Intel reported a surprisingly good set of Q2 earnings yesterday. Revenue came in at $12.9 billion, above expectations, and they turned a profit of $0.36 per share, while the Street was expecting them to lose money. They guided Q3 EPS ahead of estimates as well, while their revenue guide was a bit light. With a guidance like that, most companies would have faced a mixed reaction for their stock, but Intel is not a normal company in normal times, and after-hours the stock was up strongly. There was a lot of good news in the stock, but Intel is not out of the woods yet.
Our thesis on Intel holds that the company faces three sequential challenges – fix their manufacturing process, reignite their product portfolio, and then find ways to fill up their fabs with Intel Foundry Services (IFS). The Intel team gave a pretty solid update on that first challenge. Their process plan – 5 nodes in 4 years – is on track. Their first EUV node is entering into production, and the critical 20A node looks to be ready to enter production soon. This is really good news for the company, solving an existential problem. In fairness, we should note that they still have two years to go before we really know if they can reach those 5 nodes, but probably the most important thing about the call was management’s tone on this topic. Past conference calls sounded forlorn, on this call they sounded confident, the Intel of Old is back. As we have noted all along, the company has plenty of talent to accomplish its goals, and these seems to be on display, finally.
Of course, that still leaves the two other challenges. It is far too early to touch on IFS. We believe they are making good progress, but it is going to be many years before we can even start to assess how well Intel can rise to the challenge of providing customer service to other companies. We will likely hear announcements this year about one customer or another signing up, and we have to read those as very preliminary endorsements. None of the major fabless chip companies will dedicate real volume to IFS until their manufacturing processes are fully ironed out, which is unlikely until the back half of the decade. So, we count challenge #3 as some progress but still too soon to tell.
Which leaves the company’s next challenge – can its products compete? Here we are left with more questions than answers. Truth be told, we heard more on the call that concerned than comforted. And here we return to normal company reporting patterns, reading the tea leaves of their comments. The company really wanted to assure investors that demand for their products was normalizing, but this is the type of normal where there is a long list of reasons why it looks soft. China demand has not surged back. Inventory has improved a lot but remains high in the channel. And our latest favorite, demand should recover in the second half of 2023, a line we have heard many companies use lately. Perhaps most concerning for us was their take that the demand for AI semis is displacing demand for other chips. We are fairly certain this is not quite right. There is definitely a near term issue as customers shift what is left of their 2023 budgets to accommodate AI demand, but that is at best temporary, and we have many data points that indicate budgets are actually increasing. Maybe Intel management is saying something similar to us, but the company’s Data Center segment continues to see declines and continues to lose money. There are an awful lot of moving parts in Intel’s latest financials, enough to lead us to believe that the company’s core CPU products are still struggling, and nothing in their outlook gave us much confidence that this would change soon.
That being said, we think Intel looks to be in much better shape. They still face a mountain of trouble, but they are starting to demonstrate their ability to overcome those.