Trajectory just closed a $45 million Series B round at a $300 million valuation, barely a year after launching its platform to make AI model fine-tuning accessible to non-engineers. The funding, led by Sequoia Capital, signals growing investor appetite for tools that bridge the gap between powerful AI models and the domain experts who understand how to apply them—but lack the technical chops to train them.
Trajectory is betting that the next wave of AI adoption won’t come from data scientists—it’ll come from doctors, lawyers, and financial analysts who know their domains inside out but can’t write a line of Python. The one-year-old startup just landed $45 million in Series B funding led by Sequoia Capital, valuing the company at $300 million and positioning it at the center of a brewing battle over who gets to control AI model customization.
The timing is notable. While OpenAI scales back its fundraising ambitions and the broader AI investment landscape cools, Sequoia is doubling down on infrastructure plays that make existing models more useful. Trajectory’s pitch is straightforward: give subject matter experts a no-code interface to fine-tune models like GPT-4 or Claude on their proprietary data, without involving engineering teams.
The platform targets a real friction point in enterprise AI deployment. Companies sitting on mountains of domain-specific data—medical records, legal precedents, financial transactions—want models that understand their particular vocabulary and workflows. But traditional fine-tuning requires machine learning expertise most organizations don’t have in-house. Trajectory’s interface lets users upload datasets, define training objectives, and deploy custom models through a visual workflow builder.
Sequoia’s bet mirrors its recent pattern of backing AI tooling rather than foundation models themselves. The firm led Blacksmith’s $66M round for AI-powered code testing last month and poured $60 million into Mecka’s robot training data platform in October. The strategy acknowledges that while OpenAI, Google, and Meta will likely dominate base model development, there’s massive value in the picks-and-shovels layer that makes those models enterprise-ready.
Trajectory’s $300 million valuation—achieved in just 12 months—stands in sharp contrast to the broader funding environment. The startup launched into a market where rising bond yields are hammering AI infrastructure investments and even well-funded companies like world model startups are going dark despite massive war chests. But Trajectory’s focus on immediate enterprise ROI rather than speculative research appears to be resonating with investors looking for safer AI bets.
The competitive landscape is getting crowded fast. OpenAI offers its own fine-tuning APIs, while platforms like Hugging Face and Anthropic’s Claude give developers programmatic access to model customization. But Trajectory is wagering that the bigger market isn’t developers—it’s the millions of knowledge workers who understand their problems better than any engineer could, if only they had the tools to translate that expertise into AI behavior.
The Series B will fund expansion into regulated industries where data sensitivity makes generic models inadequate. Healthcare providers can’t send patient data to OpenAI’s APIs for fine-tuning without running into HIPAA issues. Law firms need models trained on their specific case history and legal strategy. Financial institutions want fraud detection tuned to their particular customer patterns. Trajectory’s on-premise deployment option and audit trails address these compliance requirements while maintaining the no-code promise.
Sequoia’s involvement also suggests confidence in Trajectory’s ability to scale beyond early adopters. The firm typically looks for companies that can dominate a category, and the fine-tuning market is projected to grow as enterprises move from experimenting with off-the-shelf AI to deploying customized solutions. If every company needs its own specialized models—and most lack the technical talent to build them—Trajectory’s positioning as the “WordPress for AI training” could prove prescient.
The funding arrives as the AI stack continues to stratify into distinct layers. Foundation model providers like OpenAI, Google, and Meta compete on raw capabilities and parameter counts. Cloud providers battle over inference infrastructure. And now a new category is emerging: companies that make those powerful-but-generic models actually useful for specific business problems, without requiring a PhD in machine learning.
Trajectory’s rapid rise from launch to $300 million valuation in 12 months reflects investor belief that this middleware layer—the translation zone between foundation models and end users—represents the next major value capture opportunity in AI. Whether that belief holds depends on how quickly enterprises adopt custom models and whether Trajectory can defend against both upmarket competition from OpenAI and downmarket pressure from open-source alternatives.
Trajectory’s $45 million raise marks a clear bet that AI’s next chapter belongs to domain experts, not data scientists. As foundation models commoditize and enterprises demand customization without complexity, the startup’s no-code fine-tuning platform addresses a genuine market gap. But with Sequoia’s backing comes pressure to prove that democratizing model training isn’t just a nice idea—it’s a defensible business in an increasingly crowded AI tooling landscape. The $300 million valuation gives Trajectory runway to find out whether every company really does need its own custom models, or whether off-the-shelf AI proves good enough for most use cases.
