Meta Platforms is still a contender in the artificial intelligence model race. As the parent company of Facebook, it had not released a major new AI product for some time, which may have caused some observers to overlook its presence in the field. However, CEO Mark Zuckerberg announced a new large language model series called Muse, along with its first product, Spark, on the Threads platform this Wednesday.
Following the announcement, Meta's stock price rose by 6.5%. While the broader market also experienced significant gains that day, the new model contributed positively to the sentiment. This launch is welcome news. A report last month from The New York Times suggested that Meta had delayed its release plans because its new models underperformed compared to competitors. Given that Meta has spent billions of dollars to assemble a new engineering team led by former Scale AI CEO Alexander Wang, any substantial delay would have been a significant setback.
The key question is whether Spark can generate enough momentum in the AI market to alter Meta's trajectory. It is possible. Meta appears to be targeting the consumer market with this release—Zuckerberg stated that the new Meta AI performs "particularly well" in areas such as health, social content, shopping, and gaming. While many consumers may have already chosen a preferred chatbot, likely OpenAI's ChatGPT or Google's Gemini, a substantial number of users remain undecided. Meta's AI chatbot is already integrated into its family of social apps, reaching approximately 3.5 billion users. If the new AI model receives positive word-of-mouth, this massive user base could provide a significant advantage.
Of course, achieving that positive reception is a considerable challenge. It is also noteworthy to consider the potential impact on OpenAI if Meta succeeds. Currently, advertising might be the biggest hope for revenue growth for the creator of ChatGPT, especially since Anthropic seems to have secured a strong position in the enterprise market. The two companies most skilled at monetizing digital advertising globally are Meta and Google. Both have already been using AI technologies to boost their advertising businesses. If Meta makes a breakthrough in the chatbot arena, it would become much more difficult for OpenAI to build a substantial advertising business based on ChatGPT.
**Viewing Options as Debt**
As investor sentiment sours on enterprise software companies that heavily utilize stock-based compensation, the high cost of equity awards has become a more contentious topic this year. One perspective that deserves more attention is treating the potential cost of certain employee stock options and restricted stock units as a form of debt.
According to an article by my colleague Ken Brown about the HOLT valuation model, UBS's HOLT framework accounts for vested but unexercised equity awards in this manner. Consider the implications if accounting standard-setters were to adopt this method broadly; it would substantially increase the reported debt levels of many companies, with far-reaching consequences.
Take Netflix, a company known for its long-standing and extensive use of stock options to compensate employees. As of December 31st, Netflix had 127.7 million vested but unexercised stock options outstanding, with an average exercise price of $36.07. This price was far below its stock price, which was nearly $100 at the time.
Netflix's own calculations show that the total value of these options—the difference between the stock price and the exercise price—was $7.4 billion as of December 31st. With Netflix's conventional debt standing at approximately $14.5 billion at the same time, analysts treating the option value as additional debt would see the company's total debt burden increase significantly.
Comments