Enterprise AI Infrastructure Matures with Real-Time Capabilities
PLUS: Former Amazon-acquired Founder Building Something New
This edition is brought to you by VARA Network
Built on the Gear Protocol - Vara Network is a revolutionary Web3 application platform, removing many of the barriers associated with building, deploying, and operating dApps. Vara’s computer science paradigms like persistent memory and actor model, along with parallel data processing, allows for deep scalability, new design patterns and features. Vara’s unique architecture makes it completely frictionless for users. Learn more at Vara.Network
Venture Radar
Enterprise AI Infrastructure Matures with Real-Time Capabilities
Databricks' acquisition of Tecton—announced this week—marks a meaningful evolution in enterprise AI infrastructure. Tecton, founded by former Uber engineers and backed by Sequoia and Kleiner Perkins, offers a feature store that enables real-time data processing with low latency. In a fast-maturing AI ecosystem, the game isn’t just building better models—it’s delivering them seamlessly into production workflows, and real-world speed matters.
This deal reflects a broader trend: enterprises increasingly demand real-time AI that powers responsive applications like voice assistants and dynamic decision systems. Databricks, already valued north of $100 billion and fresh off a major funding term sheet, is doubling down on this vision by integrating Tecton’s capabilities into its Agent Bricks platform. That kind of integration strengthens client relationships (Coinbase being a shared client) and deepens the functionality of Databricks' stack.
For investors and operators, the takeaway is clear: AI success now hinges on infrastructure—not just algorithm performance, but speed, deployment stability, and operational readiness. Next-gen AI platforms must offer developer-friendly pipelines, data fidelity, and responsiveness at scale. The Tecton integration signals where enterprise AI is going—and who’s best positioned to lead.
Geeks of the Week
Startup Name: KlavisAI
Geography: US
One-liner: Easiest way to connect to high-quality, secure MCP Servers at scale.
Founder(s) Background: Senior Software Engineer on Google Gemini, Tech Lead at Lyft.
Thoughts:
Simplifying MCP Integration for AI Agents
Klavis offers hosted, enterprise-grade Model Context Protocol (MCP) servers with built-in OAuth and multi-tenancy authentication—eliminating the heavy development burden of client-side setup and security. This allows developers and AI systems to connect to tools like Gmail, Slack, Notion, and more with a simple API call, accelerating deployment and reducing risk.Technically Strong Founders with Deep AI Credibility
The startup was founded by Xiangkai Zeng, a former Google DeepMind engineer who co-authored the Gemini architecture and led function-calling infrastructure, alongside Zihao Lin from Lyft and Nordstrom with systems-scale expertise. Their combined background in AI tooling and infrastructure positions Klavis to execute confidently in this niche.
Startup Name: PgDog
Geography: US
One-liner: PgDog is the missing routing layer for PostgreSQL. It moves data between nodes and directs query traffic, transparently to the application.
Founder(s) Background: Former Staff Software Engineer at Lattice, Senior Software Engineer at Instacart.
Thoughts:
Effortless Multi-Tenant Architecture Without Code Refactoring
PgDog enables multi-DB, multi-tenant PostgreSQL deployments—directing queries based on tenant ID—without requiring application changes. Its routing happens transparently via a proxy configuration, making it ideal for SaaS platforms serving multiple clients from the same codebase.Resilience and Safe Production Testing Through Traffic Replication
PgDog supports mirroring live query traffic to secondary replicas—letting teams safely "warm up" new instances or test the impact of configuration changes in real time. This supports risk-free experimentation without touching production systems.
Startup Name: Paloma
Geography: US
One-liner: AI-native CRM for billing, renewals, and expansions.
Founder(s) Background: Former Head of Deel’s EOR Business Unit, First product hire at Deel.
Thoughts:
Revenue on Autopilot with Intelligent CRM
Paloma transforms post-sale operations by turning contracts, communication logs, and service usage into a unified, AI-powered system of record. It's designed to work seamlessly with your existing tools — contracts, billing systems, even unstructured inputs like PDFs or emails — allowing teams across sales, success, and finance to operate from the same real-time context. From churn detection to handling renewals and upsells, Paloma automates the heavy lifting and lets revenue run smoothly behind the scenes.
Founder(s) building in stealth
Deals of the Week
TinyFish – AI (Web Agents)
Raised $47 million Series A (Aug 20). Led by ICONIQ Capital, with participation from USVP, MongoDB Ventures, and Sandberg Bernthal Venture Partners.August – LegalTech (AI for Mid‑Size Law Firms)
Secured $7 million Seed funding (Aug 5). Led by NEA and Pear VC, with strategic input from law schools, OpenAI, Ramp, and Bain Capital Ventures affiliates.Abridge – Healthcare AI
Raised a total of $700 million over 18 months, across multiple rounds. Now valued at around $5.3 billion, Abridge is investing 80% into AI development and reserving 20% for strategic acquisitions.
Notable Mergers and Acquisitions
Databricks → Tecton
Acquisition of the ML feature-store startup for an undisclosed amount via Databricks’ private shares. This move adds real-time data processing and deployment capabilities to Databricks’ Agent Bricks platform—accelerating enterprise adoption of AI agents.Thoma Bravo → Dayforce
A $12.3 billion take-private deal for the HR software provider, representing a substantial ~32% premium. A clear strategic consolidation in the HR tech space.
This edition is brought to you in partnership with Stella Capital.