Preview image. Unlock full-res
Light split pricing section pairing an oversized condensed headline with a tabbed compute card listing instance rows, green checkmark features, and a green-strip logo bar.
Summary
A light split pricing section that places an oversized condensed headline on the left and a tabbed "serverless compute" pricing card on the right, listing GPU instance rows with per-hour and per-second rates.
Visual description
On a warm off-white background, the left column carries a three-line oversized condensed all-caps headline, "Pricing for intensive infrastructure", with a two-line muted sub-headline about pay-as-you-go usage. The right column holds a white rounded card topped by a two-segment toggle: a green "Compute" tab (active) and a muted "Postgres" tab. Inside, a bold "Serverless Compute" heading sits beside a short explainer paragraph, then a small "/Popular Instances" label with a dark GPU/Standard/Eco segmented control. Three feature pills with green checkboxes read Autoscaling, Scale to Zero, and Per Second Billing. A table lists instance rows (RTX-A6000, L40S, A100) with vRAM specs and right-aligned per-hour prices plus a smaller per-second figure below each. A full-width dark "Get started" button closes the card. A separate rounded bar at the bottom reads "Trusted by the most ambitious teams" beside faded Bosch, eToro, Intel, Meilisearch, and MongoDB logos.
Key takeaway
Tabbing the pricing card by product type (Compute vs Postgres) keeps multiple priced services in one compact card. Showing both per-hour and per-second rates per instance row communicates granular usage-based billing transparently.
Reuse notes
Best for usage-based or per-instance infrastructure pricing where rows of hardware tiers replace fixed plan columns. The condensed headline and monospace figures keep the engineering tone. Needs real instance names and rates to be useful. Works on light; pairs naturally with the same brand's hero.
