April Newsletter 2019
Patience is THE Key to 2X Outperformance in an Ancient Bull Market
Well, today was made ultra sweet with our 2X overweight in AMD spiking 6% after market close on analyst beats on revenue and EPS ramps. Especially sweet since our investment thesis and $34 target is based on AMD
1) stealing market share from Intel in PCs with their supply constraint and lack on 7mm chip production --check
2) ramping GPU sales with new PlayStation ramp and new GPU chip intro--check
3) The new AI data center deals with Google highlighted in Google's disappointing earnings announcement--check.
But here is the lesson for this month--it's not the stock you own that makes you money, its the price you paid for the stock that makes you wealthy. We have had AMD buy under recs at $24-$26 and now $28. If you don't own it yet for some reason--don't chase it. We will undoubtedly get a fear-based sell-off for one reason or another. Next time we get one, I will add a Jan 2020 $30 call option if reasonably priced.
A subscriber sent me a note "Tobe--with the selling of call options on AMD I basically have a $2 cost basis on my 5000 shares--any thoughts on what do do?" My answer is "I don't know your portfolio or circumstances-- but I have a lot of low priced AMD after selling calls too--and I am collaring the position into January 2020. We don't sell advise on how to collar positions--yet--so call your broker and ask them how to "collar" you shares so you have 10% protection and 20% upside left--don't get greedy!"
My point here is my favorite investing axioms--"If you are not selling a stock/ETF, you are tacitly buying it" and " when you are so jacked up you feel you have to buy something, don't--and when you are so fearful of disaster (and our Transformity MacroMarket Index says 2-4% chance of a recession in the next 4 months) buy your favorite stocks/ETFs/options."
BTW--our TI Macro Market Index is up to 16.2 with the new blended NYC/Atlanta Fed 1.8% Q1 GDP analysis and Q2 NYC Fed Forecast of 2%.
Our BULLISH 16.2 reading means less than 2.5% chance of a recession in the next 4-6 months.
The initial GDPNow model estimate for real GDP growth (seasonally adjusted annual rate) in the second quarter of 2019 is 1.3 percent on April 29. The initial estimate of first-quarter real GDP growth released by the U.S. Bureau of Economic Analysis on April 26 was 3.2 percent
We have been served very well in this late stage of this bull market by buying on the big sell-offs and holding on to our favorite 2020 stocks. We have low volatility again--which provides the cover for $50 billion IPOs--but again one has to ask oneself since these companies are going public 10-12 years into their lifespan (with most having no shot at EPS for at least 5 years) is this an opportunity for us or a liquidity event for $billionaires and VC funds?
A few market mavens --Lary Fink CEO of money manager giant Blackrock as the leader-- are calling for a FOMO melt-up scenario for the overall market because 1) retail investors built a lot of cash in the December Flash Crash and 2) the Fed is done for the year. All logical for sure with real chain-weighted inflation stuck as 1.2-1.4% that the Fed is on hold but meltup?
We are back to levels and valuations of October 2018 but without a hawkish out-of-touch Fed Chairman. We all expect a "Worlds Greatest Trade Deal Ever in History" China trade deal with enough upfront goodies to help the Farm Belt who are being killed (and losing 2020 votes) but with the IP issues to be phased in and subject to inspection. What makes me most patient, however, is the SP Growth stock index and overall SPY are both way overbought--which tells me we have already had the "FOMO (fear of missing out) melt-up. AMD and Apple up tomorrow will take us even more overbought--and for the last 7 years of editing Transformity Investor (plus 20 years behind TI Pro) buying new long positions hitting 52-week highs in an overbought growth stock market makes little sense. If anything, we will SELL some calls on AMD after tomorrows run.
Here are the charts:
Note: ANY RSI reading over 70 is over-bought--market needs a rest.
What we are focusing on is secular growth bargains riding the unstoppable 3-5 year 30-50% secular growth waves of 5G, Artifical Intelligence/Machine Learning, Internet-of-Things, and the Enterprise Cloud.
Two over-done sell-offs in our research universe (we will be talking with CHGG Chegg tomorrow and will update on their lighter than expected guidance).We have been waiting for Xlynx to take a rest--and we got it. HYper Data Center chip sales took a breather Q1 after 12 straight quarters of 20%+ growth. This was bound to happen--double and triple ordering to get enough chips left excess inventory to be worked off (notice it did not affect AMD as they are taking share). Here is the headline:
Tech sees red amid data center woes
Apr. 26, 2019
The tech sector (NYSEARCA:XLK) drops 0.8% and the Philadelphia Semiconductor Index is down 1.6% in what's turned out to be a hard weak for data center players.
Yesterday, Intel (INTC -10.2%) reported a Q1 Data Center miss with $4.9B in revenue versus the $5.1B consensus. For FY19, Intel expects data-centric revenue down low-single digits Y/Y and DCG down mid-single digits.
The news is a negative read-through for Nvidia (NVDA -5.9%), which gets roughly 30% of its revenue from the data center.
