Nancy Xing

Compute Labs

AI Infrastructure · Fintech · Crypto

Role
Co-founder, design lead
Timeline
2023–Present
Tools
Figma
Scope
Brand · Product · Design systems · Web
Compute Labs marketing site, Scale Your GPU Operations

GPU compute is the scarcest resource in AI, but investing in it used to require owning a data center. Compute Labs turns enterprise GPUs into yield-bearing digital assets, so you can invest in compute without running the hardware. Incubated by NVIDIA Inception VC Alliance and Hub71 — the UAE sovereign-fund accelerator with a 1.1% acceptance rate. As co-founder and design lead, I own the whole design surface: brand, product, the business-facing story, the design system, and even how we show up in the room at offline events.

A brand for a brand-new category

Compute Labs sits where AI infrastructure, finance, and crypto meet — three worlds with very different visual languages. I built an identity that reads as credible to institutional investors and native to the on-chain world at once: warm neutrals, a confident green accent, technical diagrams, and restrained typography that lets the numbers lead.

Compute Labs Visual Identity Guidelines: logo, product lockups, isometric illustration, typography, photography, color and data viz, interface, voice, and iconography
The Compute Labs visual identity guidelines — crest, wordmark, type, color, and tagline.

One platform, several surfaces

The product spans the full lifecycle of a compute investment — from underwriting a deal to claiming the yield it generates. I design each surface to make complex capital and infrastructure legible: show people the real numbers, clearly, and let the data earn the trust.

Compute Labs GPU Financing: live deal actuals and fin ops dashboard with monthly revenue, APY, DSCR, and GPU network health
GPU Financing — live deal actuals and fin ops: revenue vs. projection, APY, DSCR, and GPU network health.

Designing the investor story

Beyond the product UI, I shape how Compute Labs makes its case to the market — mainly on the investor side. That means the pitch narrative, data visualizations, one-pagers, and decks that translate a novel asset class into something an institution can underwrite with confidence.

Examples of Compute Labs investor-facing material: Capital Markets for Compute deck with deal sourcing, model summary, underwriting criteria, and risk management slides
Investor-facing material — the Capital Markets for Compute deck, from deal sourcing to underwriting criteria.

A system the whole team builds on

I built and documented the design system, then packaged it into tooling so the team ships on-brand by default — shared tokens, components, and patterns that keep every new surface coherent as the company grows.

Compute Labs design system board: color tokens, Geist/Inter/JetBrains Mono typography, 4px spacing grid, radius, tech shadows, buttons, risk badges, data table, and dark mode
The token system at a glance — color, type, spacing, depth, components, and dark mode.

Bringing the brand into the room

The brand doesn't stop at the screen. I design how Compute Labs shows up in person — booths, signage, printed collateral, and event moments that carry the same clarity and confidence as the product.

Compute Labs offline events: the Token2049 Singapore climbing-wall booth, the 'Future of AI is Collaborative' installation, and the AI After Hours cocktail signage
In the room — the Token2049 climbing-wall booth, event installations, and AI After Hours.