Broadcom is emerging as a major player in the rapidly evolving AI networking sector, boasting an impressive non-accelerator AI backlog of approximately $20 billion. This surge in demand for reliable, high-speed networking solutions is largely fueled by the increasing scale of AI clusters, which are now frequently exceeding 100,000 compute nodes. Hyperscale cloud providers are adopting these massive infrastructures to facilitate extensive training and inferencing workloads, necessitating significant advancements in networking capabilities.
Recent trends indicate that the bottleneck in AI infrastructure is no longer solely constrained by computational power; rather, issues of bandwidth, latency, and reliability in networking interconnects are becoming increasingly critical. To effectively manage data transmission among vast arrays of GPUs, CPUs, and switches, hyperscalers are deploying tens of thousands of optical modules.
In response to these demands, there has been a marked shift toward 800-gigabit (800G) optical modules, with early adoption of 1.6-terabit (1.6T) silicon photonics-based interconnects on the rise. While the investment community has largely concentrated its focus on GPUs in this area, there are significant opportunities for companies like Broadcom and Coherent, which are poised to capitalize on the burgeoning AI networking trend anticipated to materialize by 2026.
Broadcom’s recent performance underscores its strategic positioning within the data center infrastructure landscape, particularly as hyperscalers advance their projects in anticipation of tailored accelerators. The company exited its fiscal 2025 with an AI switch backlog surpassing $10 billion, bolstered by unprecedented bookings for its high-capacity Tomahawk 6 switch, which can handle 102 terabits per second. Furthermore, Broadcom enjoys a substantial total AI-related backlog of $73 billion slated for delivery over the next 18 months, with nearly $20 billion attributed to non-accelerator products—indicating a significant revenue potential from networking and optical components.
In addition to its strong networking offerings, Broadcom’s custom accelerator segment continues to drive growth. The company’s revenue from AI semiconductors surged 74% year-over-year, reaching $6.5 billion, with expectations of doubling to $8.2 billion in the upcoming quarter. This diverse product portfolio positions Broadcom favorably to capture a significant market share amid the global AI infrastructure renaissance.
Similarly, Coherent is carving out its niche in the AI networking space by providing crucial optical and photonic components that facilitate high-speed data transmission within data centers, cloud networks, and telecom infrastructures. For Coherent, the first quarter of fiscal 2026 showed a 17% year-over-year revenue increase to $1.58 billion, with data center revenues specifically rising by 23% due to heightened demand for its advanced optical connectivity solutions. The company also achieved record bookings during this period, underscoring a positive outlook with orders stretching well into the future.
Coherent is advancing its product offerings with optical transceivers capable of supporting several cutting-edge technologies, including silicon photonics and electro-absorption-modulated lasers (EML). As more customers expand their AI networking capabilities, the broader adoption of these technologies is anticipated in 2026, positioning Coherent for significant growth as the market trends toward faster and more efficient network solutions.
As the landscape of AI-driven networking continues to shift, investors may want to explore opportunities beyond established players like Broadcom. Analysts recommend considering a list of ten other companies poised for strong performance in the sector, suggesting that while Broadcom is well-positioned, there may be untapped potential in alternative investments that could yield considerable returns.
