Category: Semiconductors
Intel’s misery continues…
Intel’s (NASDAQ: INTC) significantly weaker-than-expected guidance for the coming quarter overshadowed better-than-expected first-quarter results.
For the coming second quarter, the Pat Gelsinger-led firm expects revenue to be between $12.5B and $13.5B, well below the $13.61B analysts were anticipating.
It also anticipates earning an adjusted $0.10 per share with adjusted gross margins of 403.5% and a tax rate of 13%. Analysts were anticipating adjusted earnings of $0.25 per share.
Shares fell more than 6% in extended-hours trading.
For the period ending March 30, Intel earned an adjusted $0.18 per share on $12.7B in revenue. The quarter is Intel’s first period in changing its reporting structure to focus more on its foundry business. Intel products, which now include client computing, data center, and network and edge, came in at $11.9B, including a 31% year-over-year rise in Client Computing revenue to $7.5B.
Datacenter and AI revenue came in at $3B, while revenue attributed to Mobileye (MBLY) was $239M, down 48% year-over-year. The Network and edge segment generated $1.4B, while the company’s foundry segment saw revenue decline 10% year-over-year to $4.4B.
Analysts expected a year-over-year increase in both the top and bottom lines, with the Pat Gelsinger-led firm expected to earn $0.14 per share on $12.78B in sales.
A consensus of analysts expected Intel to earn an adjusted $0.14 per share on $12.78B in revenue.
Comment made on the SA article.
ASML – Lithography, Monopoly in EUV machines, each machine costs north of $200 Mn, absolutely vital for lower node semis like 3nm, which is powering the latest I-phone among others.
ASML is at the pole position since it’s a critical component for AI and/or high-end chips. Every chip being planned or in production for AI acceleration or incorporating AI acceleration is produced on a 5 nm or smaller node that requires EUV
They also have the lower-end D-EUV machines, which are also successful.
Here’s a good link to the equipment technology market.
Lam Research on the other hand is strong in Etch and Deposition, which is exposed to cyclicality because it’s more of a mass market commodity with competition for AMAT, KLAC, and Tokyo Electronics. But the pickup from customers like Micron, which itself is riding the AI boom for high-grade memory equipment is a big benefit for Lam. To me the biggest strength is the resilience in the last 10 years with an EPS CAGR of 28% – that is a huge deal, cyclicals/commodity producers never get that.
Marvell Technology (MRVL) $74
Networking infrastructure – It had a down year, with cyclicality in China and slower data center growth. China is about 50% of revenues.
That said, the forecast for the next three years is good, with 24% revenue growth and 35 to 40% adjusted earnings growth.
A lot of this is riding on growth from Nvidia and other AI investments in data centers.
Like most companies in the sector Marvell has also appreciated 73% in the past year so valuations are a bit expensive, 11x sales, with cyclicality and China exposure, it’s definitely on the higher side. It has also financed its acquisitions with debt, carrying a lot of interest burden.
I would prefer to buy 10% lower in the mid sixties for a better return – there is an AI event on April 11th, which may have more specifics/catalysts. Will keep a look out for that.
Intel could lose as much as $1.5B in sales due to China ban; AMD less: Bernstein
The news that China is looking to ban purchases of Intel (NASDAQ:INTC) and AMD (NASDAQ:AMD) chips for government use could be a sizable topline hit to the two semiconductor companies, but perhaps not as much as some believe, investment firm Bernstein said.
On a nominal basis, Intel had roughly 27% of its sales in China in 2023 (or $15B), while AMD had roughly 15% or ($3.4B). However, government purchases were much smaller than that — approximately 10% or so — Bernstein analyst Stacy Rasgon said. As such, Intel could see a revenue impact between $1B and $1.5B, while AMD could lose a “few hundred million,” Rasgon said.
Here are ETFs that mirror the growing Semiconductor, Cybersecurity and Technology sectors, they have 4 to 5 Star ratings from Morningstar and have done exceedingly well this past year – over 40%. Which is a great performance but a drag going forward, because we’re entering at fairly high levels and very little chance of those gains. Nonetheless these have performed in the mid to high teens per year over a five year period and some have over a ten year – basically the underlying stocks are strong so in the long run, as we can see from their consistent performance.
Most Important: Spread your buying out in installments, on declines, anything we’re buying in 2024 is priced above their mean so we want as much of a bargain as possible.
Cybersecurity
CIBR – cybersecurity, mostly large cybersecurity companies, has the biggest names like PANW, CRWD, OKTA, FTNT etc, a good proxy for cyber security,
HACK – also cybersecurity, some small companies, but it has companies that specialize in military grade products, which is a bit of an advantage.
Technology
VGT – VGT is part of the Vanguard family, very well regarded and has all the biggest names in tech, half of the M-7, several cybersecurity companies, huge returns – even the 10 year return is like 18%, going forward if big tech performs this fund do very well. But given how well it’s done again don’t expect too much, anything in the 10-12% range per year for the next 5 should be good.
Semiconductor
SOXX – Largest semiconductor ETF also very successful, having Nvidia as a large holding will do that, but there are several semiconductor companies that haven’t done as well, which can be a drag. But that is common for a sector or any mutual fund or ETF, there will be mediocre and weaker companies, but they also tend to be less volatile, it’s not all bad.
Micron (MU) Rating Upgrade to Buy from Hold, $100.
Results expected this afternoon were very good, and I am more optimistic about the guidance. I was hesitant to add or recommend buying because it looked overpriced compared to its historical average and it had doubled in the past year.
Nvidia’s comments on needing more high bandwidth memory (HBM) vendors like Samsung, suggest the Micron is more likely to have challenges meeting demand. Unlike the past year when they had to discount inventory.
