{"id":22241,"date":"2023-06-13T20:38:44","date_gmt":"2023-06-14T00:38:44","guid":{"rendered":"https:\/\/ifintechworld.com\/news\/amd-a-pause-before-the-real-ai-ramp\/"},"modified":"2023-06-13T20:38:46","modified_gmt":"2023-06-14T00:38:46","slug":"amd-a-pause-before-the-real-ai-ramp","status":"publish","type":"post","link":"https:\/\/ifintechworld.com\/?p=22241","title":{"rendered":"AMD: A Pause Before The Real AI Ramp"},"content":{"rendered":"
\n

<\/picture>
<\/figcaption><\/figure>\n<\/p>\n

The excitement over AI chips had Advanced Micro Devices, Inc.<\/b> (NASDAQ:AMD<\/span>) trading back to $130 heading into the Data Center and AI event on June 13. The stock swooned following the event, with the chip company not announcing any surprise new chips or partnerships to accelerate the shift to AI. Regardless, my <\/span>investment thesis remains ultra Bullish on AMD stock, which is now trading $40 below all-time highs in late 2021 while <\/span>Nvidia Corporation<\/strong> (<\/span>NVDA) stock is already far above those levels.<\/span><\/p>\n

\n

\"Finviz <\/picture>
\n

Source: Finviz<\/span><\/p>\n<\/figcaption><\/figure>\n<\/p>\n

More Than Hype<\/h2>\n

AMD fell after the AI presentation due to the event not necessarily matching the hype leading up to the event. The chip company provided some general details regarding the new AI chips heading to the market, but the management team didn’t provide any boosts to the financials to attract more investors to push the stock higher.<\/span><\/p>\n

What matters is that AMD has the Data Center chips to play in the surging demand for AI. The company forecasts the Data Center total addressable market (“TAM”) hitting $150 billion by 2027, up from $30 billion now, for massive 50% annualized growth for the next 5 years.<\/p>\n

\n

\"TAM<\/span>
\n

Source: AMD Data Center & AI presentation <\/span><\/p>\n<\/figcaption><\/figure>\n<\/p>\n

AMD had previously presented a $300 billion total TAM opportunity over the next 5 years with only $125 billion assigned to Data Center. The one big revenue boost at the AI event was the suggestion this TAM jumps to $150 billion now in 2027.<\/p>\n

What concerns investors the most is this analysis of the swing in CPU demand with the implementation of AI chips. The product requires a huge shift to spending on GPUs (highlighted by Nvidia numbers) while not necessarily using additional CPUs in the process.<\/p>\n

MI300 Monster<\/h3>\n

As SemiAnalysis<\/em> highlights, AMD is the only other company with a track record of delivering a chip for high performance computing. Intel<\/strong> Corporation<\/strong> (INTC) is nowhere to be found in the discussion with failed AI hardware acquisitions and no GPU coming to market to compete with the Nvidia GPUs. The news leaves the MI300 from AMD as the only competitor to Nvidia in a GPU chip where the company is estimated at charging 5x the production cost.<\/p>\n

At the AI event, CEO Lisa Su talked prominently about the new MI300 chips. The MI300A (CPU+GPU) is sampling to customers now with the MI300X (GPU only) sampling in Q3. Both chips will see production ramps in Q4 with the following configurations:<\/p>\n