Datadog Vets' Niteshift: AI Coding Startup Takes On Big AI Lock-in

Niteshift, an AI coding agent startup, has successfully secured $7 million in a seed funding round, spearheaded by Greylock's Jerry Chen. While this figure might appear modest within the current AI investment landscape, the company, established by two former early engineers from Datadog, has garnered significant backing from notable angel investors including Reid Hoffman, Datadog's Olivier Pomel and Alexis Lê-Quôc, Ankur Goyal of Braintrust, and Misha Laskin from Reflection AI.

Founded by Sajid Mehmood and Conor Branagan, both instrumental in scaling Datadog from its nascent stages to a multi-billion dollar valuation, Niteshift enters the competitive AI coding sector with a provocative premise: questioning the rationale behind companies entrusting their most critical assets—the proprietary code powering their products—directly to foundational model providers such as OpenAI and Anthropic, particularly when these providers frequently launch competing applications, effectively disrupting existing startups and businesses.

CEO Mehmood draws a parallel to Datadog's formative years, recalling how the monitoring firm attracted e-commerce clients who were hesitant to develop their infrastructure on Amazon Web Services. This apprehension was well-founded, as Amazon was simultaneously driving many of those same retail businesses into obsolescence during the era termed the "retail apocalypse."

According to Mehmood, an analogous scenario is already unfolding within the AI domain. Companies like Anthropic, OpenAI, and others are rapidly expanding into specialized vertical software markets, a trend some observers have dubbed the "SaaSpocalypse."

"At Datadog, we witnessed this phenomenon firsthand," Mehmood stated. "A significant portion of our multicloud revenue originated from e-commerce enterprises that preferred not to operate on Amazon's infrastructure. We are unequivocally poised to observe an identical dynamic as Anthropic extends its competitive reach into sectors such as legal, healthcare, finance, and various others."

Niteshift's core premise is that businesses will progressively gravitate towards infrastructure solutions that effectively decouple the core coding model from the extensive orchestration required to rigorously vet and sustain AI-generated code. Crucially, they anticipate a preference for vendors who do not harbor conflicting business objectives.

It is important to clarify that Niteshift does not aim to supersede popular coding agents like Claude Code or Codex. Instead, its proposition is to diminish organizational reliance on these specific tools.

Niteshift’s AI coding cloud is designed to intelligently route tasks among these various models—including open-source alternatives and other proprietary options—tailoring the selection to the specific requirements of each individual project.

"The capability to seamlessly transition between GPT and other cloud models is paramount," Mehmood emphasized, adding, "Everyone is apprehensive about being overshadowed by these dominant industry players."

This strategic vision was precisely what captivated Greylock's Jerry Chen, leading to his investment.

"As the leading AI research labs ascend the technology stack, a significant opportunity emerges to provide customers with an alternative approach: disaggregating their agents from the underlying infrastructure," Chen explained to TechCrunch. "Niteshift is developing the platform that facilitates this for coding agents, empowering clients to make substantial investments in their developer tooling without being confined to a singular model or agent provider."

Furthermore, Niteshift's business model deviates from selling tokens; instead, it offers infrastructure services, adopting a cloud provider's pricing structure with per-minute usage rates.

"While competitors are largely focused on selling intelligence designed to replace human labor," Mehmood clarified, "we are providing software to agents rather than directly to humans—fundamentally, we remain a software vendor."

Despite its unique approach, Niteshift is entering an already saturated market for AI coding tools. The concept of model independence, while compelling, is not entirely new, and numerous competitors possess a substantial early advantage. This competitive landscape includes companies such as Cursor, potentially subject to acquisition by SpaceX; Cognition, which recently secured $1 billion in funding at a $26 billion valuation; Amazon Bedrock; and the AI gateway platform OpenRouter, which just raised $113 million, valuing it at $1.3 billion. The list of established players continues.

Mehmood's counter to these market challenges rests firmly on the profound experience of the founding team. Both Mehmood and Branagan did not merely theorize about these issues; they intimately experienced and overcame them, successfully scaling Datadog through the precise types of growing pains that large engineering organizations are now encountering with AI-generated code. He asserts that modern teams require the capability to autonomously run, test, and verify software within their live production environments, demanding robust infrastructure crafted by individuals with proven expertise in operating at scale.