxEl = el('hype-market-context');
+    if (ctxEl) ctxEl.innerHTML = '';
+  }
+
   buildDOM(el('hype-table'), techs.map(function(t) {
     var color = PC[t.phase] || '#8888a4';
-    // FIX: adoptionPct is already a percentage (0-100), do NOT multiply by 100
+    var ms = marketSignals[t.technology];
-    var adoptionDisplay = (t.adoptionPct != null ? t.adoptionPct : 0) + '%';
+    var signalHtml = '—';
+    var recHtml = '—';
+    if (ms) {
+      var sc = ms.marketSignalScore;
+      var scColor = sc >= 70 ? '#16a34a' : sc >= 50 ? '#ca8a04' : sc >= 30 ? '#f97316' : '#94a3b8';
+      var driversTip = ms.drivers && ms.drivers.length ? ms.drivers.join(' · ') : 'No signal data';
+      signalHtml = '<span class="tip" data-tip="' + esc(driversTip) + '" style="font-weight:700;color:' + scColor + ';font-family:monospace">' + sc + '</span>'
+        + '<div style="width:60px;height:4px;background:var(--surface2);border-radius:2px;margin-top:3px;display:inline-block;vertical-align:middle;margin-left:6px"><div style="width:' + sc + '%;height:100%;background:' + scColor + ';border-radius:2px"></div></div>';
+      if (ms.recommendation) {
+        recHtml = '<span class="tip" data-tip="' + esc(ms.recommendation.detail) + '" style="font-size:0.72rem;font-weight:700;white-space:nowrap;color:' + ms.recommendation.color + '">' + esc(ms.recommendation.label) + '</span>';
+      }
+    }
     return '<tr class="clickable" data-tech="' + esc(t.technology) + '">'
       + '<td style="font-weight:600;color:var(--text-bright)">' + esc(t.technology) + '</td>'
       + '<td><span class="b tip" data-tip="' + esc(PHASE_DESC[t.phase] || '') + '" style="background:' + color + '18;color:' + color + ';border:1px solid ' + color + '33">' + esc(t.phase) + '</span></td>'
       + '<td><div class="hype-bar"><div class="hype-fill" style="width:' + t.positionPct + '%;background:' + color + '"></div></div></td>'
-      + '<td class="mono tip" data-tip="' + esc(HYPE_TIPS['Adoption']) + '">' + adoptionDisplay + '</td>'
+      + '<td>' + signalHtml + '</td>'
-      + '<td class="mono">' + esc(t.peakYear || '—') + '</td>'
+      + '<td>' + recHtml + '</td>'
-      + '<td class="mono">' + (t.yearsToPlateauFromNow != null ? t.yearsToPlateauFromNow + 'y' : '—') + '</td>'
       + '<td class="mono">' + (t.aspCurrentUsd != null ? '$' + Number(t.aspCurrentUsd).toLocaleString() : '—') + '</td>'
       + '<td class="mono">' + (t.rSquared != null ? Number(t.rSquared).toFixed(2) : '—') + '</td>'
       + '</tr>';
@@ -3369,6 +3436,9 @@ async function loadHypeCycle() {
 
   // Render Sourcing Hype Cycle
   loadSourcingHypeCycle();
+
+  // Render Hyperscaler CapEx + eBay panels
+  loadHypeCapexAndEbay();
 }
 
 // ── SOURCING HYPE CYCLE ──────────────────────────────────────────────────────
@@ -3395,7 +3465,8 @@ async function loadSourcingHypeCycle() {
   var meta      = document.getElementById('sourcing-hype-meta');
   if (!container) return;
 
-  var seedData = [
+  // Static fallback (only used if API fails)
+  var seedFallback = [
     { label: 'SFP+ 10G',     speed_gbps: 10,  form_factor: 'SFP+',    obs: 1402, avg_price: 94.86  },
     { label: 'SFP 1G',       speed_gbps: 1,   form_factor: 'SFP',     obs: 1103, avg_price: 35.11  },
