Today, Elon Musk's xAI The startup announced an enormous $6 billion funding round.
Reports from April confirm that the Series B investment involves several well-known enterprise capital firms and investors, including Valor Equity Partners, Vy Capital, Andreessen Horowitz (A16z), Sequoia Capital, Fidelity Management & Research Company, Prince Alwaleed Bin Talal and Kingdom Holding, to call a number of.
Musk said the funding was raised at a pre-money value of $18 billion.
While it will not be unusual for young startups within the AI ​​space to receive large funding rounds and high valuations, this development is exclusive simply due to enormous amount of capital invested – it might be one among the biggest Series B investments ever made on this category.
xAI, which is aggressively moving forward with its Grok-1 model and chatbot, hopes to make use of the cash to speed up its work and higher compete with OpenAI, Anthropic, Google and others in the synthetic generalized intelligence (AGI) race.
“There might be more announcements in the approaching weeks,” Musk wrote on X.
Musk's aggressive push with xAI and Grok
Musk co-founded OpenAI in 2015 together with current CEO Sam Altman, President Greg Brockman and others, but spoke to the corporate about its priorities and direction in 2018left the corporate and eventually became one among its harshest critics – particularly due to its move away from open-source AI models and toward “closed” or proprietary models to which the corporate sells access.
Musk officially re-entered the generative AI space with the announcement of xAI in July last yr. Since then, the corporate has been working aggressively to develop advanced AI systems which might be “truthful, competent, and maximally useful to all of humanity.”
The company launched Grok-1 in November 2023, saying the AI ​​was modeled after The Hitchhiker's Guide to the Galaxy and will answer almost any query to assist humanity in its quest for understanding and knowledge – no matter background or political beliefs.
In benchmarks equivalent to GSM8K, HumanEval and MMLU divided According to xAI, Grok-1 outperformed Llama-2-70B and GPT-3.5, but lagged behind leading models equivalent to Anthropics' Claude3, OpenAI's GPT-4, Google's Gemini, and Meta's open-source Llama 3 model.
In the months that followed, the corporate expanded access to the model through the chatbot Grok – making it available to paying Premium and Premium+ subscribers on Musk’s separate social network X – and showed the way it could provide answers using the newest data from X posts while being witty, funny or sarcastic. Even
Musk repeatedly emphasized Grok's unique humorousness, despite many Identification of cases of hallucinations at the identical time.
In March of this yr, the corporate began to shift gears again, first making Grok-1 open source after which releasing its more powerful successors – Grok-1.5 and Grok-1.5V.
Grok-1.5 introduced improved reasoning and problem-solving capabilities, approaching the performance of well-known open and closed LLMs, including GPT-4 and Anthropics' Claude 3.
Grok-1.5V has since added multimodality by having the ability to understand things seen in documents, charts, tables, screenshots, and photos. Around the identical time, it was reported that the corporate was in talks with investors to boost $6 billion.
Now that the funding has been officially announced, xAI is anticipated to deploy it on graphics processing unit (GPU) clusters and hire staff to enhance Grok's capabilities.
This would make it a more direct competitor to OpenAI's ChatGPT, which recently received a significant upgrade with a brand new GPT-4o model that makes real-time judgements on text, speech, and pictures. Google has also raised the bar for Gemini's multimodal capabilities.
“xAI will proceed this steep progression in the approaching months and can shortly announce several exciting technology updates and products. The funds from the round might be used to bring xAI's first products to market, construct advanced infrastructure, and speed up research and development of future technologies,” the corporate wrote in a blog post.
An xAI supercomputer by 2025?
While it stays to be seen what updates are planned for Grok, work on the infrastructure will likely revolve around an enormous supercomputer. According to a different report from The informationMusk told investors that he desires to have this GPU cluster, or “Gigafactory of Compute,” ready by fall 2025.
The system will reportedly use Nvidia's H100 GPUs and might be no less than 4 times the scale of the biggest GPU clusters built to coach and run high-performance AI models, he said.
Apparently, OpenAI and Microsoft are also reportedly planning an enormous AI supercomputer. According to initial estimates, the project, called Stargate, will use hundreds of thousands of specialty chips to power OpenAI's models and will cost as much as $100 billion. Back in March, Musk sued OpenAI for abandoning its founding mission and turning right into a closed-source de facto subsidiary of Microsoft that develops AGI for profits quite than for the good thing about humanity.