@@ -63,7 +63,7 @@ Returns `{ chartData: InferenceData[][], hardwareConfig: HardwareConfig }`.
 - `tokenRevenuePerGpuHour` uses cache-aware token sale pricing. `buildDerivedChartFields` emits a normalized placeholder; `applyTokenRevenuePricing` applies the selected normalized or OpenRouter prices to uncached input, cached input, and output. Revenue is the priced streams per GPU-second scaled to one hour. Normalized pricing uses $1/M for uncached input and output and $0.10/M for cached input. OpenRouter uses its published cache-read price when available, otherwise 10% of the prompt price. Agentic cache share uses the same measured GPU-plus-external-or-CPU rule as Fleet Lifecycle. Aggregate rows use their measured input/output split; disaggregated rows derive a like-for-like split from fixed ISL:OSL or measured agentic prompt:generation totals because their raw input/output rates use different prefill/decode GPU denominators. Historical Trends splines total throughput, token share, and cache hit independently before applying these prices; a partly measured cache frontier receives no cache discount.
 - `tpPerMw` — `(tputPerGpu * 1000) / hardwarePower` (GPU power in kW, result in tok/s/MW).
 - Cost-per-million fields — GPU hourly cost divided by tokens-per-hour (in millions): `costh` / `costr` for owning-at-large-hyperscaler-volume / 3-year-rental pricing respectively. Three token variants exist: combined (`costh`/`costr`), output-only (`costhOutput`/`costrOutput`), and input-only (`costhi`/`costri`). The former Neocloud tier (`costn`/`costni`/`costnOutput`) was removed; its share-link keys alias to the hyperscaler-volume metric.
-- Infrastructure purchasing-power fields divide tokens per hour by the same hourly costs. Total (`tokensPerDollarH`/`R`), output-only (`outputTokensPerDollarH`/`R`), and input-only (`inputTokensPerDollarH`/`R`) are separate USD Y-axis metrics. The former ¥-priced `tokensPerRmb*` axes and the Neocloud `*PerDollarN` axes were removed; their share-link keys alias to the matching hyperscaler-volume $ metric. The dashboard defaults to the hyperscaler-volume total-tokens-per-dollar axis (`y_tokensPerDollarH`). Links created for the removed API-price `i_metric=y_tokensPerDollar` axis resolve to the same hyperscaler-volume variant.
+- Infrastructure purchasing-power fields divide tokens per hour by the same hourly costs. Total (`tokensPerDollarH`/`R`), output-only (`outputTokensPerDollarH`/`R`), and input-only (`inputTokensPerDollarH`/`R`) are separate USD Y-axis metrics. The former ¥-priced `tokensPerRmb*` axes and the Neocloud `*PerDollarN` axes were removed; their share-link keys alias to the matching hyperscaler-volume $ metric. The dashboard defaults to the hyperscaler-volume total-tokens-per-dollar axis (`y_tokensPerDollarH`); with a measured-power tier filter (`i_power`) it defaults to the measured power scatter (`y_measuredAvgPower`) instead. Links created for the removed API-price `i_metric=y_tokensPerDollar` axis resolve to the same hyperscaler-volume variant.
 - Energy fields — `jTotal` / `jOutput` / `jInput`: `(hardwarePower * 1000) / tputPerGpu` (Joules per token, where power in kW is converted to W). These divide one GPU's all-in power by that GPU's reported throughput; for disaggregated rows `output_tput_per_gpu` is per decode GPU, so `jOutput` ignores the prefill pool. Its semantics are frozen; the whole-deployment view lives in the power boundaries below.
 - Power-boundary fields — see [Power boundaries](#power-boundaries).
 
 
@@ -113,6 +113,7 @@ them.
 Axis and presentation state:
 