     { label: 'QSFP28 100G',  speed_gbps: 100, form_factor: 'QSFP28',  obs: 369,  avg_price: 409.46 },
@@ -3406,15 +3477,29 @@ async function loadSourcingHypeCycle() {
     { label: 'QSFP-DD 800G', speed_gbps: 800, form_factor: 'QSFP-DD', obs: 40,   avg_price: 749.08 }
   ];
 
-  var items = seedData;
+  var items = seedFallback;
+  // Try real price observation counts per form_factor+speed_gbps (live DB data)
   try {
-    var d = await api('/api/procurement/signals?limit=30');
+    var priceCountRes = await api('/api/procurement/signals?limit=200');
-    if (d && d.signals && d.signals.length > 0) {
+    var sigItems = priceCountRes.data || [];
-      d.signals.forEach(function(sig) {
+    if (sigItems.length > 0) {
-        var key = (sig.form_factor || '') + String(sig.speed_gbps || '');
+      // Build a map from form_factor+speed_gbps → {obs, avg_price}
-        var ex = items.find(function(it) { return (it.form_factor || '') + String(it.speed_gbps || '') === key; });
+      var realMap = {};
-        if (ex && sig.observation_count > 0) ex.obs = sig.observation_count;
+      sigItems.forEach(function(row) {
+        var key = (row.form_factor || '') + '|' + String(row.speed_gbps || '');
+        if (!realMap[key] || row.observation_count > realMap[key].obs) {
+          realMap[key] = { obs: row.observation_count || 0, avg_price: parseFloat(row.avg_price) || 0 };
+        }
       });
+      // Merge into items
+      items = items.map(function(it) {
+        var key = (it.form_factor || '') + '|' + String(it.speed_gbps || '');
+        if (realMap[key] && realMap[key].obs > 0) {
+          return Object.assign({}, it, { obs: realMap[key].obs, avg_price: realMap[key].avg_price || it.avg_price });
+        }
+        return it;
+      });
+      if (meta) meta.setAttribute('data-source', 'live');
     }
   } catch(e) {}
 
@@ -6769,7 +6854,7 @@ var procAiClustersData = [];
 var procAiClustersMinTx = 0;
 
 function showProcSection(name) {
-  ['signals','abc','demand','ai-clusters','market','lifecycle'].forEach(function(s) {
+  ['signals','abc','demand','marketplace','ai-clusters','market','lifecycle'].forEach(function(s) {
     var sec = el('proc-section-' + s);
     var btn = el('proc-btn-' + s);
     if (sec) sec.style.display = s === name ? '' : 'none';
@@ -6778,6 +6863,7 @@ function showProcSection(name) {
   // Lazy-load on first visit
   if (name === 'demand' && procDemandData.length === 0) loadInternalDemand();
   if (name === 'ai-clusters' && procAiClustersData.length === 0) loadAiClusters();
+  if (name === 'marketplace' && !el('proc-marketplace-grid').querySelector('.card')) loadProcMarketplace();
 }
 
 async function loadProcurement() {
@@ -7144,6 +7230,167 @@ function filterAiClusters(minTx) {
   }).join('');
 }
 
+// ─── MARKET SIGNALS — HYPE CYCLE ENRICHMENT ──────────────────────────────────
+
+function renderHypeMarketContext(ctx, technologies) {
+  var el2 = el('hype-market-context');
+  if (!el2) return;
+  var capexYoy = ctx.capexYoyAvg || 0;
+  var capexColor = capexYoy > 80 ? '#16a34a' : capexYoy > 40 ? '#ca8a04' : '#64748b';
+  var totalAiTx = ctx.totalAiClusterTx90d || 0;
+  var fastTrend = ctx.internalFastMoverTrend || 0;
+
+  // Summary stat cards
