"Standing in the light", Broadcom Inc. (AVGO.US) solidifies its cash flow base with a $30 billion debt repurchase. The bullish market logic of "optical interconnect + AI ASIC" computing power is further strengthened.

One of the big winners in the global AI boom, Broadcom Inc. (AVGO.US), announced on Thursday that its cash tender offer to purchase its six important series of outstanding senior notes has expired and the final results have been released. The cash tender offer was aimed at buying back its high-yield senior notes. The chip giant, focused on AI computing infrastructure, also announced that the total purchase price (excluding accrued interest payments) has been significantly increased from the previously announced $25 billion to $30 billion. The increased offer amount is enough for Broadcom Inc. to purchase all validly submitted 4.926% Senior Notes due 2037 and 4.900% Senior Notes due 2038, including a series of notes submitted through a guaranteed delivery procedure. The latest $30 billion note purchase is aimed at cleansing part of the long-term debt, optimizing duration and credit curve without sacrificing growth investment, and retaining strong financial flexibility for future investments in AI ASICs, Ethernet switch core chips, optical interconnect systems, and advanced packaging. The $30 billion note tender offer itself is not a major catalyst related to revenue generation in AI, but it reinforces Broadcom Inc.'s growth narrative of "cash flow manufacturing machine + AI computing infrastructure seller shovel." While NVIDIA Corporation dominates the data center AI GPU computing architecture core, Broadcom Inc. is increasing its efforts to compete for the large incremental market share of AI ASIC custom chips, open Ethernet infrastructure, and CPO such as data center optical high-speed interconnect infrastructure. As Broadcom Inc. repurchases senior notes with a $30 billion cash tender offer and optimizes its balance sheet, it is also positioning itself with AI ASIC/XPU, 102.4T Ethernet switch, CPO interconnect, and 400G/lane optical DSP technologies at the critical intersection of "computing power + interconnect" in the AI factory. The latest statement shows that the cash tender offer is not for Broadcom Inc. to spend $30 billion to acquire another company, but for Broadcom Inc. to issue a "cash tender offer" to its bondholders: whoever is willing to sell back their Broadcom Inc. senior notes, Broadcom Inc. will buy them back in cash, essentially repurchasing and reducing some of the outstanding debt. Originally, Broadcom Inc. planned to spend up to $25 billion to buy back these bonds, but now the cap has been raised to $30 billion because the scale of the notes submitted for sale by investors is larger. Ultimately, about $5.5 billion of principal notes were validly submitted, with approximately $2.9 billion of principal being accepted for purchase. In essence, Broadcom Inc. is using cash to repurchase its long-term bonds, optimizing its debt structure and reducing future interest and refinancing pressure, rather than engaging in equity mergers or asset acquisitions. Broadcom Inc. is one of the core chip suppliers for Apple Inc. and other major tech companies, as well as a key provider of high-performance Ethernet switch chips and high-speed optical interconnect core devices in global large-scale AI data centers. With the announcement of the expiration and final results of the latest cash tender offer, Broadcom Inc.'s stock price rose nearly 3% in after-hours trading on Wednesday, highlighting the market's increasingly optimistic bullish sentiment. As AI data centers increasingly rely on the power of light, the "optical interconnect + AI ASIC" engine is igniting a new round of bull market for Broadcom Inc. Broadcom Inc. is making a significant breakthrough in the direction of self-developed AI ASIC accelerators by Google TPU (TPU technology route is a typical AI ASIC chip), OpenAI/Anthropic and other cloud giants, entering the largest incremental market of custom AI chips outside of NVIDIA Corporation's GPUs; AI ASIC is undoubtedly the first core pillar of valuation and fundamental growth for Broadcom Inc., and the strong financial reports announced recently by Broadcom Inc. and Marvell are sufficient to demonstrate the unprecedented strong growth logic of AI ASIC is being rapidly confirmed by "financial report-level evidence." Broadcom Inc. expects AI chip sales to exceed $100 billion by 2027, with Q2 FY2026 AI semiconductor revenue expected to reach $10.8 billion, a year-on-year increase of 143%, indicating that the demand for AI ASIC/XPU has become a core growth curve. On the other hand, all GPU, TPU, and ASIC routes cannot be separated from high-speed networks and optical interconnects. Broadcom Inc.'s Tomahawk 6 has achieved a single-chip 102.4Tbps switching capacity and showcased 102.4T Ethernet, CPO, 400G/lane optical DSP core devices, PCIe Gen6, and other AI interconnection solutions at OFC 2026, precisely bridging the bottleneck from "single-chip computing power" to "thousands of cards/tens of thousands of card cluster efficiency" in AI data centers. The bottleneck in AI data centers is shifting from single-card computing power to cluster interconnection, east-west flow, power efficiency, optical module power consumption, and network stability, which is the strategic territory of Broadcom Inc. A single chip determines the limit of computing power, but the real efficiency of thousands, tens of thousands, and even terawatt-level AI clusters depends on whether there is sufficient bandwidth and low-power communication between chips, within AI data center internal server racks, and between multiple data centers. This is where the strategic value lies in Broadcom Inc.'s Ethernet switches, CPO, optical DSP, and network silicon photonics routes. Therefore, Broadcom Inc.'s new strong bull market logic is not simply about challenging NVIDIA Corporation, but about becoming a supplier of massive custom cloud AI inference resources, data center optical interconnect systems, and network infrastructure bases at the "AI factory" level: benefiting from the cost reduction of custom ASICs by cloud providers while benefiting from the strong demand for optical interconnect, Ethernet switch, and data center network infrastructure driven by all types of AI computing clusters. Like its competitor Marvell, Broadcom Inc. also occupies a core position in data center optical interconnect and network infrastructure, which can be said to be the key moat that sets them apart from other AI ASIC or AI chip suppliers. Broadcom Inc. not only produces custom AI chip/XPU for cloud providers but also covers critical interconnect components in AI clusters such as Ethernet switch chips, SerDes, optical DSPs, re-timers, AEC, PCIe switches, co-packaged optics (CPO), etc.; for example, Tomahawk 6 has achieved a single-chip 102.4Tb/s switching capacity, geared towards AI cluster scale-up and scale-out networks, and Broadcom Inc. recently showcased 102.4T CPO Ethernet switches, 400G/lane optical DSPs, 200G/lane Ethernet re-timers/AEC, PCIe Gen6 switches, re-timers, and other end-to-end AI connectivity solutions at OFC 2026.