 - selected x-axis and y-axis metrics, percentile, and effective x-axis mode
+- the y-axis default depends on the power tier quick filter: `defaultYAxisMetric` in `url-state.ts` opens a link that has `i_power` but no `i_metric` on `y_measuredAvgPower`, and `InferenceContext` omits `i_metric` from share links only when the metric equals that contextual default (or the metric a route seeds, as on the compare pages), so an explicit `i_metric` always wins. Turning the tier filter on from empty moves a metric outside the measured family to `y_measuredAvgPower`; tier edits after that, clearing the filter, and restoring `i_power` from a link leave the metric alone
 - the Measured controls (Boundary / Per / Scope / Statistic / Display / Unit) own no state: `measured-metric-config.ts` resolves each selection to the nearest registered metric key and writes it back to `selectedYAxisMetric`, so the power boundary rides on `i_metric` (there is deliberately no `i_pbasis`; see [PowerX Permanent View](./powerx-permanent-view.md))
 - the Measured Power Display value `timeline` (`y_measuredPowerTimeline`) swaps the chart body for `PowerTimeline`; its axis mode, line mode, window-only display, focused trace, all-in reference switch and concurrency filter are component state that `PowerTimeline` itself reads and writes through `useUrlState` as `i_ptaxis` / `i_ptlines` / `i_ptwindow` / `i_ptfocus` / `i_ptutility` / `i_ptconc`, so they never pass through the display domain
 - token-revenue price source (`i_revenue`): normalized uncached/cached/output pricing or the selected model's live OpenRouter catalog prices
 
@@ -1,3 +1,5 @@
+import { assertShareLinkParams } from '../support/share-link';
+
 // Deterministic intercepted rows exercise the validated/historical power UI:
 // one row has power_valid=1 and one has no validation verdict.
 
@@ -95,6 +97,15 @@ function visitCertifiedPowerChart(extraParams = '', benchmarks = powerBenchmarks
   cy.get('[data-testid="chart-figure"]').should('have.length.at.least', 1);
 }
 
+const MEASURED_POWER_OPTION = 'Measured Power';
+const MEASURED_POWER_TITLE = 'Measured Average Power per Chip';
+const TOKENS_PER_DOLLAR_TITLE = 'Total Tokens per $1 TCO';
+
+const assertYAxisMetric = (option: string, title: string) => {
+  cy.get('[data-testid="yaxis-metric-selector"]').should('contain.text', option);
+  cy.get('[data-testid="chart-figure"] h2').should('contain.text', title);
+};
+
 const visiblePowerPoints = () =>
   cy.get<SVGGElement>('.dot-group').filter((_, element) => element.style.opacity !== '0');
 
@@ -173,6 +184,39 @@ describe('Validated vs historical measured power', () => {
     cy.get('[data-testid="quick-filters-selected-count"]').should('contain.text', '1 selected');
   });
 
+  it('opens a shared i_power link without i_metric on the measured power scatter', () => {
+    visitCertifiedPowerChart('&i_power=certified');
+
+    assertYAxisMetric(MEASURED_POWER_OPTION, MEASURED_POWER_TITLE);
+    cy.get('.dot-group[data-hw-key^="mi300x"]').should('exist');
+    cy.get('.dot-group[data-hw-key^="b200"]').should('not.exist');
+    assertShareLinkParams({ i_power: 'certified', i_metric: null });
+  });
+
+  it('keeps an explicit tokens-per-dollar metric next to a power tier filter', () => {
+    visitCertifiedPowerChart('&i_metric=y_tokensPerDollarH&i_power=certified');
+
+    assertYAxisMetric(TOKENS_PER_DOLLAR_TITLE, TOKENS_PER_DOLLAR_TITLE);
+    assertShareLinkParams({ i_metric: 'y_tokensPerDollarH', i_power: 'certified' });
+  });
+
+  it('moves the default chart to measured power when Validated turns on, not when cleared', () => {
+    visitCertifiedPowerChart();
+    assertYAxisMetric(TOKENS_PER_DOLLAR_TITLE, TOKENS_PER_DOLLAR_TITLE);
+
+    cy.get('[data-testid="scatter-quick-filters"]').click();
+    cy.get('[data-testid="quick-filter-power-certified"]').click();
+    cy.get('[data-testid="quick-filters-selected-count"]').should('contain.text', '1 selected');
+    assertYAxisMetric(MEASURED_POWER_OPTION, MEASURED_POWER_TITLE);
+
+    cy.contains('button', 'Clear filters').click();
+    cy.get('[data-testid="quick-filters-selected-count"]').should('not.exist');
+    assertYAxisMetric(MEASURED_POWER_OPTION, MEASURED_POWER_TITLE);
+    cy.get('[data-testid="quick-filters-dialog"]').contains('button', 'Done').click();
+
+    assertShareLinkParams({ i_metric: 'y_measuredAvgPower', i_power: null });
+  });
+
   it('uses Optimal Only for measured markers while preserving the power boundary', () => {
     // Each chip has two boundary vertices and one lower-power interior point.
     const curveBenchmarks = powerBenchmarks.map((row) => ({
 
