Tiny fabless semiconductor companies MaxLinear (MXL -0.14%) and Silicon Motion (SIMO -1.41%) are still awaiting approval for their merger. For MaxLinear, the acquiring company, getting its hands on Silicon Motion could be a game-changer as it battles against its far larger rival Broadcom (AVGO 2.41%) -- a top dividend stock, and one of the best-performing semiconductor businesses around over the last decade. 

Chart showing Broadcom's total return higher than MaxLinear's and Silicon Motion's since 2015.

Data by YCharts.

But MaxLinear's pending takeover of Silicon Motion has been a drawn-out process, and since the deal was announced, the chip market has fallen on hard times. Do either of these stocks offer value right now, or should investors just keep plowing cash into Broadcom stock? 

MaxLinear and Silicon Motion hit the skids, while Broadcom keeps advancing

With their merger still awaiting a nod from Chinese regulators (both companies expect approval by the end of this summer), MaxLinear and Silicon Motion reported ugly first-quarter 2023 earnings.

MaxLinear is a more broadly diversified company that provides internet and network infrastructure chips (roughly along the same lines as what Broadcom does). After a fantastic 2022, the current chip downturn that began with consumer electronics like PCs and smartphones has spread and finally come for MaxLinear too. Q1 revenue fell 15% from last quarter and fell 6% year over year to $248 million. Free cash flow fell even more sharply, down 71% year over year to $36.9 million.  

Things will get worse in the second quarter. Revenue is expected to fall 32% year over year (at the midpoint of guidance) to $190 million as its broadband internet service provider customers (which made up about one-third of revenue last quarter) work through excess inventory.  

As for Silicon Motion and its portfolio of memory control chips, it has been in the midst of a downturn for far longer, owing to its more square focus on consumer device markets (again, PCs and laptops) via the memory chips it accompanies. Nevertheless, Q1 was still ugly, with revenue down 49% year over year to $124 million. After maintaining positive free cash flow for several quarters of severe downturn, even this small and efficient business fell into the red. Free cash flow was negative $6.6 million. Given that it agreed to the acquisition, Silicon Motion is no longer providing specific financial guidance.

Meanwhile, one of my top semiconductor holdings, Broadcom, has been more than fine. Though its growth has also slowed, Broadcom has continued to chug higher. The chip design giant will report its last quarter on June 1, but it has forecasted year-over-year revenue growth of 7% to about $8.7 billion. That's not bad at all for an industry giant, while the industry it plays in is in decline.

Broadcom is the AI stock here, but is it the best buy now?

During its last earnings call, Broadcom had hinted at its data center networking chips getting a boost from the current arificial intelligence (AI) boom spawned by ChatGPT. For every Nvidia (NASDAQ: NVDA) GPU used in generative AI services housed in a data center, infrastructure needs to be strung together to help coordinate how all that data will flow through the system.

Broadcom CEO Hock Tan had said chip shipments for these AI systems brought in $200 million last year, but that sales from this end market could reach $800 million in 2023. Subsequently, Broadcom announced its new Jericho3 networking chip, custom-designed with AI in mind, that will no doubt aid in this rapid expansion.

The only problem is, Broadcom is huge, and data center AI is just a small segment of its overall business. With many of its peers already sliding into a cyclical slump, I worry that Broadcom could be next. Indeed, even companies with more focus on data centers are feeling a bit of a sting right now. Of course, Broadcom will be just fine even if it loses some steam. Cyclical slumps are normal every few years in the chip market. 

But for the next six months to a year, I wanted a rival chip stock to complement my investment in Broadcom -- rather than just plowing more money into the chip design titan. I decided to take a small position in Silicon Motion. The reason?

Silicon Motion seems to be further along in its slump, and is anticipating a gradual rebound in business in the back half of 2023. MaxLinear is early in its slump, and Broadcom could still get hit as 2023 progresses. If Silicon Motion gets acquired, shareholders will get a premium well above the current share price in the form of cash and MaxLinear stock (as of this writing, a more than 70% premium to where Silicon Motion is currently trading). Thus, owning Silicon Motion is also an investment in MaxLinear.

If Silicon Motion isn't acquired, I like it as a stand-alone business with its own well-established track record of profitable expansion.

With the chip industry overall expected to remain bumpy in the coming six to 12 months, I want a little diversification, and Silicon Motion looks attractive right now after getting clobbered by the market.

However, over the long term, Broadcom remains my primary bet in this infrastructure semiconductor segment. The company is well-diversified, highly profitable, and doles out tons of cash in the form of its dividend (currently yielding 2.9% a year) and share repurchases. It has all the ingredients to continue outperforming many of its peers. While I'm not currently buying more Broadcom, I remain a happy long-term shareholder.