Those in the IT world will have noticed the name Discoveryhatmakertechcrunch, which has become synonymous with a brand of AI system that enables users to build applications faster and easier than ever before. The company has a team of Stanford professors behind it, and SoftBank Vision Fund 2 has invested in the venture. Here’s a look at what they’re doing.
SoftBank Vision Fund 2
Earlier this week, SoftBank Vision Fund (SVF) 2 announced it has invested in SambaNova Systems Inc., a startup developing hardware and software for artificial intelligence applications. This is the first time SoftBank has invested in the company, and it is expected to value SambaNova Systems at $5 billion.
SambaNova’s flagship product is Dataflow-as-a-Service, which is a software and hardware platform that enables organizations to rapidly deploy AI solutions to support their business objectives. It’s subscription-based, and can be used to accelerate existing data centers or augmented by AI-powered hardware.
Other investments include Predictim, which is a company that uses AI to develop mathematical algorithms that analyze billions of data points. Photomath, meanwhile, uses artificial intelligence to solve mathematical equations. Predictim’s algorithms can analyze data within minutes, compared to hours or days for traditional data mining.
Other investments include Meesho, a company that offers logistics and payment processing services, and Trax, a company that provides a granular view of changing stock prices. Trax’s offerings include an image recognition and a platform that can detect anomalies and provide real-time insights.
Stanford professors behind company
Whether it is through a new breed of silicon or by building a new class of software, SambaNova is on the ball when it comes to AI. Their new DataScale platform will deliver high performance computing and efficiency across applications. The company has raised a cool $676 million in a series D round led by SoftBank Vision Fund 2*. The company also landed a number of new investors including Temasek and GIC.
The company has also got its eye on the AI community with its latest hire, Rodrigo Liang, former SVP of development at Oracle. Liang brings over 20 years of experience to the table, and his team has a knack for building products that help companies run smoothly. SambaNova will also be aggressively tackling legacy competition in the AI space. This is the company’s second round of funding, and the company has raised a total of more than $1 billion in funding.
AI system
Founded by Stanford professors Kunle Olukotun and Chris Re, SambaNova is a startup that aims to build systems for running data-intensive AI applications from the data center. It has developed a machine learning platform and chips for AI workloads. Its flagship product is Dataflow-as-a-Service, a subscription-based platform that gives enterprises access to SambaNova’s AI system. The system includes pre-trained deep learning models, and users can recompile their workloads with SambaNova’s compiler in just an hour.
SambaNova has developed a system that has set world records for AI performance. The company claims its AI processor can improve performance by up to five times over GPUs running the same models. The company is also working on cognitive simulation. SambaNova’s models are built from parts learned from historical training data. They include natural language models, recommender models, and computer vision models. SambaNova’s AI system is aimed at accelerating the shift to an AI-enabled world.
SambaNova Systems is a Palo Alto, California-based startup that aims to create hardware and software platforms to run AI applications. It is backed by Silicon Valley’s most prominent investors, including Intel Capital and BlackRock.
Dataflow-as-a-Service subscription model
Using Dataflow-as-a-Service, enterprises can access the power of SambaNova’s AI system, without the need for training and maintaining the system themselves. Rather than train custom deep learning models, customers can benefit from pre-trained models with industry data, which can accelerate AI capabilities by weeks or even months.
In order to train a deep learning model, enterprises often have to invest in expensive personnel and advanced technology. Moreover, building complex deep learning models often takes months or years, and they can be out of date the moment they are released. Dataflow-as-a-Service provides customers with state-of-the-art deep learning models, and eliminates the need for training and maintaining the system. This approach can help enterprises accelerate their AI capabilities and jump start their enterprise-level AI initiatives.
Conclusion
SambaNova’s Dataflow-as-a-Service combines an advanced software and hardware systems platform with a full-stack solution, making it easy for customers to access the power of SambaNova’s machine learning model. This AI system can be customized to fit any industry, and it provides users with the ability to access pre-trained models and industry data through easy-to-use APIs.
