Tech

Niteshift secures $7m seed to build AI coding infrastructure independent of model vendors

Startup aims to provide routing infrastructure for coding tasks across OpenAI and Anthropic models, charging for usage rather than selling tokens.

Author
Owen Mercer
Markets and Finance Editor
Published
Draft
Source: TechCrunch · original
Datadog veterans launch AI coding startup Niteshift on a bet against Big AI lock-in
Former Datadog engineers Sajid Mehmood and Conor Branagan back bet against Big AI lock-in with Greylock-led round

Niteshift, an artificial intelligence coding agent startup founded by former Datadog engineers Sajid Mehmood and Conor Branagan, has closed a $7 million seed funding round. The investment was led by Jerry Chen of Greylock Partners, with participation from a notable group of angel investors including LinkedIn co-founder Reid Hoffman, Datadog co-founders Olivier Pomel and Alexis Lê-Quôc, Ankur Goyal of Braintrust, and Misha Laskin of Reflection AI.

The company has entered the competitive AI coding market with a strategy focused on providing infrastructure that allows businesses to route coding tasks between various large language model providers, including those from OpenAI and Anthropic, without becoming locked into a single vendor. Mehmood, who serves as chief executive, argues that as major AI labs move into vertical software markets, companies will increasingly seek independent infrastructure to protect sensitive code assets.

Mehmood draws parallels to Datadog’s early growth, where the monitoring company won e-commerce customers who refused to build on Amazon Web Services due to concerns about competing with their own retail clients. He describes a similar dynamic emerging in the AI sector, which he refers to as the "SaaSocalypse," where firms like Anthropic and OpenAI are expanding into legal, healthcare, and finance software. This shift creates a conflict of interest for businesses relying on these labs for foundational models.

Niteshift does not aim to replace popular coding agents such as Claude Code or Codex, but rather positions itself as a layer that reduces dependence on them. The platform charges for infrastructure usage on a per-minute basis, similar to a cloud provider, rather than selling tokens. Chen noted that this approach offers an alternate path by unbundling agents from the underlying infrastructure, allowing customers to invest in developer tooling without locking themselves into a single model or agent vendor.

The startup faces significant competition from established players with substantial valuations and funding. Competitors include Cursor, which may be acquired by SpaceX; Cognition, valued at $26 billion following a $1 billion raise; Amazon Bedrock; and OpenRouter, which recently raised $113 million at a $1.3 billion valuation. Mehmood contends that the founding team’s experience scaling Datadog provides the necessary depth to address the growing pains large engineering organisations face with AI-generated code.

Continue reading

More from Tech

Read next: Satirical ‘Data-Free’ Filesystem Project πFS Released on GitHub
Read next: CISA Mandates Three-Day Patching for US Federal Agencies Amid AI Threats
Read next: Logitech enters portable peripheral market with $80 Mobi Fold mouse