
A day after AMD smashed both top and bottom-line earnings projections and provided a positive outlook for its artificial intelligence chip business through 2024, the company’s stock shot up more than 8% on Wednesday.
The chipmaker slightly outperformed the LSEG, previously Refinitiv, projection of 68 cents per share with adjusted earnings per share of 70 cents. Moreover, revenue surpassed expectations, coming in at $5.8 billion as opposed to the projected $5.7 billion.
Wall Street seemed to have been enthused by one comment: According to AMD CEO Lisa Su’s remarks on Tuesday night, the company anticipates GPU sales to reach $2 billion by 2024 and to be roughly $400 million in the fourth quarter.
The stock recovered and rose on Wednesday as a result of that prediction.
Wall Street analysts are optimistic about AMD’s chances in the Nvidia-dominated AI space. Nevertheless, AMD is among the few companies that can produce powerful GPUs that drive artificial intelligence models.
“Anticipating MI300 accelerator is anticipated to generate $400 million in revenue in Q4E and $2 billion or more in CY24E, with robust adoption among hyperscalers, enterprises, OEMs, and AI startups,” analysts from Bank of America stated in a letter to investors. According to AMD, volume production of its new MI300A and MI300X GPUs is scheduled for this year’s fourth quarter.
Raymond James analysts maintained AMD as a “strong buy” despite cutting their price objective from $145 to $125, mostly because of its AI division.
“We are cutting our projections, but the tale appeals to us because of the long-term potential for AI and ML. The Raymond James analysts stated, “AMD is off to a solid start, and we see no reason why the company can’t capture 10–20% share of the $100B+ AI accelerator market longer term.”
Jeffries analysts expressed a similar opinion regarding AMD’s GPU division.
Perhaps the most exciting development during the earnings call was AMD’s revised forecast for its data center GPU family (the MI300), which will start at $400 million in both the fourth quarter of 2023 (for HPC) and the first quarter of 2024 (for inferencing and training).” This information was shared by Jeffries analysts in a note to investors.