@@ -47,6 +47,7 @@ import type {
   TokenRevenuePriceSource,
 } from '@/components/inference/types';
 import { resolveMetricConfigKey } from '@/components/inference/metric-registry';
+import { getMeasuredMetricConfig } from '@/components/inference/measured-metric-config';
 import { Button } from '@/components/ui/button';
 import {
   Dialog,
@@ -74,7 +75,7 @@ import {
 } from '@/components/inference/perf-ruler-store';
 import { useParetoHighlightToggle } from './hooks/useParetoHighlightToggle';
 import { useOpenRouterPricing } from '@/hooks/api/use-openrouter-pricing';
-import { DEFAULT_Y_AXIS_METRIC } from '@/lib/url-state';
+import { defaultYAxisMetric } from '@/lib/url-state';
 import { computeToggle } from '@/hooks/useTogglableSet';
 import { buildAvailabilityHwKey } from '@/lib/chart-utils';
 import { getHardwareConfig, getModelSortIndex, isKnownGpu } from '@/lib/constants';
@@ -125,6 +126,9 @@ import {
   type SpecMode,
 } from './utils/quickFilters';
 
+const splitUrlList = (value: string | undefined): string[] =>
+  value ? value.split(',').filter(Boolean) : [];
+
 const InferenceDataContext = createContext<InferenceDataContextType | undefined>(undefined);
 const InferenceFiltersContext = createContext<InferenceFiltersContextType | undefined>(undefined);
 const InferenceDisplayContext = createContext<InferenceDisplayContextType | undefined>(undefined);
@@ -445,7 +449,11 @@ export function InferenceProvider({
     }
   }, [selectedGpuResolution]);
   const [selectedYAxisMetric, setSelectedYAxisMetric] = useState<string>(() =>
-    resolveMetricConfigKey(getUrlParam('i_metric'), initialYAxisMetric ?? DEFAULT_Y_AXIS_METRIC),
+    resolveMetricConfigKey(
+      getUrlParam('i_metric'),
+      initialYAxisMetric ??
+        defaultYAxisMetric(parsePowerTiers(splitUrlList(getUrlParam('i_power')))),
+    ),
   );
   const [tokenRevenuePriceSource, setTokenRevenuePriceSource] = useState<TokenRevenuePriceSource>(
     () => (getUrlParam('i_revenue') === 'openrouter' ? 'openrouter' : 'normalized'),
@@ -557,10 +565,8 @@ export function InferenceProvider({
   const [quickFilterPower, setQuickFilterPower] = useState<PowerTier[]>([]);
   const [quickFilterTopologies, setQuickFilterTopologies] = useState<string[]>([]);
   useEffect(() => {
-    const parse = (key: 'i_vendor' | 'i_fw' | 'i_disagg' | 'i_spec' | 'i_power' | 'i_topology') => {
-      const v = getUrlParam(key);
-      return v ? v.split(',').filter(Boolean) : [];
-    };
+    const parse = (key: 'i_vendor' | 'i_fw' | 'i_disagg' | 'i_spec' | 'i_power' | 'i_topology') =>
+      splitUrlList(getUrlParam(key));
     const vendors = parse('i_vendor');
     const frameworks = parse('i_fw');
     // Preserve old shared links: `agg` used to mean every non-disaggregated
@@ -990,6 +996,22 @@ export function InferenceProvider({
     },
     [setSelectedYAxisMetric, clearPresetOnChange],
   );
+  // A tier filter is only meaningful on a power chart. The move happens once, on
+  // entry, so a metric picked while the filter is on survives later tier edits.
+  // The URL restore keeps the raw setter so a link's explicit `i_metric` wins.
+  const setQuickFilterPowerAndMetric = useCallback(
+    (tiers: PowerTier[]) => {
+      setQuickFilterPower(tiers);
+      if (
+        quickFilterPower.length === 0 &&
+        tiers.length > 0 &&
+        !getMeasuredMetricConfig(selectedYAxisMetric)
+      ) {
+        setSelectedYAxisMetricAndClear(defaultYAxisMetric(tiers));
+      }
+    },
+    [quickFilterPower, selectedYAxisMetric, setSelectedYAxisMetricAndClear],
+  );
   const setSelectedGPUsAndClear = useCallback(
     (next: string[]) => {
       if (
@@ -1677,7 +1699,10 @@ export function InferenceProvider({
 
