NextFin News - Former Datadog engineers Sajid Mehmood and Conor Branagan have launched AI coding startup Niteshift and raised a $7 million seed round led by Greylock’s Jerry Chen, according to TechCrunch.
Niteshift is entering one of enterprise software’s most crowded markets as coding agents from OpenAI, Anthropic and others spread quickly. The company’s pitch centers on a risk many customers already know well: the model makers behind those tools could also become direct competitors.
Mehmood and Branagan, who helped Datadog grow from its early days into a multi-billion-dollar company, argue that customers will not want their most sensitive assets — the code that runs their products — tied too tightly to a single frontier AI vendor. In TechCrunch’s account, Mehmood said Datadog learned a similar lesson years ago from multicloud customers that did not want to run entirely on Amazon. He expects the same dynamic to emerge as Anthropic, OpenAI and others move into more vertical software categories.
Niteshift is also making a narrower claim than the headline might suggest. It is not trying to replace Claude Code or Codex, the two best-known coding agents. Instead, the company says it reduces dependence on them by providing the surrounding infrastructure needed to vet, maintain and orchestrate AI-generated code.
That puts Niteshift in a less visible part of the AI coding stack. The market has already split into two layers: the model layer, where the brand names and compute budgets sit, and the control layer, where security, workflow integration and governance often determine whether a tool is adopted inside a company. Niteshift wants to sell into that second layer, not compete head-on with the model providers.
Greylock’s Jerry Chen led the seed round, and the cap table includes Reid Hoffman, Datadog chief executive Olivier Pomel, Datadog co-founder Alexis Lê-Quôc, Braintrust founder Ankur Goyal and Reflection AI’s Misha Laskin. The roster gives the startup credibility. It does not make the business any easier. A $7 million seed round is modest by 2026 AI standards, especially in a category where startups often compete with one another and with the platform giants whose models they rely on.
The funding amount points to discipline, but it also shows how early the company still is. Niteshift has yet to prove that the lock-in problem is large enough, urgent enough and persistent enough to support a standalone vendor.
Datadog matters here for more than the founders’ resumes. The observability company built its reputation on selling infrastructure software that sat above, or across, multiple cloud providers instead of tying itself to one. That gives Mehmood and Branagan a credible story to take to enterprise buyers who worry that AI coding could become another layer of hidden dependence. But the comparison only goes so far. Multicloud was a procurement and architecture problem with clear budget owners, while AI coding is more experimental and more politically charged inside companies because it touches developer productivity, security reviews, software ownership and, increasingly, compliance.
Winning in that environment requires more than a warning about vendor lock-in. Niteshift will need to show that its tools save time without creating new risk.
The market is receptive, but not forgiving. OpenAI and Anthropic are pushing their own agents, while smaller companies are trying to win by being more specialized, more secure or more enterprise-friendly. Niteshift is trying to occupy a specific niche: not the model, not the editor, but the infrastructure that sits in between. That could appeal to buyers who like the productivity promise of AI coding but do not want to hand production code to a direct platform rival.
The same logic also sets limits on the business. If customers do not believe model vendors are likely to become hostile competitors in their own industries, the lock-in argument gets weaker. If OpenAI, Anthropic or another large provider improves governance features quickly enough, the need for a separate middle layer could shrink. And if companies remain in pilot mode rather than moving AI coding into production at scale, infrastructure spending may lag the headlines.
Investors are backing more than a coding startup. They are backing a view of how enterprise buyers may respond as AI agents become more powerful and more vertically integrated. For now, that proposition remains plausible but unproven, and Niteshift still has to show that buyers will pay for its layer instead of using the one offered by the AI giants it is trying to keep at arm’s length.
Explore more exclusive insights at nextfin.ai.