With this beat and these upgrades from Wall Street analysts in Barron, I would start buying.
“Micron is likely to report continued soaring demand for “high bandwidth memory,” or HBM—parts that combine multiple DRAM chips to improve data-processing speeds.
TD Cowen analyst Krish Sankar wrote in a recent research note previewing the quarter that when it comes to Micron, “HBM remains the centerpiece of attention.” Last week, he lifted his target for the stock price to $120, from $100. He said there is a “potential scenario” where the stock can reach $150, for a gain of more than 50% from current levels.”
For the May quarter, the Street is projecting revenue of $5.98 billion, with an adjusted profit of 8 cents a share. Analysts expect the rebound to continue from there. Estimates for the August quarter now point to $6.86 billion in revenue and an adjusted profit of 81 cents a share.
FQ3-24
Revenue 6.6Bn Expected 5.8Bn
EPS $0.17 Expected $0.08
| FQ3-24 | GAAP(1) Outlook | Non-GAAP(2) Outlook |
| Revenue | $6.60 billion ± $200 million | $6.60 billion ± $200 million |
| Gross margin | 25.5% ± 1.5% | 26.5% ± 1.5% |
| Operating expenses | $1.11 billion ± $15 million | $990 million ± $15 million |
| Diluted earnings per share | $0.17 ± $0.07 | $0.45 ± $0.07 |
Wedbush analyst Matt Bryson wrote in a recent research note that recent trends in prices for both DRAM and NAND memory chips suggest Micron will beat its guidance for the quarter. Bryson, who has an Outperform rating on Micron shares, said he expects positive commentary from the company on the outlook for HBM demand.
“Since last summer, management has provided consistently optimistic commentary around anticipated progress with HBM in light of the technology being a derivative of their highly successful standard DRAM nodes,” Bryson writes.
Meanwhile, analysts say the balance between supply and demand has stabilized following a supply glut that spanned multiple quarters.
“Customer inventories have largely normalized, demand conditions across markets appear stable, and supply growth remains muted,” Raymond James analyst Srini Pajjuri wrote in a research note previewing the quarter. “In addition, HBM is a significant secular driver that could add $1.5-$2 in incremental EPS at the next peak.”
Pajjuri maintains an Outperform rating on the stock.
SOXX, The premier semiconductor ETF, naturally done very well with the AI, Nvidia boom. Up 62% in the last twelve months. Solid for the long term, but I would wait for a lower entry point, 10-20% lower to get meaningful returns. Given the froth in the market for all things AI, we may not get that decline, though, I think patience is better.
XLK, Very similar to SOXX 48% up in the last 12- same thing wait for a better price, but again long term very solid. The big issue is so much of the future gains have already been priced in so going forward it’s going to be much lower than the 48% plus you have the risk of the ETF dropping or staying sideways.
VOO – follows the S&P 500, is about 28% up in the past year, and 8% YTD – not surprisingly because the S&P 500 has about a 30% technology influence, its a market capitalization weighted index so the big tech bellwethers like Nvidia, Microsoft dominate its movements.
Many analysts have already crossed their S&P 500 targets for 2024, and just a few are revising it upwards, therefore not likely to see too many Buy Calls from these levels.
The longer-term average annual move in the S&P 500 is usually around 8%. We’ve already advanced 8% in the first two months, so the same story, currently overbought – would prefer a better entry point on a correction.
Nvidia’s results exceeded expectations as usual, kind of becoming a habit! The previous quarter’s (Jan 2024) revenue beat by $1.6Bn and it guided Q1-FY25 (April 2024) revenue 10% or 2Bn higher to $22Bn.
Shares took off from $675 to $725. Everybody’s happy.
Now comes the tough part.
Nvidia had a net profit margin of 55% = $12.2Bn in profits on $22Bn in sales. That is drug lord margin territory! Simply, they can charge whatever they want for the H100s, the new Grace Hopper, and the H200s that are coming down the pike. I’m confident that these margins will continue for at least a year untill competitors get their act together.
However, to assume that these margins will continue beyond that is difficult to swallow, and most of the street estimates for earnings are based on at least 52% in NPM, which if not achieved can be a huge disappointment.
So I modeled earnings at a 40% Net Profit Margin, which is similar to a big pharma company’s patented drug margin that also charges as much as the market can pay for it.
With that NPM, Earnings come down naturally; three years down the road in the 40% model, EPS is $26 compared to the street estimate of $33. Assigning a P/E of 40, that gets us to $1,030 from today’s price of $725 or an annualized gain of 12%. And if the street is correct, we’re looking at 40*33 =$1,320 or an annualized gain of 22%.
The counter argument to the lower margin thesis is – Nvidia can lower prices and sell more, and at some point this is likely to happen – the overall growth doesn’t reduce – especially if you’re changing the whole paradigm of accelerated computing replacing the way data centers are built now.
At the moment, I’m not planning to add any more, my exposure to Nvidia is already very high, and the long-term thesis doesn’t change.
Nvidia (NVDA) $715-$720. Planning to take profits, and reduce position by about 10% this week.
Nvidia reports Q4-FY24 earnings after market tomorrow, 02/21 and expectations are high for an impressive beat and raise for FY2025. Nvidia has a January year-end.
Nvidia is the largest holding in my portfolio and I need to reduce it a bit to keep my risk rules and parameters intact. I also believe that expectations are a little too rich for my liking and anything less may be hammered, instead of the usual earnings pop there could be a selloff. Nonetheless, it remains a great investment over the long term and I wouldn’t sell more than 10-15% of my position. Just pure profit-taking and risk control.