Intel Options For
Total Page:16
File Type:pdf, Size:1020Kb
EQUITY RESEARCH Intel Corporation (INTC) Intel Target Change Options For Pat USA | Semiconductors January 14, 2021 RATING HOLD Key Takeaway PRICE $56.95^ We update our "Focus, Frugality and Fabless" framework originally published Jan 2020, MARKET CAP $239.8B and estimate a move to a fabless model could add $27 to the stock. We assess three PRICE TARGET (PT) $53.00 (FROM $50.00) strategic tracks INTC's new CEO, Pat Gelsinger, could pursue, and believe that a move UPSIDE SCENARIO PT $100.00 to a fabless front-end, integrated manufacturer back-end model would have the highest DOWNSIDE SCENARIO PT $40.00 long-term positive impact to shareholder value. We will assess Gelsinger's plan when it ^Prior trading day's closing price unless is articulated. otherwise noted. FY Dec Updated Focus, Frugality and Fabless Framework. In our Jan 2020 note, we argued that USD 2018A 2019A 2020E 2021E INTC should do three things to increase shareholder value. In this note we update that EPS 4.58 4.87 4.89 4.72 analysis: 1) Focus - we estimate that by divesting all of its memory business, it could FY P/E 12.4x 11.7x 11.6x 12.1x add $0.13 to EPS and $6 to its stock price; 2) Frugality - by driving its SG&A to industry *EPS changes reflect lower estimates and change from GAAP to standards, we estimate it could add $0.78 to EPS and $14 to the stock; 3) Fabless - we non-GAAP estimates estimate that by moving to a front-end fabless model, it could lower R&D by $4bn, CapEx by $10bn, morph into a cash flow and capital return machine that would add $1.36 to EPS Links to Related Research and $28 to the stock. We believe the Focus and Frugal steps could be executed within Jan 20, 2020 Table Set for Change: Focus, six months, but the move to a fabless model could take 2-to-4 years. Frugality, ...and Fabless? 3 Options for Pat. Intel announced that Pat Gelsinger would assume the CEO position Jul 23, 2020 Our Fabless Thesis Now a starting February 15th (see "Gelsinger to the Rescue"). He has yet to articulate his vision Higher Probability Event for INTC, but we think there are one of three paths he could take the company - we examine the pros and cons of each: Aug 03, 2020 A Butterfly Flaps Its Wings in 2005... Option 1: Front End Fabless, Back End IDM. In this option, INTC would move its front end manufacturing to a fabless model, but would continue to operate as an Oct 20, 2020 Finally Ditching Loss-Making integrated manufacturer on the back end. Pros: INTC has among the best back-end Memory...or Not? advanced packaging, assembly and testing (PAT) capabilities in the world, including Oct 19, 2020 Focus (sort of) is WIP, silicon photonics, and given the slowing transistor cost curve, we believe that increased Frugality and Fabless Hopefully Next differentiation in CPUs will come more from PAT and design than from transistor improvements. On the front end, we believe Intel has permanently lost transistor Oct 23, 2020 Would TSMC Take INTC...? leadership to TSMC and by proxy its competitors, and by moving its chip production to Jan 11, 2020 Jef U Graviton ARM CPU TSMC, it would neutralize its transistor deficit and potentially arrest its share loss. Cons: Expert Call + Transcript a move to fabless would likely take 2-to-4 years. If executed properly, we estimate that this could add $27 to the stock. Jan 13, 2021 Gelsinger to the Rescue...? Option 2: Stay an IDM. Pros: lowest execution risk. Cons: we've argued INTC has Mark Lipacis * permanently lost transistor leadership and appears to be falling farther behind. We would Equity Analyst expect this strategy to translate to continued share loss. (415) 229-1438 [email protected] Option 3: Move to a Fab-lite Model, where it pursues both its own internal leading edge Vedvati Shrotre * development and production, as well as an outsourced model. Pros: keeps the door open Equity Associate for internal manufacturing in the event it catches up to TSMC on the transistor, also, (415) 229-1574 doesn't preclude a move to a fabless model later if it determines it can never catch up [email protected] to TSMC. Cons: this is the most expensive option, that would require it to maintain $4bn Natalia Winkler, CFA * of annual R&D for its own process development, and incur additional R&D expense for Equity Associate its foundry design rules. (415) 229-1511 [email protected] Please see analyst certifications, important disclosure information, and information regarding the status of non-US analysts on pages 17 to 23 of this report. * Jefferies LLC / Jefferies Research Services, LLC EQUITY RESEARCH Intel Corporation (INTC) INTEL (INTC) The Long View | Scenarios Estimates Base Case USD 2018A 2019A 2020E 2021E Rev. (MM) 70,848.0 71,965.0 75,332.0 69,781.0 • Management