AI coding agent startup Niteshift has raised $7 million in a seed round led by Greylock’s Jerry Chen. That’s a modest amount by AI standards, but the startup was founded by two former early Datadog engineers and has attracted high-profile angels like Reid Hoffman, Datadog’s Olivier Pomer and Alexis Lecoq, Braintrust’s Ankur Goyal, and Reflection AI’s Misha Raskin.
Founded by Sajid Mehmood and Conor Branagan, who helped grow Datadog from its early days to a multi-billion valuation, the company entered the crowded AI coding space with a compelling idea. Why would model makers like OpenAI and Anthropic directly trust their most sensitive assets: the code that runs their products? Why, considering these companies are constantly “crushing” startups and businesses by launching competing apps?
CEO Mehmood likens this to Datadog’s early growth. This is when the monitoring company acquired an e-commerce customer who refused to build on Amazon Web Services. It became known as the “retail apocalypse,” and it was a legitimate concern given that Amazon was simultaneously putting many similar retailers out of business.
As Mehmood sees it, the equivalent of AI is already underway. Anthropic, OpenAI, and others are rapidly entering the vertical software market called SaaSpocalypse.
“At Datadog, we clearly recognized this,” says Mehmood. “A big part of our multi-cloud business comes from e-commerce businesses that didn’t want to run on Amazon, right?…We’re definitely going to see the same kind of moves that Anthropic is trying to make to compete in legal, medical, financial, and all sorts of other areas.”
Enterprises will increasingly seek infrastructure that separates their coding models from all the other orchestration needed to ensure that AI-generated code is properly inspected and maintained (and will seek vendors without competing agendas).
To be clear, Niteshift does not replace Claude Code or Codex, two of the most popular coding agencies. They claim that this will reduce dependence.
Niteshift’s AI coding cloud routes between these models, along with open source options and more, based on each project’s needs.
“Being able to switch between GPT and cloud models is important. Everyone is worried about being trampled by these giants,” Mehmood said.
That idea intrigued Greylock’s Chen.
“As Frontier Labs moves higher up the stack, there is an opportunity to offer customers an alternative to decoupling the agent from the infrastructure it runs on,” Chen told TechCrunch. “Niteshift is building a platform that enables this for coding agents, allowing customers to invest deeply in development tools without being locked into a single model or agent vendor.”
Beyond that, Niteshift does not sell tokens. They sell their infrastructure and charge per minute usage rates like cloud providers.
“Everyone else is selling labor alternatives,” Mehmood said. “We’re selling software to agents, not humans, but we’re still selling software.”
Still, Niteshift is entering a crowded market of AI coding tools. Model independence is not a new idea, and Niteshift’s competitors are well ahead. This includes Cursor, which may soon be swallowed up by SpaceX. Cognition just raised $1 billion at a valuation of $26 billion. Amazon bedrock. AI gateway platform OpenRouter just raised $113 million at a valuation of $1.3 billion. The list goes on.
Mehmood’s answer to all of this is the depth of the founding team. Mehmood and Branagan not only studied these problems, they lived them and scaled Datadog through the very growing pains that large engineering organizations are currently facing with AI-generated code. He said teams need to autonomously run, test and validate software in real production environments, and they need an infrastructure built by the people who have done it at scale.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