+  var cards = [
+    statCard('💰', capexYoy.toFixed(0) + '% YoY avg', 'Hyperscaler CapEx',
+      (ctx.capexBoom ? '🚀 Boom — above +50% threshold' : 'Moderate growth') + ' (' + (ctx.hyperscalerCapex || []).map(function(c) { return c.company; }).join(', ') + ')'),
+    statCard('🤖', totalAiTx > 0 ? '~' + (totalAiTx / 1000).toFixed(0) + 'k Tx' : '—', 'AI Cluster Demand (90d)',
+      ctx.aiClusterCount90d + ' cluster builds with transceiver demand estimates'),
+    statCard('📈', fastTrend > 0 ? '+' + fastTrend.toFixed(1) + '%' : fastTrend.toFixed(1) + '%', 'Internal Fast-Mover Trend',
+      'Average demand trend across ' + (ctx.hyperscalerCapex || []).length + ' fast-moving SKUs'),
+  ];
+
+  // Per-technology recommendation summary
+  var recCards = (technologies || []).map(function(ms) {
+    if (!ms.recommendation) return '';
+    return '<div style="background:var(--surface2);border:1px solid var(--border);border-radius:8px;padding:0.6rem 0.75rem;min-width:160px">'
+      + '<div style="font-size:0.75rem;font-weight:700;color:var(--text)">' + esc(ms.technology) + '</div>'
+      + '<div style="font-size:0.72rem;font-weight:700;color:' + ms.recommendation.color + ';margin-top:0.2rem">' + esc(ms.recommendation.label) + '</div>'
+      + '<div style="font-size:0.65rem;color:var(--text-dim);margin-top:0.15rem;line-height:1.4">' + esc(ms.recommendation.detail.substring(0, 80)) + (ms.recommendation.detail.length > 80 ? '…' : '') + '</div>'
+      + '<div style="margin-top:0.4rem;display:flex;align-items:center;gap:0.4rem">'
+      + '<div style="flex:1;height:3px;background:var(--border);border-radius:2px"><div style="width:' + ms.marketSignalScore + '%;height:100%;background:' + ms.recommendation.color + ';border-radius:2px"></div></div>'
+      + '<span style="font-size:0.65rem;font-family:monospace;font-weight:700;color:' + ms.recommendation.color + '">' + ms.marketSignalScore + '</span>'
+      + '</div></div>';
+  }).join('');
+
+  el2.innerHTML = '<div style="display:flex;gap:0.75rem;flex-wrap:wrap;margin-bottom:0.75rem">' + cards.join('') + '</div>'
+    + '<div style="font-size:0.72rem;font-weight:700;color:var(--text-dim);margin-bottom:0.5rem;letter-spacing:0.05em;text-transform:uppercase">Recommendations per Technology</div>'
+    + '<div style="display:flex;gap:0.75rem;flex-wrap:wrap">' + recCards + '</div>';
+}
+
+// ─── CAPEX TABLE — in Hype Cycle tab ─────────────────────────────────────────
+
+async function loadHypeCapexAndEbay() {
+  try {
+    var d = await api('/api/procurement/hyperscaler-capex');
+    renderCapexTable(d.data || [], d.summary || []);
+  } catch(e) {
+    var el2 = el('capex-table-container');
+    if (el2) el2.innerHTML = '<div style="color:var(--text-dim);padding:0.5rem">Could not load capex data.</div>';
+  }
+  try {
+    var e = await api('/api/procurement/marketplace-velocity');
+    renderEbayVelocity(e.hot || e.data || []);
+    renderProcMarketplace(e.hot || e.data || []);
+  } catch(e2) {}
+}
+
+function renderCapexTable(rows, summary) {
+  var badge = el('capex-avg-badge');
+  if (badge && summary.length) {
+    var avgYoy = summary.reduce(function(s, r) { return s + (parseFloat(r.latest_yoy_growth) || 0); }, 0) / summary.length;
