Generative AI Startups Secure Record Funding to Advance Open Source and Commercial Projects
Generative AI companies are attracting significant investment to support both commercial goals and open-source initiatives. Together, a startup focused on developing open-source generative AI and infrastructure for AI model development, has successfully completed a $102.5 million Series A funding round. This funding was led by Kleiner Perkins, with contributions from Nvidia and Emergence Capital, representing more than five times the amount raised in its previous round. The funds will be utilized to enhance Together’s cloud platform, allowing developers to create custom and open AI models, according to co-founder and CEO Vipul Ved Prakash.
In a blog post on Together’s website, Prakash stated, “Startups and enterprises are eager to develop a generative AI strategy that is not tied to a single vendor.” He emphasized the value of open-source AI, which serves as a robust foundation for applications, given the frequent release of increasingly powerful generative models. “We believe generative AI represents a new operating system for applications, profoundly impacting human society. The future AI landscape will encompass both proprietary and open models, making it essential to offer choices and options.”
Co-founded by Prakash, Ce Zhang, Chris Re, and Percy Liang in June 2022, Together aims to provide organizations with the tools to integrate AI seamlessly into their applications. Together's cloud platform is designed for running, training, and fine-tuning AI models, which the founders claim provides more scalable computing at competitive prices compared to leading providers like Google Cloud, AWS, and Azure.
As outlined in their blog, Together is focused on significantly reducing the costs associated with interactive inference workloads on large models. “We optimize our infrastructure with thousands of GPUs housed across multiple secure facilities, utilizing advanced software for virtualization and model optimization to lower operating expenses.”
Together’s current infrastructure spans data centers in the U.S. and EU, partnering with providers such as Crusoe Cloud and Vultr. This setup delivers about 20 exaflops of compute power, with clusters scaling from 16 GPUs up to 2,048 GPUs. Key customers include NexusFlow, Voyage AI, and Cartesia, many of whom also utilize Together’s APIs for model deployment.
Pika Labs, which recently secured $55 million, has leveraged Together’s GPU clusters to develop a text-to-video model. The company utilizes these clusters to train new versions of its model from scratch, generating millions of videos monthly for early access users. Prakash noted, “By creating customized infrastructure, we offer significantly improved economic efficiencies. The Together platform enables developers to swiftly integrate top open-source models or create their own through pre-training and fine-tuning. Our clients are drawn to our exceptional performance and reliability, with the added benefit that they maintain ownership of their AI investments and can deploy their models on any platform.”
In addition to its cloud services, Together offers a consultation service called Custom Models, which allows clients to integrate their data into the Together cloud and collaborate with the team to design, build, and test bespoke models.
Together is also dedicated to advancing open-source AI research. One of its initial projects, RedPajama, aimed to create a suite of open-source generative AI models, including "chat" models similar to OpenAI’s ChatGPT. The company has also released a fine-tuned version of Meta’s Llama 2 text-generating model, GPT-JT (a variant of EleutherAI’s open-source text-analysis model GPT-J-6B), and OpenChatKit, designed as a ChatGPT alternative.
The demand for generative AI continues to surge among companies and investors. According to IDC, generative AI investments are expected to increase from $16 billion this year to an astounding $143 billion by 2027. Generative AI startups captured approximately 40% of all venture capital funding—around $11.9 billion—in the first half of 2023.
However, not all generative AI initiatives are guaranteed success. Stability AI, once a favored investment among firms such as Lightspeed Venture Partners and Coatue Management, is reportedly considering a sale as profitability remains elusive and its financial outlook deteriorates.
Ranking
