Updated December 25, 12:21 p.m. Pacific: Updated information on xAI’s valuation and Kingdom Holdings’ investment.
xAI, the artificial intelligence firm founded by Elon Musk, has secured $6 billion through a Series C funding round.
The company disclosed this week that it received backing from notable investors including Andreessen Horowitz, BlackRock, Fidelity, Lightspeed, MGX, Morgan Stanley, OIA, QIA, Sequoia Capital, Valor Equity Partners, Vy Capital, Nvidia, and AMD.
According to a public filing, Saudi Arabia’s Kingdom Holdings invested approximately $400 million in this funding round. The filing also indicated that xAI’s valuation has surged to $45 billion, nearly double its previous worth.
With this new capital, xAI has raised a total of $12 billion, building on the $6 billion raised earlier in May.
As reported by the Financial Times, only existing investors from xAI’s previous funding round were permitted to take part in the latest round. Allegedly, investors involved in financing Musk’s acquisition of Twitter had access to purchase up to 25% of xAI’s shares.
“xAI’s most powerful model to date is currently in training, as we concentrate on launching cutting-edge consumer and enterprise products,” xAI stated. “The proceeds from this financing will be directed towards bolstering our sophisticated infrastructure, launching groundbreaking products, and expediting research and development.”
Advancing AI Technology
Musk established xAI last year. Shortly after, the company unveiled Grok, its flagship generative AI model that now facilitates various features on X, the rebranded platform formerly known as Twitter. This includes a chatbot available to both X Premium subscribers and limited markets for free users.
Grok has been described by Musk as having “a rebellious streak,” demonstrating a willingness to engage with “spicy questions” that many AI systems typically reject. For instance, when prompted to be vulgar, Grok complies without restraint, using language that would be inappropriate for models like ChatGPT.
Musk has criticised ChatGPT and similar AI systems for being overly “woke” and “politically correct,” even though Grok has shown hesitancy to cross certain thresholds and avoidance of specific political themes. He has also characterised Grok as “maximally truth-seeking” and less biased than other models, though there is evidence suggesting that Grok may exhibit a left-leaning bias.
In the past year, Grok has embedded itself deeply within X, the social network previously known as Twitter. At its launch, Grok was exclusively accessible to X users, including developers adept at implementing the “open-source” edition.
Due to an integration with xAI’s proprietary image generation model, Aurora, Grok can create images on X (controversially without restrictions). The model also possesses the capability to analyse images and summarise news and trending topics—though not always accurately.
Reports suggest that Grok may soon manage additional functions within X, enhancing its search abilities and account bios, along with assistance in post analytics and reply configurations. Recently, X introduced a “Grok button” to facilitate users in uncovering “relevant context” and engaging more deeply with trending topics and real-time discussions.
xAI is racing to compete against formidable rivals such as OpenAI and Anthropic in the generative AI sector. The company launched an API in October, enabling customers to integrate Grok into third-party applications, platforms, and services. It has also recently released a standalone Grok iOS app to a select group of testers.
Musk claims that the competition has not been equitable.
In a lawsuit against OpenAI and Microsoft, with OpenAI being its close associate, Musk’s legal team alleges that OpenAI is “actively attempting to eliminate competitors” like xAI by securing agreements from investors not to support them. Musk’s lawyers further contend that OpenAI gains undue advantages from Microsoft’s infrastructure and expertise, describing it as a “de facto merger.”
Despite these challenges, Musk asserts that the data from X provides xAI with a competitive edge over its adversaries. Last month, X amended its privacy policy, allowing third parties, including xAI, to train AI models using posts made on the X platform.
It is noteworthy that Musk was among the original co-founders of OpenAI but departed the organisation in 2018 due to fundamental disagreements regarding its direction. He has previously claimed that OpenAI has benefited from his early contributions while failing to uphold its non-profit commitment to make the results of its AI research accessible to everyone.
OpenAI has expressed strong disagreement with Musk’s viewpoint. In a press release issued in mid-December, OpenAI labelled Musk’s lawsuit as misleading, baseless, and a case of resentment.
Understanding the xAI Ecosystem
xAI has presented an ambitious vision for its models, which will be trained using data from Musk’s various enterprises, including Tesla and SpaceX. The intention is for these models to enhance technology across these businesses. According to reports from The Wall Street Journal, xAI is already supporting customer service for SpaceX’s Starlink internet service and is claimed to be in discussions with Tesla to provide research and development in return for a share of the automotive company’s revenue.
However, many Tesla shareholders have voiced their concerns regarding these intentions. Several have taken legal action against Musk, arguing that he has redirected both talent and resources away from Tesla to pursue what they perceive as a competing enterprise.
Despite this, xAI’s revenue has surged to approximately $100 million annually, driven by its agreements and its range of developer and consumer products. In contrast, Anthropic is expected to achieve $1 billion in revenue this year, while OpenAI aims for $4 billion by the end of 2024.
Musk announced this summer that xAI is developing the next generation of Grok models at its Memphis data centre, which reportedly was constructed in a mere 122 days and is currently partially reliant on portable diesel generators. The company intends to upgrade the data centre, which houses 100,000 Nvidia GPUs, next year, with plans to double that capacity as stated in a press release. (GPUs are preferred for model training and operations due to their capability to perform numerous calculations simultaneously.)
In November, xAI received approval from the Memphis regional power authority for an additional 150MW of power supply — sufficient to cater to approximately 100,000 homes. To gain the agency’s approval, xAI promised to enhance the quality of local drinking water and provide discounted Tesla-made batteries to the Memphis electric grid. Nonetheless, some locals have expressed their concerns, claiming it might burden the power grid and could degrade air quality in the area.
Additionally, Tesla is anticipated to utilise the upgraded data centre to enhance its autonomous driving technologies.
Since its inception, xAI has seen rapid expansion, increasing its workforce from around a dozen employees in March 2023 to over 100 today. In October, the startup relocated to the former headquarters of OpenAI in San Francisco’s Mission district.
xAI has indicated to investors its intentions to secure additional funding in the upcoming year.
It is not alone in its pursuit of substantial financial backing. Anthropic recently obtained $4 billion from Amazon, raising its total funding to $13.7 billion, while OpenAI garnered $6.6 billion in October, increasing its total to $17.9 billion.
Significant investments like those from OpenAI and Anthropic propelled AI venture capital to reach $31.1 billion across more than 2,000 transactions in Q3 2024, as reported by PitchBook.
StartupSuperb offers an AI-focused newsletter! Sign up here to receive it in your inbox every Wednesday.