This is the same day that Deloitte comes out to match the IDC research report that AI spending worldwide goes from $40 billion this year to nearly $80 billion in 2022. This growth is the once a generation convergence of 1) cloud computing, 2) Big data sets 3) Massive power data analytics. With early successes with AI, big biz is in the same "Get it done yesterday" mode like the late 90's exhortation of the CEOs to CTOs "Get us on the god damn Internet--whatever that means!" The business process optimization is aggressively being sold into health sciences, life sciences, healthcare, consumer and industrial goods, and of course media and telecom.
But Xilinx's biggest opportunity now is their 5G radios and technology stack for 5G rollouts worldwide. Field Gate Arrays are the play in the data center--in 5G Xilinx has a virtual monopoly on programmable chipsets. 5G is not the big driver yet, and the chips sell at a 8% lower margin, but we should see a shortage and "double ordering" wave come in Q1 2020. China slow down is real for data centers, but not 5G--they indend to have Huwaei be the world leader in 5G--and they can only use XLNX chips!
We had a nice run with XLNX from $55 to $85 last year--protected profits into the October meltdown. Here is another bit at the apple--Buy under $125 with $160 target. For options, let's wait for the $130 January 2020 to pull back in premium.
Update on Bloom Energy (BE)
Bloom is a great place right now. #1 PG&E in Northern CA has announced blackouts in fire risk areas--Bloom already sells 25% of its nat gas zero carbon emission power servers in CA. Bloom servers are a much better solution 365 for many companies with PG&E rate hikes coming.
Bloom Energy Buy Under $14 $28 target
This from a highly concentrated fund manager I admire-- “Tim” Johnston III whose firm’s Concentrated Equity Alpha portfolio owns just 25 to 30 stocks chosen by a four-person team led by Johnston. It has returned 330% since its inception through the end of March, versus 148% for the S&P 500 index. In Barron’s this week: “Bloom is a game-changer: Normally, when I hear the term fuel cell, I run the other way. Bloom makes energy servers, or boxes that use oxygen and natural gas, without combusting the natural gas, to create electricity. We’re seeing Moore’s Law at play. As they manufacture more servers, the cost per kilowatt hour comes down; [this could become] the cheapest source of energy in the world. Each new generation, which happens every five years, reduces the cost per kilowatt hour of energy by 30% to 35%. This is a massive market. We bought 1.5 million shares, so we’re committed. It’s 2.6% of our portfolio.
The servers provide energy to commercial buildings for as low as 11.4 cents per kilowatt-hour, versus 18 cents in Connecticut, 16 cents in California, and 14 cents in New York. And that cost per kilowatt hour goes down 33% in 2020 with their new iteration.
You will find Bloom servers on top of Morgan Stanley ’s trading headquarters in Times Square, and on the Staples Center in Los Angeles, and at the eBay/ PayPal data center in Utah (among many other buildings). "
Notes from factory visit. I recently met with Bloom Energy management in March and toured their giant full-tilt factory production line as well. Takeaways: 1) Gen 7.5 cost reduction is significant as the company reduces cell thickness from 115 microns to 100 microns, has less parts as natural gas conduit is etched out in the cell, uses a longer cell stack (200 vs 31 in 5.0) that improves throughput, and has a 50% higher power density.
Gen 7.5 testing is expected to end by 2019, initial prototypes in 2020, and commercial launch in 2021 - in line with prior expectations. The increased power density of Gen 7.5 will gradually lead to reduced overall product costs (per kW) by 33% through 2021 as shipments ramp through the year (4-5 quarter ramp).
■ Biogas and landfill opportunity are massive growing: Bloom Energy’s fuel cells can operate at even higher efficiencies with even landfill and biogas. The company estimates that renewable landfill gas in California alone can support 1.26 GW of electricity using Bloom Energy’s fuel cells, indicating that future landfill gas installations could tap into SGIP beyond 2020. This translates to $5B total addressable market (TAM) for CA landfills alone. Biogas is used in 10% of Bloom’s installations today. California tax credit has been renewed one before and ends in 2022--with the PG&E debacle and increased fire danger, my bet is it is extended to 2030.
We got a nice entry under $12.50 in April--hope we get another opportunity. But this one is truly a "hold forever" stock in my portfolio.
PS--my old real estate developer friend Wayne Ratkovich is building a massive 400,000 s.f. healthcare and government and college facility in the old CF Braun campus in So Cal--Alhambra to be exact. His team has ordered enough Bloom Servers to power the entire facility 24/7/365---and sell extra capacity to the CA grid.
I expect many new developments like the amazing new Hudson Yard facility on the West Side of NYC using Bloom Energy servers as both primary and back up power. Wal Mart has all their CA distribution centers 100% with Bloom--they are converting their CA stores every month. If you are in natural disaster areas--fire, flood, hurricane, snow, tornado--the Bloom server is cheaper to buy (per KW) and all nat gas pipelines under the ground.
More to come--updates on Nvidia and AMD after conference calls.