news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Baidu holds Create AI Developer Conference in Shenzhen

According to Phoenix Technology, on August 23, Beijing time, Baidu released its second-quarter financial report on Thursday, with a slight decline in revenue. Bloomberg pointed out that this financial report shows that Baidu has difficulty in monetizing artificial intelligence (AI), and the transition of revenue-generating business from search advertising to AI has encountered difficulties.

Baidu's second-quarter revenue was 33.9 billion yuan, down 0.4% year-on-year, and below the average analyst estimate of 34.1 billion yuan. As of the close of U.S. stocks on Thursday, Baidu's stock price fell 4.40%.

AI is difficult to monetize

Bloomberg said that the poor performance highlights the challenges Baidu faces in converting its leadership in the field of generative AI into substantial revenue. With the help of advertising and cloud services, Baidu's "Wenxin" large language model has gradually begun to bring in additional sales, but it has also been involved in an AI price war with companies such as Alibaba Group and Tencent Holdings. It may take Baidu several years to comfortably reduce its reliance on advertising.

"Baidu's business appears to be at a crossroads,"“Its AI initiatives have not yet produced the expected results, and AI has not yet become a growth driver for Baidu,” Hou Xiaotian, an analyst at Tianhao Capital, wrote in a report before Baidu’s earnings report.

Robin Li

Baidu founder Robin Li has high hopes for a Chinese version of ChatGPT, but he faces an uphill battle competing with other big tech companies and emerging startups. IDC estimates that Baidu had about one-fifth of China's $250 million generative AI market last year.

But that leadership is rapidly being eroded. Bytedance’s chatbot Doubao surpassed Baidu’s Wenxin Yiyan in downloads last year and has more monthly active users on iOS in China, according to data from mobile app analytics firm Sensor Tower.

Doubao's iOS monthly active users surpass Wenxinyiyan

Bloomberg Industry Research analysts Robert Lea and Jasmine Lyu said in a research report: Baidu's prospects still face huge challenges. As Tencent and Alibaba continue to narrow the gap, Baidu's AI business is expected to continue to lose money in the next three years. We expect Baidu's main source of cash, the search engine business, to continue to be under pressure due to the increasingly fierce competition in the short video field. "

"The worsening AI price war is likely to cause Baidu to lose more market share this year, hindering its ability to leverage its technological expertise to generate revenue and reverse its loss-making AI business. We expect Baidu's adjusted net profit to fall 5%-10% this year," they said in the report.

AI drives down software profit margins

In addition to Baidu, US technology giants are also investing heavily in generative AI. Data from research firm Synergy Research Group shows that Amazon, Microsoft and Google spent more than $48 billion in the second quarter, mainly on data centers.

Big tech companies are expected to spend a staggering $1 trillion on generative AI as they increasingly see it as the next big tech wave after cloud computing.

But how much return can such a huge investment in AI bring? This week, analysts at RBC Capital Markets gave a preliminary answer: Not that high.

Software gross profit margins decline in the AI ​​era

"The result of generative AI is that the long-term gross margin of software will be structurally reduced," they said in a research report. According to data from Royal Bank of Canada Capital Markets, when software moves from "on-premises deployment" (enterprises run software on their own computers) to "cloud" (running remotely on rented servers), gross margins drop from 90% to 75%.

The analysts estimate that the switch from cloud computing to generative AI will further reduce software gross margins, to about 60%.

In the software industry, gross margins are typically around 90%. That sounds like a lot, which is why the industry is so attractive to investors and why valuations of software companies are so high. The upfront costs of developing new software are high. But once it's developed, the costs of making new versions and distributing them to customers are next to nothing. So, the more software a business sells, the greater the profits.

High cost

So why might software businesses see declining profits in the coming AI era?

Generative AI is expensive to develop and expensive to run. AI models need to be trained, which involves buying extremely expensive GPUs from Nvidia and then putting those AI chips into servers. Those servers need special cooling and networking in large data centers. These facilities consume a lot of electricity, are also expensive, and require expensive upgrades.

That doesn’t include the cost of the data used to train AI models. Big tech companies and startups are trying to avoid paying much of that, but collecting and cleaning that data is still expensive.

Once AI models are trained, they need to be run. This is the inference step, where the model processes new data or requests to infer useful information from them. This step also requires expensive chips and is an ongoing expense.

This is different from the previous on-premises software business. Before, every new sale was almost 100% profit. However, every time an AI customer uses a generative AI service, the provider incurs a lot of costs. For example, industry analyst Dylan Patel estimated last year that ChatGPT costs $700,000 a day to operate.

Revenues make up for margin decline

However, analysts at RBC Capital Markets are not all pessimistic.

They expect generative AI to become so revolutionary that customers will spend more on new AI software, which they estimate will double or even triple future software revenues from current levels.

The analyst also explained that as the software market size increases, more "profit dollars" may be created even with lower margins.

When profit margins decline, “profit dollars” is a metric that management and analysts use to measure a company’s absolute profits. For example, if a company has $100 million in revenue and a 10% profit margin, its absolute profits are $10 million. If the company’s revenue increases to $300 million but its profit margin drops to 8%, its profits are still $24 million, which is more than before.”

“While we expect generative AI to put pressure on margins, we believe long-term gross margins will be higher in the post-generative AI era,” analysts at RBC Capital Markets concluded.

However, there is an important assumption here, that is, generative AI can bring huge revenue growth. Otherwise, these huge AI investments may produce "pretty bad" economic benefits.(Author/Xiao Yu)

For more first-hand news, please download the Phoenix News client and subscribe to Phoenix Technology. If you want to read in-depth reports, please search "Phoenix Technology" on WeChat.