+    badge.textContent = 'Avg YoY: +' + avgYoy.toFixed(0) + '%  ·  ' + summary.length + ' hyperscalers';
+    badge.style.color = avgYoy > 80 ? '#16a34a' : avgYoy > 40 ? '#ca8a04' : '#2563eb';
+    badge.style.background = avgYoy > 80 ? '#16a34a11' : '#2563eb11';
+    badge.style.borderColor = avgYoy > 80 ? '#16a34a33' : '#2563eb33';
+  }
+
+  var container = el('capex-table-container');
+  if (!container || !rows.length) return;
+
+  var html = '<div style="overflow-x:auto"><table style="width:100%;border-collapse:collapse;font-size:0.8rem">'
+    + '<thead><tr style="border-bottom:2px solid var(--border);color:var(--text-dim);font-size:0.7rem;font-weight:700;text-transform:uppercase">'
+    + '<th style="text-align:left;padding:8px 6px">Company</th>'
+    + '<th style="text-align:left;padding:8px 6px">Period</th>'
+    + '<th class="tip" data-tip="Total quarterly CapEx in millions USD (all infra)" style="text-align:right;padding:8px 6px">CapEx ($M)</th>'
+    + '<th class="tip" data-tip="Estimated DC portion (≈55% of total CapEx based on analyst estimates)" style="text-align:right;padding:8px 6px">DC Est. ($M)</th>'
+    + '<th class="tip" data-tip="Year-over-year growth of total CapEx vs same quarter prior year" style="text-align:right;padding:8px 6px">YoY %</th>'
+    + '<th style="text-align:left;padding:8px 6px">Filing</th>'
+    + '</tr></thead><tbody>';
+
+  rows.forEach(function(r) {
+    var yoy = parseFloat(r.yoy_growth_pct);
+    var yoyColor = yoy > 80 ? '#16a34a' : yoy > 40 ? '#ca8a04' : yoy > 0 ? '#2563eb' : '#ef4444';
+    var yoyStr = isNaN(yoy) ? '—' : (yoy > 0 ? '+' : '') + yoy.toFixed(1) + '%';
+    html += '<tr style="border-bottom:1px solid var(--border)">'
+      + '<td style="padding:7px 6px;font-weight:700;text-transform:capitalize">' + esc(r.company) + '</td>'
+      + '<td style="padding:7px 6px;color:var(--text-dim)">' + esc(r.period_label) + '</td>'
+      + '<td style="padding:7px 6px;text-align:right;font-family:monospace;font-weight:600">$' + parseFloat(r.capex_usd_millions).toLocaleString(undefined, {maximumFractionDigits: 0}) + '</td>'
+      + '<td style="padding:7px 6px;text-align:right;font-family:monospace;color:var(--text-dim)">$' + parseFloat(r.dc_capex_est_millions).toLocaleString(undefined, {maximumFractionDigits: 0}) + '</td>'
+      + '<td style="padding:7px 6px;text-align:right;font-weight:700;color:' + yoyColor + '">' + yoyStr + '</td>'
+      + '<td style="padding:7px 6px;font-size:0.68rem;color:var(--text-dim)">' + esc(r.filing_type || '—') + '</td>'
+      + '</tr>';
+  });
+
+  container.innerHTML = html + '</tbody></table></div>';
+}
+
+function renderEbayVelocity(rows) {
+  var container = el('ebay-velocity-container');
+  if (!container || !rows.length) {
+    if (container) container.innerHTML = '<div style="color:var(--text-dim);padding:0.5rem">No marketplace data.</div>';
+    return;
+  }
+
+  var html = '<div style="display:grid;gap:0.75rem;grid-template-columns:repeat(auto-fill,minmax(220px,1fr))">';
+  rows.forEach(function(r) {
+    var sold = parseInt(r.sold_count_30d) || 0;
+    var listings = parseInt(r.active_listings) || 0;
+    var sellThrough = listings > 0 ? ((sold / listings) * 100).toFixed(1) + '%' : '—';
