Conversations with engineers and insights from the TypeScript Congress 2023 highlight its growing prominence. JetBrains’ Developer Survey shows TypeScript’s user share has tripled from 12% in 2017 to 34% in 2022 and 2023.

TypeScript is no longer just a frontend language. It is becoming the orchestration layer for AI native applications.

Java dominated the infrastructure era of the internet. It became a standard that triggered a massive retooling across industries, leading to the rise of Databricks, Confluent, and Atlassian. Java’s influence was particularly profound in data infrastructure, underpinning projects like Spark, Kafka, Hadoop, and Cassandra.

The same pattern may now be emerging around TypeScript.

Python still dominates AI model training through frameworks like PyTorch and TensorFlow. But AI applications increasingly look different from AI models. They are realtime, streaming, asynchronous systems sitting between users, APIs, agents, and inference layers.

That environment maps unusually well to TypeScript.

The shift matters because the economic centre of AI may move away from training and toward orchestration: agent workflows, collaborative interfaces, edge-native compute, and AI application infrastructure.

This is already visible in the ecosystem. Vercel launched its AI SDK around TypeScript. Cloudflare and edge runtimes increasingly optimise for JavaScript execution. New frameworks are emerging TypeScript-first rather than treating JavaScript as a secondary wrapper around Python infrastructure.

Both languages share impressive enterprise adoption rates. The key question is: who will benefit most from this TypeScript retooling? While Vercel has emerged as a significant player, opportunities for TypeScript extend beyond front-end development. Who becomes the Databricks or Confluent of the AI application era?

You track it the same way you would have tracked Java in the early 2000s: not by asking “is the language popular?” but by asking:

“What operational complexity emerges once this becomes the default developer environment?”

Java itself was not the big winner economically. The winners were the companies solving the second-order problems Java created:

  • distributed compute → Hadoop/Spark/Databricks

  • messaging/event streams → Kafka/Confluent

  • collaboration/dev workflows → Atlassian

  • deployment/orchestration/tooling → entire middleware layer

Where TypeScript becomes the “default assumption”

Not “can it do AI?”

But: are startups defaulting to TS first? Are SDKs launching TS-native? Are infra providers optimising for TS runtimes? Are AI application engineers mostly writing TS?

Once enough developers coordinate around one runtime/environment, massive infrastructure opportunities emerge around the friction that coordination creates.

That is the real lesson from Java.