The Design-Build Wall Comes Down
seansalmon 🤷♂️@seanaes
−5.3%
since Oct 14
Practiced to +5.3% over 14 days before going real.
Practice chapter: simulated results; no money was invested.
The idea
I believe AI is dissolving the wall between designing software and building it, letting one person do what used to take a designer and an engineer, and that shift plays out over a decade, not a quarter. This basket spreads across every layer of that change: design platforms, AI coding tools, the chips and clouds underneath, and the databases and monitoring every new app needs. It stays broad on purpose, anchored by large stable companies rather than single bets, with a small cash cushion. I'd rethink it if designers and engineers stop converging into the same workflow.
What it owns
- Every AI design and coding tool, regardless of which platform wins, ultimately runs on the same concentrated set of chip manufacturers' hardware. · NVIDIAEvery AI tool in the design and coding workflow runs on its compute, so growth in AI-assisted software creation translates directly to demand for its chips., Taiwan SemiconductorIt manufactures the advanced chips for every major AI compute provider, making it the underlying beneficiary regardless of which chip designer gains share., BroadcomIt builds custom AI chips for cloud providers seeking alternatives to single-source chip dependency, plus the networking infrastructure that connects data centers., Advanced Micro DevicesIt provides the credible alternative to single-source AI compute dependency, and growth in the chip market benefits it as buyers seek competitive options.23%
- The software where designers already spend their time is where AI-generated prototypes actually get built and shipped, making it the natural capture point for this shift. · FigmaIt is the workspace where designers already work and increasingly ship AI-generated prototypes directly, making it the natural center of the design-to-code shift., AdobeIt owns the creative toolchain designers use and is integrating AI throughout the suite, positioning it to retain users as they evolve from designers into builders., AutodeskIt sits at the point where architectural and manufacturing designs become real-world objects, positioning it to capture the same design-to-build shift in physical industries., Wix.comIt embodies the core shift most directly: users describe what they want and AI builds it, requiring no engineering knowledge, which expands the population of software creators.20%
- As designers begin shipping code directly, the platforms that help them write and deploy AI-assisted software expand their addressable market significantly. · MicrosoftIt owns GitHub and Copilot, the dominant AI coding layer, plus the Azure cloud where built software runs, making it the foundational anchor for AI-assisted code creation., AtlassianWhen designers ship working code, the resulting work still requires planning and tracking, and its products remain the standard platform for that collaboration., GitLabIt is the complete pipeline for building, testing, and shipping AI-written code, so more software creation means more pipelines to operate.17%
- A surge in software creation and deployment means more applications require hosting, and the largest cloud providers bill for all of it. · AlphabetIt captures the shift twice: Google Cloud hosts the software created through this process, while Gemini powers some of the AI models these tools rely on., AmazonAWS hosts the majority of new applications, so an explosion in AI-assisted software creation means more deployment workloads across its cloud infrastructure., OracleIt has become the primary cloud rental destination for AI model training and hosting, so growth in AI workloads drives demand for its compute infrastructure.17%
- Every new application still requires databases, observability, and deployment infrastructure, so an increase in software creation directly increases customer acquisition here. · CloudflareIt is the deployment platform where small, fast projects built by designers typically land, so an increase in rapid software creation flows directly to its user base., DatadogAll the new software created through this process, including AI-generated code, requires production monitoring and observability, which is its core business., MongoDBIt is the default database choice for new applications and for AI tools writing code, so an increase in software creation drives proportional growth in its adoption., SnowflakeIt stores and processes the data that new applications and AI systems run on, and growth in software creation means more data flowing through its platform.17%
- Cash6%