Check out the latest Intel earnings call transcript.
Under previous CEO Brian Krzanich, Intel (INTC 0.62%) was fairly acquisitive. The company spent a whopping $16.7 billion on programmable-chip specialist Altera and another $15.3 billion for self-driving-car chip specialist Mobileye during his tenure.
These two acquisitions were supposed to expand the chip giant's total addressable market (TAM), and they did. But I'd argue that simply getting more aggressive in leveraging its core technologies in new markets translated into more TAM expansion than these acquisitions did.
How much did these acquisitions grow TAM?
First, let's look at how the Altera acquisition expanded Intel's TAM. In the world of field-programmable gate arrays (FPGAs), there are two major players -- Intel's programmable solutions group (PSG), which is what Altera is now known by, and rival Xilinx (XLNX). In 2018, PSG generated $2.12 billion in revenue and $466 million in operating income. During its past four quarters, Xilinx raked in a little over $2.9 billion. That means the TAM Intel is going after was worth a little over $5 billion over the past year, and it looks set to grow.
When Intel acquired Altera, it sold investors on the notion that having Altera's technology in-house could help boost its data center group business by allowing it to build processors with integrated FPGAs. The company reportedly shipped a specialized Xeon data-center part that combined a Xeon processor with an FPGA on the same package, but the company has been mum about further levels of integration.
It's harder to get a clear picture of the TAM that Mobileye opens up for Intel. In its acquisition announcement, Intel said that "we estimate autonomous driving could be up to a $70B TAM by 2030, when you factor in vehicle systems, data, and services."
Realistically, though, that number is inflated. In the acquisition announcement, Intel noted that Mobileye consensus revenue estimates for 2020 were less than $2 billion against a projected "vehicles systems and services" TAM of a little over $20 billion.
It seems the addition of Mobileye to Intel's portfolio may have expanded the company's nearer-term TAM by, at most, a few billion dollars. The company's longer-term projections, the ones that talk about the 2030 time-frame, don't mean much to me because it's almost impossible to make reliable predictions that far out.
Compare that with organic efforts
Let's consider some of the ways Intel is growing its TAM organically. The company signaled its intention in late 2017 to enter the market for standalone graphics processing units (GPUs), targeting both PC gaming and the data center. While those efforts will require incremental investments, the company is fundamentally expanding on graphics technology it already builds, rather than acquiring its way into a new market.
For perspective, graphics specialist NVIDIA (NVDA -0.86%), which commands the lion's share of the standalone GPU market, generated $8.14 billion from sales of standalone GPUs during its fiscal 2018 and nearly $8.2 billion in GPU sales through the first three quarters of its fiscal 2019. Although that figure was inflated by sales of GPUs to cryptocurrency miners, and that business is basically over, the point is that the standalone GPU TAM is a large one that Intel is, through organic means, going after.
That's not all, though, as Intel has also doubled down on its efforts to build memory products based on both 3D NAND and 3D XPoint technology. The memory market is in pretty rough shape right now, with Intel's non-volatile memory solutions group (NSG) falling short of expectations last quarter because of a "weaker NAND pricing environment" and suffering a $5 million operating loss in 2018, but the TAM the company is going after here is still significant. Intel didn't need to buy its way into the memory market, either.
On a related note, Intel is also leveraging the 3D XPoint memory technology NSG developed to build dynamic random-access memory (DRAM) substitutes, known as Optane DC Persistent Memory, in a bid to try to capture some of the DRAM TAM while simultaneously building a moat around its Xeon data center processor platforms. That means if customers want the benefits of Optane DC Persistent Memory, they need to use Intel's latest Xeon processors.
Acquisitions aren't all bad
I don't want to suggest that Intel not acquire technologies and businesses to help it grow. It's just that Intel spent $32 billion to get its hands on two companies that haven't exactly been game-changers so far. Perhaps those acquisitions will prove more valuable over time, but I can't help thinking the company spent a lot of money for a pretty modest TAM expansion.