Niteshift Raises $7M to Fight AI Model Lock-In
Key insights
- Niteshift raised $7 million led by Greylock's Jerry Chen, co-founded by former early Datadog engineers Sajid Mehmood and Conor Branagan.
- The per-minute pricing model positions Niteshift as cloud infrastructure rather than a model vendor competing on per-token costs.
- Direct competitors include Cognition at a $26 billion valuation, OpenRouter at $1.3 billion, Cursor, and Amazon Bedrock.
Why this matters
Vendor lock-in with major AI model providers is the explicit problem Niteshift is built to address, and its launch signals that well-connected founders believe enterprises will pay for a neutral routing layer rather than accepting whatever a single frontier lab offers. Per-minute pricing is a structural departure from per-token models, positioning Niteshift as infrastructure rather than a reseller, which changes how enterprise AI coding costs get projected and allocated. The competitive field already includes Cognition at a $26 billion valuation and OpenRouter at $1.3 billion, meaning Niteshift's $7 million seed is entering a market where incumbents are heavily capitalized and the window for differentiation is narrowing.
Summary
Niteshift launched with a $7 million seed led by Greylock's Jerry Chen, founded by two former early Datadog engineers betting enterprises will pay to avoid vendor lock-in on AI coding models.
The company routes requests across GPT, Claude, and open-source models at per-minute rates rather than per-token. CEO Sajid Mehmood is direct: "Everybody's worried about getting stepped on by these giants."
Essentially: (Niteshift, Greylock) are building neutral routing infrastructure as a hedge against OpenAI and Anthropic expansion.
- Angels: Reid Hoffman, Datadog's Olivier Pomel and Alexis Lê-Quôc, Ankur Goyal of Braintrust, and Misha Laskin of Reflection AI.
- Per-minute pricing frames this as cloud infrastructure rather than a model reseller competing on margin.
- Competing against Cursor, Cognition ($26 billion valuation), Amazon Bedrock, and OpenRouter ($1.3 billion valuation).
Whether enterprise teams actually need routing-layer neutrality before one dominant tool consolidates the market is the core bet Niteshift is making.
Potential risks and opportunities
Risks
- OpenAI and Anthropic could release native multi-model routing features within 12 months, eliminating the neutrality gap Niteshift is building into before it signs enterprise contracts.
- Cognition ($26 billion valuation) and Amazon Bedrock have distribution reach to bundle routing into existing enterprise agreements, marginalizing a $7 million-funded newcomer on procurement timelines.
- OpenRouter, already valued at $1.3 billion, is a direct competitor with a significant head start that Niteshift's seed round may be insufficient to match on engineering and go-to-market resources.
Opportunities
- Datadog co-founders Olivier Pomel and Alexis Lê-Quôc as angels could open enterprise CTO networks at existing Datadog customers, shortcutting early sales cycles for Niteshift.
- Enterprise teams seeking predictable AI coding costs have a structural reason to adopt per-minute pricing as an alternative to variable per-token billing from OpenAI and Anthropic.
- Open-source model providers gain enterprise distribution if Niteshift's routing layer defaults to including them, bypassing the direct enterprise sales motion those providers currently lack.
What we don't know yet
- No enterprise pilot customers or launch partners are named in the article, leaving Niteshift's early commercial traction unverified.
- Whether Niteshift's per-minute pricing is cheaper or more expensive than per-token alternatives for typical enterprise coding workloads is not addressed.
- How Niteshift handles model version updates or quality drift when routing across multiple providers simultaneously is not explained.
Originally reported by techcrunch.com
Read the original article →Original headline: Niteshift Raises $7M Seed to Route AI Coding Requests Across GPT, Claude, and Open Source — Betting Against Big-AI Lock-In