Unlock Editor's Digest free of charge
Roula Khalaf, editor of the FT, picks her favorite stories on this weekly newsletter.
Microsoft boss Satya Nadella expressed a standard view within the technology industry on the time he said It was recently noted that enormous language models, the engines behind the generative AI boom, have gotten “more like a commodity.”
With a handful of leading modelers vying for honors with each recent iteration of their AI, it's becoming increasingly difficult to tell apart OpenAI's latest GPT from Anthropic's Claude or Google's Gemini.
It is all of the more remarkable that Nadella's Microsoft has just backed OpenAI's latest round of financing and its value has risen to $150 billion. Will this moment be considered the height of generative AI mania?
Valuing a fast-growing technology company in a brand new market is notoriously difficult. But the extent to which generative AI has modified the technology landscape and the speed at which OpenAI is emerging has left investors in search of benchmarks and historical comparisons.
First, consider what built it. Launched almost two years ago, ChatGPT became a successful consumer brand almost overnight and now boasts 250 million weekly users. The $20 monthly subscription fee, paid by a small minority, has increased the corporate's annual revenue to $3.6 billion.
OpenAI is also on its strategy to becoming a broader technology platform. Many other corporations have integrated AI into their very own services and products. The tools the corporate is developing to make its technology more useful within the business world have given it a rare opportunity within the enterprise market.
It's tempting to attract parallels with previous hot startups like Google. When the search company's market value first reached $150 billion in 2006, with lower than half the market, it was not the clear winner within the search business that it later became. This 12 months's $10 billion in revenue was roughly according to the $11 billion OpenAI forecast for next 12 months.
But that is where the comparisons fail and the size of the challenge ahead for OpenAI becomes clearer. Google spent money back in 2006. OpenAI, and not using a viable business model, is on the right track to burn greater than $5 billion in money this 12 months and has little prospect of stemming the inflow within the near term.
Along with the soaring costs of coaching ever-larger models, the numerous computing power required to answer user prompts will proceed to weigh heavily on margins as the corporate grows. It also doesn't appear to have the option to make use of pricing as a weapon. Although prices have been rapidly reduced to enable more efficient query response, the fee of querying other LLMs available through major cloud services has largely fallen in parallel.
That points to OpenAI's biggest challenge: the shortage of deep moats around its business and the extraordinary competition it faces.
On the patron side, Meta said last week that 500 million people now watch Meta.AI no less than once a month, an indication of the large, competitive markets available to OpenAI's big tech rivals. Google and Meta even have ready-made promoting corporations which have proven to be the very best strategy to monetize large digital audiences.
ChatGPT can claim a preferred position on the iPhone due to a cope with Apple. But Apple only makes the chatbot available through its Siri assistant, and even then only to handle questions beyond the present capabilities of its own AI models – hardly a recipe for long-term success as OpenAI seeks to capitalize on its early customer benefits to consolidate.
Competition can be increasing sharply on the company side. Close ally Microsoft is diversifying its initial reliance on OpenAI, while the capabilities of open source AI models have rapidly evolved, making them viable alternatives. Metas Llama has not yet change into the “Linux of AI,” as Mark Zuckerberg suggested last week, however the danger of commercialization that Nadella warned about is great.
At this point, it’s price remembering that generative AI continues to be in its infancy and that the big resources being poured into the technology still hold big surprises and will introduce significant unexpected disruptions.
OpenAI's latest models hint on the potential. Its voice-controlled GPT-4o is credited with breaking recent ground in naturalistic voice interaction and potentially opening recent consumer markets for AI. And it claims its GPT-o1 is the primary model able to breaking down a fancy problem and finding the trail to an answer. This could point to a future where AI models themselves take over more of the work in a business application, taking value away from traditional software because it becomes more central to working life.
It is not possible to say how far such capabilities will advance and whether OpenAI can maintain a meaningful lead in modeling. But with essentially the most powerful tech corporations closing fast, investors backing the $150 billion group will need strong courage.