divests memory • No actions are taken to streamline SG&A costs Previous • Share loss to AMD in desktop and server and NVDA in Consensus EPS 4.58 4.87 4.86 4.68 datacenter with increased pricing pressure Previous • Decline in PC unit shipments • 2021E EPS: $4.72; P/E FWD: 11.2x; Price Target: $53 EPS Upside Scenario Q1 0.87 0.89 1.45A 0.95 • Management moves to a fabless model and Previous streamlines SG&A costs Q2 1.04 1.06 1.23A 1.10 • Better data center and corporate PC demand from Previous COVID-19 driven work-from-home trends • Design wins in datacenter for parallel processing and Q3 1.40 1.42 1.11A 1.27 accelerator products (Nervana, Altera, GPU) Previous • Share gains with new GPU products Q4 1.28 1.52 1.10 1.42 • AMD mis-executes • 2021E EPS: $6.20; P/E FWD: 14x; Price Target: $100 Previous Downside Scenario FY Dec 4.58 4.87 4.89 4.72 COVID-19 outbreak extends beyond 2020 causing Previous • greater than expected demand destruction and supply chain constraints Valuation • Re-accelerating decline in PC unit TAM 2018A 2019A 2020E 2021E • Continued share loss to AMD, NVDA, XLNX and P/Rev 3.4x 3.3x 3.2x 3.4x homegrown solutions from hyperscale players EV/Rev 3.6x 3.5x 3.4x 3.6x • Mis-execution of manufacturing process technology transitions FY P/E 12.4x 11.7x 11.6x 12.1x • Gross margin compression due to under-utilization *EPS changes reflect lower estimates and change from GAAP to non-GAAP estimates • Valuation compression associated with lower growth outlook Market Data • 2021E EPS: $3.66; P/E FWD: 11x; Price Target: $40 52-Week Range: $69.29 - $43.61 | Investment Thesis / Where We Differ Total Entprs. Value $254.4B • Near term, we think INTC's Data Center and Notebook Avg. Daily Value MM (USD) 2,382.84 PC business is benefiting from Work-from-Home trends driven by COVID-19 shelter-in-place restrictions. Float (%) 97.3% • Longer term, we argue there is a tectonic shift in computing toward a parallel model, and as the incumbent with dominant share, we think INTC has the most to lose. • Our "4th Tectonic Shift in Computing" thesis argues that parallel processors will take share from x86 MPUs in the data center going forward. We also expect AMD will take x86 server MPU share from INTC in the data center. | Catalysts • Breakthroughs in manufacturing process technology or chip design. • Corporate PC refresh cycles. • Design wins for its parallel co-processor (Xeon Phi) or artificial intelligence products (Nervana, Movidius). • Adoption and ramp of 3D XPoint memory. • Design wins in Networking products for virtualized network appliances. January 14, 2021 2 Please see important disclosure information on pages 17 - 23 of this report. EQUITY RESEARCH Intel Corporation (INTC) Chart 1 - INTC Stock Price We estimate that: 120 1) Divesting memory could add $6 to INTC's stock price, $28 100 2) Lowering SG&A to sales-adjusted industry EPS +$1.36 norms could add $14 to INTC's stock price, P/E +2.0x and 80 $14 EPS +$0.78 3) Moving to a fabless manufacturing model Bull Case @ $100 $6 P/E +1.0x could add $28 to INTC's stock price. 60 $53 Fabless Model EPS +$0.13 EPS: $7.0 Base Case - ex P/E +1.0x P/E: 14x 40 Restructuring @ $53 FCF: $27.7bn INTC Stock Price ($) Stock INTC Consensus EPS $4.49 20 P/E 11.8x 0 ng Divest - ex - Fabless Move to Move Memory SG&A Discipline Base Case Base . Restructuri Source: Jefferies, Company data *Note that the sum of the individual restructuring vectors is slightly greater than our bull case January 14, 2021 3 Please see important disclosure information on pages 17 - 23 of this report. EQUITY RESEARCH Intel Corporation (INTC) Chart 2 - INTC/TSMC Lead on Nodes Historically, Intel would deliver a new transistor every 2 years, which would 45nm 12 Qs TSMC overtook INTC in translate to an 18-to-24 month lead over transistor leadership TSMC 10 Qs During the transition to 14nm, Intel's 32nm 11 Qs historical 8 quarter cadence extended to 11 quarters, and then the transition to 10nm 22nmFF/20nm, 6 Qs took 18 quarters, more than 2x its historical cadence of 8 quarters 16nm FF 6 Qs 9 Qs In July-20, INTC announced yields on its 7nm 14nm/10nm transistor were 12 months behind schedule, and that its 7nm launch would be delayed by 10nm/7nm 6 months to mid-2022, or three years after it 7nm/5mn 5nm/3nm had launched 10nm 8 Qs 8 Qs 11 Qs 18 Qs 12 Qs During its 4.5 year transition to 10nm, TSMC accelerated its transistor cadence to 6 quarters and in 3Q18, TSMC beat Intel to Jul-11 Jul-21 Jul-22 Jul-23 Jul-10 a smaller transistor for the first time ever, Jan-10 Jan-11 Jan-22 Jan-23 Jun-24 Jun-12 Jun-13 Apr-16 Apr-17 Feb-09 Feb-20 Feb-21 Oct-15 Oct-16 Sep-17 Sep-18 Dec-11 Dec-12 Dec-23 Dec-24 Mar-08 Mar-18 Mar-19 Aug-08 Aug-09 Aug-20 Aug-19 Nov-13 Nov-14 May-15 May-14 shipping its 7nm node (equivalent to Intel .