A pair of Amazon alum raised $12.5 million for a new startup that helps enterprises build and deploy secure artificial intelligence applications, preventing company data from slipping through the cracks.
Founded earlier this year, Griptape is developing an open-source Python framework and cloud platform. Its tools give companies a way to safely build large language models, including conversational, copilot, and autonomous agents. The idea is to provide enterprises with security controls, allowing them to use AI models such as OpenAI’s GPT or Anthropic’s Claude 2 without compromising data security. It’s currently in private preview and plans for a commercial launch in October.
A new SEC filing revealed the fresh cash. Kyle Roche, the startup’s co-founder and CEO, confirmed the funding to GeekWire.
Griptape is backed by Seattle venture firms Fuse and Acequia Capital; Crosslink Capital; Range Ventures; and Peterson Ventures. It has 12 employees, according to Roche.
Roche spent more than eight years at Amazon Web Services (AWS) in various roles. In his position as head of immersive technology, Roche led teams that built and ran AWS services including Amazon Location Service, Amazon Nimble Studio, Thinkbox, and others. He previously founded 2lemetry, an IoT startup that Amazon acquired back in 2015.
Vasily Vasinov, the company’s co-founder and CTO, most recently was principal product manager at AWS. He met Roche as a senior software engineer at 2lemetry. The co-founders are joined by chief product officer Derek Pai, who served as principal product manager at AWS, and chief creative officer Jason Schleifer, a former principal creative director at AWS.
Engineers building large language model apps can also use LangChain. However, there are concerns about its readiness for enterprise use. Griptape pitches itself as an “enterprise grade alternative to LangChain.” Customers include Torc, ThingLogix, and Zennify.
Seattle startup Fixie is also building tools to help companies fuse large language models into their software stack. It raised a $17 million seed round in March.
Earlier this summer, Griptape was among 21 startups selected for the inaugural cohort of the AWS generative AI accelerator.
Griptape looks to address a pain point in enterprise AI app development: data vulnerability. Ilana Stern, a general partner at Peterson Ventures, wrote in a blog post that using large language models for enterprise apps is akin to leaving your front door open during a home remodel, giving every subcontractor unrestricted access to rummage through your fridge and closet.
“Which is exactly why an enterprise wants to limit data access and control privacy when building applications using generative AI models,” she wrote.
Tech giants Microsoft, Amazon and Google are in a race to integrate large language models into their core business strategies. The companies are rolling out tools in online search, software development, and office software.
An IBM survey found that more than 60% of enterprise CEOs report facing “significant pressure from investors, creditors, and lenders to accelerate adoption of generative AI.”
