The Hallway Track
15 tracked signals on agent-infrastructure.
LangChain argues AI models are commoditizing, making domain-specific harnesses the new competitive differentiator
“the models themselves are becoming somewhat commoditized, and it's all of this stuff around the model, in particular the data...This is becoming the differentiator”
AWS launches new AgentCore runtime with elastic memory reclamation and consistent cold starts for production agents.
“Agents are no longer experiments.”
AI agents transacted $24M via USDC in 30 days using the X402 autonomous payment protocol
“we believe this year, 2026, is when agents actually start paying for services that they want”
AWS launches Agent Registry and open ARD specification for cross-environment AI agent discovery
“if every registry describes resources in the same format and exposes discovery through a common protocol, publishers describe once and consumers discover everywhere”
LangChain launched SmithDB and LangSmith Engine to power a full Agent Development Lifecycle platform.
“Today, we launched SmithDB as a foundational database to power all of this, and we built LangSmith Engine as an agent that sits on top and helps you spin this flywheel faster and faster.”
LangSmith Sandboxes gives AI agents isolated, scalable compute environments spinning up in under a second
“The question isn't whether agents need computers. It's how you give them one, safely.”
Browser agents fail due to poor infrastructure around models, not model intelligence; better page representation makes cheap models fast and reliable.
“models are pretty smart, but it's the infra around them that sucks”
Ampersend built a pay-per-intelligence routing layer on Amazon Bedrock AgentCore Payments letting AI agents pay autonomously per request.
“agents should pay for intelligence the same way they already call APIs, programmatically, instantly, and without human intervention.”
Amazon Bedrock AgentCore harness is now generally available, letting developers run production-grade agents via two API calls.
“Getting it into production is where the work explodes”
Daytona enables AI agents with cloud computers accessible via API, signaling end of localhost development
“The end of localhost”
Microsoft Foundry model router achieves 37.7% cost savings by routing prompts to appropriately-sized models
TypeSafe's Jev is a new 'System 1' model class making fast type-safe decisions rather than generating text.
“Jev does not generate text. It is not a generative AI in that sense... Jev is about making fast decisions from specifically well-scoped questions with type safe schemas for each of them.”
LangChain launches Managed Deep Agents to handle production agent infrastructure at scale.
“this will make it far more easy than before to build powerful agents and actually run them at scale”
Amazon Bedrock AgentCore lets coding agents run in isolated cloud microVMs instead of on developers' laptops.
“The laptop won the job by being the nearest machine, not the right one.”
Box CTO warns agent infrastructure has a short half-life requiring continuous adaptation
“we are now in the ballpark of trillion tokens probably will be 10 trillion tokens sometime soon”