+    var demandColor = sold > 200 ? '#16a34a' : sold > 100 ? '#ca8a04' : sold > 50 ? '#2563eb' : 'var(--text-dim)';
+    var avgPrice = parseFloat(r.avg_sold_price);
+    html += '<div style="background:var(--surface2);border:1px solid var(--border);border-radius:8px;padding:0.8rem">'
+      + '<div style="font-size:0.82rem;font-weight:700">' + esc(r.form_factor || r.keyword || '?') + ' ' + esc(r.speed_label || '') + '</div>'
+      + '<div style="display:flex;align-items:baseline;gap:0.4rem;margin-top:0.4rem">'
+      + '<span style="font-size:1.5rem;font-weight:800;color:' + demandColor + '">' + sold.toLocaleString() + '</span>'
+      + '<span style="font-size:0.72rem;color:var(--text-dim)">sold / 30d</span>'
+      + '</div>'
+      + '<div style="font-size:0.72rem;color:var(--text-dim);margin-top:0.2rem">'
+      + listings.toLocaleString() + ' active listings · sell-through ' + sellThrough
+      + '</div>'
+      + (avgPrice > 0 ? '<div style="font-size:0.72rem;color:var(--text-dim)">Avg sold: $' + avgPrice.toLocaleString(undefined,{maximumFractionDigits:0}) + '</div>' : '')
+      + '</div>';
+  });
+  container.innerHTML = html + '</div>';
+}
+
+// ─── PROC MARKETPLACE SECTION ─────────────────────────────────────────────────
+
+async function loadProcMarketplace() {
+  var grid = el('proc-marketplace-grid');
+  try {
+    var d = await api('/api/procurement/marketplace-velocity');
+    renderProcMarketplace(d.hot || d.data || []);
+  } catch(e) {
+    if (grid) grid.innerHTML = '<div style="color:var(--text-dim);padding:1rem">Could not load marketplace data.</div>';
+  }
+}
+
+function renderProcMarketplace(rows) {
+  var grid = el('proc-marketplace-grid');
+  if (!grid) return;
+  if (!rows.length) { grid.innerHTML = '<div style="color:var(--text-dim)">No data yet.</div>'; return; }
+  grid.innerHTML = rows.map(function(r) {
+    var sold = parseInt(r.sold_count_30d) || 0;
+    var listings = parseInt(r.active_listings) || 0;
+    var demandColor = sold > 200 ? '#16a34a' : sold > 100 ? '#ca8a04' : sold > 50 ? '#2563eb' : 'var(--text-dim)';
+    var borderColor = sold > 200 ? '#16a34a' : sold > 100 ? '#ca8a04' : 'var(--border)';
+    var avgP = parseFloat(r.avg_sold_price);
+    return '<div class="card" style="border-left:3px solid ' + borderColor + '">'
+      + '<div style="font-weight:700;font-size:0.9rem">' + esc(r.form_factor || '?') + ' ' + esc(r.speed_label || '') + '</div>'
+      + '<div style="font-size:0.72rem;color:var(--text-dim);margin-bottom:0.5rem">' + esc(r.keyword || '') + ' · ' + esc(r.marketplace || 'eBay') + '</div>'
+      + '<div style="display:flex;align-items:baseline;gap:0.5rem">'
+      + '<span style="font-size:1.8rem;font-weight:800;color:' + demandColor + '">' + sold.toLocaleString() + '</span>'
+      + '<span style="font-size:0.75rem;color:var(--text-dim)">sold / 30d</span>'
+      + '</div>'
+      + '<div style="font-size:0.72rem;color:var(--text-dim);margin-top:0.25rem">'
+      + listings.toLocaleString() + ' active listings'
+      + (avgP > 0 ? ' · avg $' + avgP.toLocaleString(undefined,{maximumFractionDigits:0}) : '')
+      + '</div>'
+      + '</div>';
+  }).join('');
+}
+
 // INIT
 loadOverview();
 loadChangelog();