   useUrlStateSync(
     {
-      i_metric: selectedYAxisMetric,
+      i_metric:
+        selectedYAxisMetric === (initialYAxisMetric ?? defaultYAxisMetric(quickFilterPower))
+          ? ''
+          : selectedYAxisMetric,
       i_revenue: usesTokenSalePricing(selectedYAxisMetric) ? tokenRevenuePriceSource : 'normalized',
       i_mstat: fixedSequenceStatistic,
       i_gpus: selectedGPUs.join(','),
@@ -1716,6 +1741,7 @@ export function InferenceProvider({
     },
     [
       selectedYAxisMetric,
+      initialYAxisMetric,
       tokenRevenuePriceSource,
       selectedXAxisMetric,
       selectedE2eXAxisMetric,
@@ -2028,7 +2054,7 @@ export function InferenceProvider({
     setQuickFilterFrameworks,
     setQuickFilterDeployment,
     setQuickFilterSpec,
-    setQuickFilterPower,
+    setQuickFilterPower: setQuickFilterPowerAndMetric,
     setQuickFilterTopologies,
     setIsLegendExpanded,
     setHideNonOptimal,
 
@@ -67,7 +67,7 @@ import {
 import { XAxisModeSelector } from './XAxisModeSelector';
 import { showsTcoBasisSelector, Sequence, type Model } from '@/lib/data-mappings';
 import { useLocale } from '@/lib/use-locale';
-import { DEFAULT_Y_AXIS_METRIC } from '@/lib/url-state';
+import { defaultYAxisMetric } from '@/lib/url-state';
 
 const STRINGS = {
   en: {
@@ -207,8 +207,14 @@ export default function ChartControls({
   useEffect(() => setMounted(true), []);
 
   const { openDropdown, handleDropdownOpenChange } = useOpenDropdown<string>();
-  const { selectedModel, selectedSequence, selectedPrecisions, selectedGPUs, selectedDateRange } =
-    useInferenceFilters();
+  const {
+    selectedModel,
+    selectedSequence,
+    selectedPrecisions,
+    selectedGPUs,
+    selectedDateRange,
+    quickFilters,
+  } = useInferenceFilters();
   const {
     graphs,
     availableGPUs,
@@ -449,7 +455,7 @@ export default function ChartControls({
   };
 
   const secondaryCount =
-    (selectedYAxisMetric === DEFAULT_Y_AXIS_METRIC ? 0 : 1) +
+    (selectedYAxisMetric === defaultYAxisMetric(quickFilters.power) ? 0 : 1) +
     (showXAxisMode && selectedXAxisMode !== 'interactivity' ? 1 : 0) +
     (selectedXAxisMetric === undefined || selectedXAxisMetric === 'p90_ttft' ? 0 : 1) +
     (scaleType === 'auto' ? 0 : 1) +
 
@@ -46,15 +46,13 @@ describe('PARAM_DEFAULTS', () => {
     expect(PARAM_DEFAULTS.i_seq).toBe('');
   });
 
-  it('strips i_metric against the same default the dashboard opens on', async () => {
-    // A share link omits any value equal to PARAM_DEFAULTS. If this drifted
-    // from DEFAULT_Y_AXIS_METRIC, a link captured on the *other* metric would
-    // be written without `i_metric` and reopen on the dashboard default.
-    const { PARAM_DEFAULTS, DEFAULT_Y_AXIS_METRIC } = await import('@/lib/url-state');
+  it('opens on tokens per dollar, or on measured power under any power tier filter', async () => {
+    const { defaultYAxisMetric } = await import('@/lib/url-state');
     const { DEFAULT_METRIC_CONFIG_KEY } = await import('@/components/inference/metric-registry');
-    expect(PARAM_DEFAULTS.i_metric).toBe(DEFAULT_Y_AXIS_METRIC);
-    expect(DEFAULT_Y_AXIS_METRIC).toBe('y_tokensPerDollarH');
-    expect(DEFAULT_Y_AXIS_METRIC).toBe(DEFAULT_METRIC_CONFIG_KEY);
+    expect(defaultYAxisMetric([])).toBe('y_tokensPerDollarH');
+    expect(defaultYAxisMetric([])).toBe(DEFAULT_METRIC_CONFIG_KEY);
+    expect(defaultYAxisMetric(['certified'])).toBe('y_measuredAvgPower');
+    expect(defaultYAxisMetric(['legacy'])).toBe('y_measuredAvgPower');
   });
 
   it('has expected default for r_range', async () => {
@@ -375,6 +373,19 @@ describe('writeUrlParams + buildShareUrl', () => {
     expect(url).toContain('g_model=test-model');
   });
 
