Description

Arrow distributes semiconductors and related electronic components (Global Components - 77% of sales) and distributes IT hardware and software to value-added resellers (Enterprise Computing Solutions - 23% of sales) through 220 sales offices and 39 distribution centers in 85 countries. Arrow is a classic distributor, with 600 suppliers and 210,000 customers. 

Products distributed include high value semis like GPUs, but mostly consist of a huge variety of cheap electronic components (think more like the electronics in a children’s toy and less like Nvidia H100s). Arrow's website highlights the breadth and depth of the categories they touch:

[A prior write-up by singletrack might provide more helpful background]

The role of distributors

Distributors aggregate customer demand and allow suppliers to ship to distribution facilities in bulk. Suppliers use distributors as an extension of their go-to-market, with a few large customers served directly and the long tail served by distributors. The long tail of small customers requires varying amounts of hand holding that suppliers would rather leave to distributors. 

Distributors reduce complexity for customers, who would rather deal with a single distributor than have a product order that includes many suppliers. Single product orders frequently span 10-50 suppliers, and Arrow’s Americas linecard alone has 430 suppliers. 

Distributors solve the mismatch between semiconductor manufacturing timelines and visibility into product demand. The standard delivery time for semiconductor manufacturers is 12 weeks, whereas customers might have only a few weeks of visibility. Distributors fill that gap, coordinating and preparing adequate safety stock to narrow the difference.

Additionally, in about a third of sales, Arrow goes beyond simple hand holding to actually helping in product design. Arrow’s 22,000+ employees include over 2,000 design engineers who assist customers in designing a product or using Arrow’s reference designs (basically a library of Arrow designed semis that provide some functionality). In this motion, which Arrow calls demand creation, Arrow registers a customer design with suppliers and Arrow captures higher margins over the life of the product. Suppliers like this because it provides them with more certainty and visibility into production needs. 

Why is Arrow cheap?

It’s hard to believe with the constant drumbeat of AI news, but we are in the middle of a downturn in semiconductors. Quarterly sales growth at suppliers like Texas Instruments has been negative for over a year, and inventory levels remain elevated. Arrow is a short cycle business with limited near-term visibility, and the financial results and stock have been whipsawed by this dynamic. Additionally, the management team is not promotional (last I-day was in 2012) and the stock has poor sentiment on the sell-side, with a single buy and a bunch of sells/holds. It is unclear when the cycle will end, but Arrow is cheap to its earnings power, and the buybacks and countercyclical cash flows ameliorate the downturn.

This downturn looks similar to prior downturns so far (chart below), which suggests we're close to the bottom of the cycle. 

Is Arrow a good business?

I’m sure a good portion of VIC members recognize Arrow as a perennial inhabitant of screens for low multiples (of anything). While you don’t need a multiple for this stock to work, I think the quality of this business has been underrated for some time. Returns have been decent over a long period:

Free Cash Flow is lumpy but countercyclical: 

Top line growth is low, but buybacks at low multiples have led to decent per-share growth:

If you compare the valuation to the CAGRs in Revenue, EPS, and share count, Arrow stacks up well (shown here relative to a histogram of S&P stocks):

Usually characteristics like these suggest a business that has secular challenges. I don't see anything imminent here. Major suppliers have at times shifted their favored distributors, or they have chosen to change the mix of direct versus distributed volumes, but Arrow has maintained market share over time.

Valuation

Arrow is a cyclical business, so I value it by looking at an appropriate multiple of mid-cycle earnings power. Management has said that more design work/demand creation activities leads to higher margins, but this mid-cycle case contemplates lower margins than their guidance. This is essentially a mean reversion bet where you don’t need a higher than historical multiple for the stock to work. Time is usually your enemy on these types of bets, but when you have large buybacks, as you do with Arrow, languishing at low multiples isn't thesis breaking. 

10x midcycle earnings of $14.74 gets you to a $147 stock:

Arrow is cheap versus its own historical multiples:

Risks

- Supplier consolidation

- Suppliers bringing more distribution in-house. A few years ago, Texas Instruments did this and competitor Avnet suffered as TI moved most of its business to Arrow. 

- Higher levels of competition from lower margin Asian competitors

Summary Financials and Projections

Catalyst

Buybacks, mean reversion