Intel: Zen And The Art Of Chip Production

What did I think about INTC and AMD last time?

While 2018 was a good year for both Intel and AMD, the first quarter of 2019 was very weak for both of them. And despite that, INTC was still above my buy under price. Given the general weakness in the semiconductor sector, I thought that patient dividend growth investors would eventually have the opportunity to get a great, growing company that is growing at a good price. AMD was still doing great even with a weak quarter, but it was still not having any significant impact on sales or revenue for Intel.

Based on the declared dividend and predicted dividend increase, I thought a good price for INTC was anything under $43. At the time, the market was offering shares a few dollars above my buy price.

Before I look at what is new for INTC, let’s look at AMD

Part of being a reliable investment partner is a management team that can predict future performance with a reasonable level of accuracy, so I always like to look at what the company said it would do this quarter. Below I show the slide from the Q1 presentation about what AMD expected for Q2.

Figure1 AMD Q1 Results

The guidance for revenue in Q2 looks pretty good, however, it is about $150 million less than what AMD actually achieved in Q2 last year. This points to a better Q2 than Q1, but still not a quarter as good as we have come to expect from AMD. As we can see from the slide below, actual results in Q2 came in a bit above the mid-point of revenue guidance. That is exactly what I like to see, management gives a range and comes in somewhere in the top half of that range. That tells me that management has a good handle on how the company is performing and that while they are conservative in their estimates they are also not deliberately low balling their estimates to make the actual results look better.

Figure 2 AMD Q2 Results

Q2 was still a tough quarter with less revenue than last year. However, on a positive note, while last year Q2 generated less revenue that Q1, this year Q2 saw an increase over the prior quarter. ASPs were pushed up, particularly on a year over year basis, by higher Ryzen sales. A greater volume of mobile chip sales in Q2 lowered APSs from the prior quarter. ASPs for GPUs were primarily driven by the pace of sales to the data center, which were higher than a year ago, but lower than Q1. While still a tough quarter for AMD, there are positive signs as well.

What did the latest earnings report say about Intel?

The first step in evaluating Intel’s performance in Q2 is to look at what Intel expected for this quarter when it reported results last quarter.

Figure 3 Intel Q1 Earnings Report

Q1 results caused 2019 expectations to be revised downward to a 3% decline in revenue with a 5% decline in EPS. Clearly, Q1 hit Intel pretty hard.

Figure 4 Intel Q1 Earnings Report

The slide above shows that Intel was clearly expecting a tough Q2 with revenue down 8% and EPS down 14% from the year ago quarter. While Q2 was a pretty impressive quarter last year, these drops are significant. Given the big hit Intel took in Q1, these projections seem quite reasonable.

So now let’s move on to the most recent earnings presentation, for Q2 2019.

Figure 5 Intel Q2 Earnings Presentation

Intel updated guidance again in Q2, and now 2019 looks a bit better. Now Intel is projecting that revenue for the year will only decline 2%, 50% less than before. And EPS is projected at $4.35, a 5 cent improvement over the Q1 projection.

So how did Q2 turn out relative to the projections from last quarter?

Figure 6 Intel Q2 Earnings Presentation

Revenue was down only 3%, not the projected 8% decline, which resulted in around $900 million more in revenue for the quarter. Operating margin, predicted to drop 4 ppt, only dropped half that much. And the EPS number came in at $1.06, a 2% increase not the decrease projected. Q2 was still a tough quarter for Intel. It turns out to have been less bad than it looked like last quarter, but a 3% revenue decline is never a good thing. Volumes still declined, even if much of that was made up by getting better prices.

Figure 7 Intel Q2 Earnings Presentation

The Client Computing group might look like good news. Revenue and operating income both increased compared to the same quarter last period. But a closer look reveals some information that isn’t so good. Look at the decline in revenue for the core CPU business. Segment revenue only increased due to a big increase in revenues for other things, like modem sales. Remember that Intel is exiting the 5G modem business so this can’t be counted on as a source of future revenue.

