First Profit in the Industry
For years, labs building large language models have been viewed as businesses doomed to permanent losses because the costs of computing power and training new models grow faster than revenue. The positive adjusted operating result for the second quarter challenges this view and suggests that previous losses stemmed more from investments in training future models than from unprofitable products.
The difference compared with the competition lies in the type of figure involved. The $11.5 billion represents actual invoiced revenue from April through June, which, together with the first quarter, amounts to roughly $16 billion for the first half of the year. While the run rate is an estimate, this is money that has already come in.
Battle with OpenAI
Path to the Stock Market
Sources: cnbc.com, finance.yahoo.com and theinformation.com



