Qualcomm FQ2 26 – The Long Game

Finally, the market seems to get it.

• Qualcomm gave weak guidance for the coming quarter but said that it could see the bottom, and by confirming its entrance into the data centre, finally gave the market the confidence that the diversification strategy is working and that handsets are becoming less important.
• FQ2 26 revenues / Adl-EPS were $10.6bn / $2.65 broadly in line with consensus at $10.6bn / $2.56, but guidance was soft with FQ3 revenues / Adj-EPS expected at $9.2bn – $10.0bn ($9.6bn) / $2.10 – $2.30 ($2.20) was below consensus at $10.2bn / $2.40.
• The reason was solely due to the memory crunch that is hurting the handset market, where Chinese OEMs in particular have been reducing inventory and buying chips below end-market demand.
• This was where the bad news ended, and the market has finally chosen to look past quarterly volatility when considering the outlook.
  • First, smartphone bottom: where the company can see the bottom of the correction and expects to see QoQ growth to resume in the September quarter.
  • This does not mean that the memory crunch is over, but it does mean that customers will, from FQ4, be buying chips in line with end-demand rather than meaningfully below it.
  • This draws a line under the correction, meaning that the market can now see where the bottom is, and fears about a major collapse should now recede.
  • Second, data centre: was featured front and centre, where the company confirmed a custom silicon deal with one of the hyperscalers (I suspect Meta) that begins shipments towards the end of the year.
  • While Qualcomm would not be drawn on the size of the deal, it is hinting that it will be material, meaning that combined with revenues from Humain, estimates for fiscal 2027 should be moving up.
  • It was mainly this confirmation that convinced the market to look through the numbers and reverse a 7% decline in the share price, ending in a 13% increase over the last close.
  • Third, Automotive: which jumped to 38% YoY growth in FQ2 26 and is on track to exit the fiscal year at a revenue run rate of $6bn.
  • This implies a further acceleration of revenue growth in the fiscal second half and is being driven by a very strong showing at the Chinese OEMs, who are launching vehicles with a lot more compute power in them.
  • So far, Momenta and Horizon have yet to make a dent in Qualcomm’s position in the Chinese automotive industry, which I suspect has to do with its flexibility as well as its ability to offer chips to suit a wide range of vehicle trim levels.
  • It is the combination of China gaining share overseas and the increasing compute in the vehicle that is driving the step change in growth rate.
  • Automotive is also now big enough to start having a material impact on the overall outlook for the company.
  • Fourth, AI at the edge: where there has been much speculation about Qualcomm supplying the silicon for OpenAI’s product that it is working with Jony Ive.
  • When it comes to AI products at the edge, AI glasses are already shipping in volume thanks to Meta’s early start in this space, and here Qualcomm has a commanding position.
  • It is still not big enough to move the numbers, but the more people use agents, the more processing will be required at the edge, which plays directly to Qualcomm’s historic strengths.
  • Hence, this segment, which is also being helped with edge AI being used for numerous industrial applications, is on track to be meaningful in the coming few years.  
• The net result is that the numbers are precisely what I was expecting, but what has caught me by surprise is the market’s sudden change of heart when looking at the stock.
• If this is finally the moment when the market is prepared to look past Apple moving off Qualcomm and focus on the longer term and non-handset revenue, then there is a lot further to go.
• Qualcomm is one of the only semiconductor suppliers that has a credible claim to benefit from the AI boom, but where the shares have not had a very large increase in valuation.
• This is clear from the fact that the shares have increased by a paltry 6% in the last 12 months, while peers like MediaTek have increased by 91%.
• It is also clear from the consensus estimate that fiscal 2027 sees no EPS growth, which is clearly now too low.
• Hence, it remains one of the cheapest semiconductor companies with exposure to AI, and I remain very comfortable in continuing to hold it.