The news on margins is also mixed. While margins increased due to a richer product mix which pushed up average sale prices, this richer product mix is due to Intel not producing enough of its lower end chips to meet demand. This shortfall was enough to reduce volumes. While this shortfall was due to converting some production over to 10nm, that is still money Intel left on the table. Some of the margin increase was due to reduced spending on 10nm, since that is nearing full ramp. This looks to be good news, provided 10nm begins shipping for the holidays in significant numbers.

Figure 8 Intel Q2 Earnings Presentation

The results from the Data Center Group, previously one of Intel’s most profitable segments, continues to show significantly poorer performance compared to last year. Unit volumes were doing 12% (50% more than last quarters drop), while prices grew only a small 2%. Operating income took a big as well, down $900 million ($100 million more than last quarter's drop). The only good news is that much of the decline in margin was still due to spending on 10mn and further roadmap investments.

In the data center, AMD and Intel both had bad Q2. Neither company had great results in the PC segment either. I don’t think either company is in any danger, and I don’t see Intel’s dividend being at risk either. But clearly, both will have some struggles going forward.

What’s a good price?

To figure out a good price, I do a DDM calculation using my Excel^® based DDM calculator (pictured below, you can see the web-based calculator I based it on here and read a discussion on how the formulas were developed here). I also found this discussion of DDM, and note that in the article the author uses a discount rate of 5%.

While I usually look to David Fish’s CCC List (which contains data on companies that have raised their dividend each year for 5 or more years) Intel hasn’t re-qualified for inclusion on that list yet. Intel has been on the list in the past, and it is again increasing its dividend. I will calculate the dividend payments for the next 12 months based on the current declared dividend (for 1 payment) and 3 payments 1.5 cents higher, based on the amount of the last increase, for a total payment of $1.30. I will set the dividend growth rate for the next 5 years at 5%, with the terminal growth rate at 3% based on current yield.

Using those parameters, I calculate that the NPV of the predicted dividend stream is $43.11, which sets my buy price at $44. At just over $45 currently, INTC hasn't dropped quite enough to move it below my buy price. But it is getting close.

Using the 4-year average yield of 2.87% produces a price of $43.90, which is in line with my DDM calculation showing that INTC is somewhat pricey still. I don’t think the price is so high as to make small purchases, like dividend reinvestment, unwise, but I think it best to wait for larger purchases. With the price being under some pressure since the earnings release, it’s entirely possible a patient investor will soon be able to pick up shares at a good value.

What to watch for going forward?

Figure 9 Intel Q2 Earnings Presentation

Clearly, 2019 is not shaping up to be like the outstanding year 2018 was. Revenue, EPS, and Operating Margin are all down. Part of that is due to spending on ramping up 10nm production. Part of that is due to Intel just not being able to produce enough chips to meet demand (which is also tied to dedicating production to 10nm). The data center, once a big source of profit from Intel has been struggling this year. Some good news has Apple buying up Intel’s 5G modem business for $1 billion.

On the 10nm front, Intel already has two FABs in full production on 10nm products and began shipping them to clients in Q2. Everything looks on track to have 10nm products available in time for the holidays. 10nm server products look to be available in 2020 with full production in the 2 ^nd half. Supposedly 7nm (which will be a match for 5nm from TSMC) is on track for 2021. There is a lot of time till then, so we will see how closing that ends up matching when those chips actually ship.

Conclusion

Certainly, the first half of 2019 has been tough for both AMD and Intel. Neither had the growth and profits they had in 2018. Going forward, AMD is providing far more positive guidance than Intel. I still think 2019 results depends on each company’s ability to deliver meaningful quantities of chips with their newest technologies at profitable levels of yield. Intel seems to be doing what it needs to do to meet its goal of delivering 10nm products by the holidays. It is now even talking about preliminary schedules for 7nm products. There is still much Intel has to prove by actually delivering product, but this is good news.

Right now the market price is just above $45, still a bit above my buy under price. I think a little patience will allow investors to buy Intel shares at a very good value. And I think that until results come out for Q3 might be too long.

If you enjoyed this article and wish to receive updates on my latest research, click "Follow" next to my name at the top of this article.