Advanced Micro Devices Inc (NASDAQ: AMD) closed higher on Thursday after HSBC analysts led by Frank Lee raised their price objective on the semiconductor firm to $200.
Lee cited AMD’s new MI350 series GPUs for the bullish view, which he believes can rival Nvidia’s Blackwell B200 chips in performance as well as pricing.
While that may be true – the MI350 does indeed offer a pricing premium and compatibility with existing data centre infrastructure – the broader market dynamics actually tell a different story.
Nvidia currently controls more than 90% of the discrete GPU market, while AMD’s share has fell to 8.0% only. Plus, the former has its annual revenue pegged comfortably at $115 billion – while the latter is at $6.7 billion only.
A 17x gap – that’s not a competitive race, it’s a landslide. Even if AMD doubles its AI revenue, it remains a distant second only. That said, AMD stock has, nonetheless, soared nearly 85% over the past three months.
AMD lacks ecosystem despite strong performance
Investors should note that Nvidia’s dominance in the AI data centre segment is reinforced by its full-stack ecosystem, including CUDA, TensorRT, and NeMo, which are deeply entrenched in enterprise and sovereign cloud deployments.
While AMD is evidently improving its ROCm stack, it still suffers from fragmented adoption and limited developer tooling.
Its latest MI355 chips do boast 288GB of HBM3E memory and impressive bandwidth, but Nvidia B200 and GB200 still outperform in sparse compute, software optimization, and, most importantly, real-world inference throughput.
Moreover, Nvidia’s chips are backed by a mature developer ecosystem, with over 5 million CUDA developers and widespread adoption across hyperscalers like AWS, Azure, and Google Cloud as well.
In short, Advanced Micro Devices may offer better tokens-per-dollar, but that metric alone doesn’t win enterprise contracts. NVDA’s ability to deliver turnkey AI infrastructure – from silicon to orchestration – gives it unmatched pricing power and margin leverage.
Nvidia vs. AMD shares: the gap is structural, not just technical
HSBC’s optimism around AMD’s MI350 series is understandable – the chips are a genuine leap forward. But calling it a true rival to Nvidia in AI is a bit premature.
AMD may be narrowing the performance gap with its new chips, but the ecosystem moat, market share dominance, and developer lock-in still favour NVDA by a wide margin.
Until AMD can match Nvidia’s full-stack integration and software defensibility, it remains only a challenger, not a peer. Therefore, investors should be cautious about extrapolating short-term performance gains into long-term market leadership.
The AI chip war sure is heating up, but Nvidia still holds the high ground. Finally, other Wall Street firms do not particularly agree with HSBC’s optimism on AMD shares either.
While the consensus rating on AMD stock currently sits at “overweight”, the mean target of about $137 actually suggests potential “downside” of more than 5.0% from here.
The post AMD stock: HSBC says it’s catching up to Nvidia – but is it really? appeared first on Invezz