Computing-for-cash funding shows no sign of slowing as Amazon tries to leapfrog into the big league
Anthropic, which is developing a generative AI (GenAI) platform has attracted more hefty investment. SK Telecom stumped up $100 million in August to help fund the development of a large language model (LLM) to help telcos leverage GenAI.
Amazon plays catch-up
Now Amazon has dwarfed that sum, saying it will invest up to $4 billion to make Anthropic’s models more useful to AWS’ customers as well as acquire an unspecified minor shareholding. The initial investment that secures the stake will be $1.25 billion.
This deal with Anthropic is Amazon’s way to leapfrog into the AI big time and an admission that its own AI developments are lagging. For example, critics have lambasted Alexa’s lack of progress in understanding language over the last five years or so. This looked like woeful neglect, given the estimated installed active base of about 1 billion.
Amazon has had no takers for its chips to train AI. Now Anthropic is to use AWS’ machine learning accelerators, Trainium and Inferentia, to build, train and deploy new foundation models for AWS as well as technology for AWS machine-learning chips. This is a necessary step as Trainium and Inferentia cannot run on Cuda, NVIDIA’s accelerator platform. NVIDIA has about 85% market share in AI-enabling silicon.
Cloud matters
Under the deal, AWS will become Anthropic’s main cloud provider for mission-critical workloads. This is interesting, specific wording given the GenAI firm’s relationships with Google.
Anthropic received about $300 million funding from Google in February 2023, according to the Financial Times [subscription needed] and Google acquired a stake in the GenAI platform of about 10%. It was also announced that Anthropic had a “large cloud contract” with Google, although no details were made public.
Anthropic is registered as a public benefit corporation with special governance arrangements to protect its mission to “responsibly develop and maintain advanced AI for the benefit of humanity”. In other words, it is determined not to put all its eggs in one basket and is determined to resist undue influence by any multi-billion dollar cloudco. Indeed, this is what propelled it into existence.
Complex relationships and safety
Three years ago, Microsoft kicked off what the FT calls “cash-for-computing investment” in OpenAI three years ago, investing $1 billion in OpenAI, which was backed by Elon Musk and Sam Altman in 2015, and went on to develop ChatGPT.
ChatGPT reportedly clocked up 1 million users at the beginning of December 2022 after being available to the public for about a week. Microsoft chucked in an additional $10 billion in January this year.
Anthropic was formed in 2021 when a group of researchers led by Dario Amodei, head of safety, left OpenAI after a disagreement about the company’s strategy and raised $124 million in its first round of funding in May 2021.
FT says OpenAI has sought to keep its research into AI safety separate from its newer commercial operations by limiting Microsoft’s presence on its board.
Let’s hope we don’t look back on these well-intentioned goals with the same cynicism as we now view Google’s lauded motto of “Don’t be evil” around the turn of the century.
The question is not if the GenAI bubble will pop, only when, and those intentions could well be among the casualties.