+  it('keeps an explicit tokens-per-dollar metric, whose default status depends on i_power', async () => {
+    setupWindow('', '/inference');
+    const { writeUrlParams, buildShareUrl, DEFAULT_Y_AXIS_METRIC } =
+      await import('@/lib/url-state');
+
+    writeUrlParams({ i_metric: DEFAULT_Y_AXIS_METRIC, i_power: 'certified' });
+    await vi.advanceTimersByTimeAsync(200);
+
+    const params = new URL(buildShareUrl()).searchParams;
+    expect(params.get('i_metric')).toBe('y_tokensPerDollarH');
+    expect(params.get('i_power')).toBe('certified');
+  });
+
   it('removes params that match their default value', async () => {
     setupWindow('', '/inference');
     const { writeUrlParams, buildShareUrl } = await import('@/lib/url-state');
 
@@ -13,6 +13,7 @@
  * Only non-default values are written to keep URLs short.
  */
 import { LIFECYCLE_DEFAULTS } from '@/components/calculator/lifecycle';
+import { MEASURED_METRIC_DEFAULTS } from '@/components/inference/measured-metric-config';
 import { dashboardRouteForPathname, getDashboardRoute } from '@/lib/dashboard-routes';
 import { routeModelForPathname } from '@/lib/model-routes';
 
@@ -135,19 +136,22 @@ export type UrlStateParams = Partial<Record<UrlStateKey, string>>;
 /** Default values for each parameter. Params matching their default are omitted from share URLs. */
 /**
  * Dashboard default y-axis: total tokens purchased per $1 of Hyperscaler
- * ownership TCO. It leads with infrastructure purchasing power, which depends
- * only on measured throughput and hardware cost, so the opening view does not
- * assume a token sale price. Token Revenue per GPU Hour remains one selector
- * click away for revenue-side questions.
- * `?i_metric=` still wins, so existing shared links are unaffected.
- *
- * Lives here rather than in `InferenceContext` because `PARAM_DEFAULTS` below
- * strips any value equal to the default from share links. If the two drifted,
- * a link captured on the *other* metric would be written without `i_metric`
- * and reopen on this one.
+ * ownership TCO. It depends only on measured throughput and hardware cost, so
+ * the opening view does not assume a token sale price.
  */
 export const DEFAULT_Y_AXIS_METRIC = 'y_tokensPerDollarH';
 
+/**
+ * The y-axis a link without `i_metric` opens on. A measured-power tier filter
+ * (`i_power`) is only meaningful on a power chart, so any tier opens the
+ * measured power scatter. `InferenceContext` resolves links and strips
+ * `i_metric` from share links with this one function; if those two defaults
+ * differed, a link would reopen on a metric other than the one it captured.
+ */
+export function defaultYAxisMetric(powerTiers: readonly string[]): string {
+  return powerTiers.length > 0 ? MEASURED_METRIC_DEFAULTS.power : DEFAULT_Y_AXIS_METRIC;
+}
+
 /** Shared defaults for the fleet lifecycle and calculator MW controls. */
 export const DEFAULT_FLEET_MW = '10';
 export const DEFAULT_LIFECYCLE_RAMP_MONTHS = String(LIFECYCLE_DEFAULTS.rampMonths);
@@ -169,7 +173,10 @@ export const PARAM_DEFAULTS: Record<UrlStateKey, string> = {
   // explicitly, so an explicit FP4 selection must survive (not be stripped as a
   // "default") or it would silently revert to the per-model auto default on reload.
   i_prec: '',
-  i_metric: DEFAULT_Y_AXIS_METRIC,
+  // No strippable default: the no-param metric depends on `i_power` and on a
+  // route's seeded metric, so `InferenceContext` writes '' when the metric
+  // equals that resolved default and the explicit key otherwise.
+  i_metric: '',
   i_revenue: 'normalized',
   i_pctl: 'p90',
   i_mstat: 'median',