Broadcom Stock Price Forecast: AVGO at $325 Ahead of AI-Driven Earnings

Broadcom (NASDAQ:AVGO) – AI infrastructure giant mispriced after deep correction

AVGO around $325 with trillion-dollar AI pipeline still not fully priced

Broadcom Inc. (NASDAQ:AVGO) trades near $325–$326 per share, versus a 52-week high around $415 and a range that starts near $138, leaving the stock roughly 20–25% below its peak while the business carries a ~$1.5T market value, a trailing P/E close to 68x and a dividend yield near 0.8%. On headline multiples AVGO looks expensive, but the internal profile is not priced like a mature, ex-growth story when you line up >50% expected EPS growth, AI semiconductor revenue projected to double, Infrastructure Software gross margin near 93%, group EBITDA margin around 67% and a disclosed backlog of roughly $73B scheduled to be delivered over about eighteen months. The recent selloff has been driven by three main narratives: fear that AI code generation will crush legacy software profits, concern that China’s push to cut foreign cybersecurity and VMware-type subscriptions will heavily damage Broadcom’s software cash flows, and worry that shipping full AI systems with third-party components will dilute gross margins and expose earnings. The actual numbers from the last reporting cycle indicate a very different story, with AI semiconductor and VMware franchises scaling aggressively, China already reduced as a share of revenue, leverage trending lower, and hyperscaler capex commitments moving sharply higher. At today’s level the stock reflects those fear narratives more than the hard data, which is why Broadcom real-time chart and the current valuation still support a Buy / bullish stance, with medium-term value closer to the $480–$500 band if management hits its own growth and margin targets.

Revenue engine – AI semis plus 93%-margin software instead of a pure chip cycle

Broadcom’s business mix is split between a semiconductor engine and a software engine, and the software side matters as much as the chips. The semiconductor solutions division covers high-speed networking silicon, AI accelerators and XPUs, custom chips for hyperscalers and other core infrastructure components. Alongside that, the Infrastructure Software segment, anchored by VMware and flanked by Symantec and other platforms, operates like an enterprise software franchise with very high incremental margins. For FY-25, Infrastructure Software generated around $27B of revenue, roughly 42% of Broadcom’s total, and grew approximately 26% year over year with a gross margin around 93%. Over the same period, the semiconductor division expanded about 22%. This combination means AVGO is not a linear bet on a chip cycle; it is a hybrid that monetizes both hardware demand and a deeply embedded software stack. Looking into FY-26, management has laid out a plan where semiconductor revenue grows around 50%, AI semiconductor revenue doubles year over year and the VMware-driven software base keeps the blended profitability profile elevated. With that product mix the company can reasonably guide to Q1 FY-26 revenue around $19.1B while keeping EBITDA margin near 67%, which is unusual for a business that still sits in the high-growth phase of a secular AI build-out.

AI capex super-cycle – $650B hyperscaler budgets feeding Broadcom’s backlog

The key macro driver for Broadcom’s next two to three years is the AI infrastructure capex cycle. Across the large platform companies, guidance and commentary now cluster around an estimated $650B of total capex in 2026, with a significant majority earmarked for AI infrastructure such as GPUs, XPUs, TPUs, networking, storage and power. Within that overall budget, one of Broadcom’s largest customers has pointed to a capex midpoint near $125B, another near $180B, explicitly emphasizing accelerated investment in AI data centers. These plans translate directly into long-dated networking and compute orders rather than abstract narratives. Broadcom sits at several points along this flow. It co-designs and supplies custom accelerators and TPUs for major hyperscalers, dominates high-speed Ethernet switches and routing silicon in AI data centers, and provides VMware Cloud Foundation as the private cloud operating layer that runs across Nvidia GPUs, AMD GPUs and Broadcom XPUs. When management disclosed a backlog of roughly $73B scheduled to ship over approximately eighteen months, that backlog reflected AI and hyperscaler demand rather than legacy on-prem businesses. A backlog of that magnitude relative to forward annual revenue guidance in the high-$70B band gives Broadcom unusually strong growth visibility into 2026 and 2027. If AI semiconductor revenue keeps doubling and moves toward three-quarters of the semiconductor mix by the end of 2026, the current flat share price since last autumn and the compression of the forward multiple from roughly fifty times earnings down to about thirty times are misaligned with the scale of the capex pipeline.

VMware and Infrastructure Software – why “AI eats software” does not apply here

The latest leg of the AVGO selloff was triggered by the idea that AI code generation will rapidly erode the economics of traditional software. Investors extrapolated from Anthropic’s and other players’ coding tools to assume broad risk for SaaS and enterprise platforms, then applied the same assumption to Broadcom’s Infrastructure Software and especially VMware. The product details contradict that view. VMware Cloud Foundation 9.0 is not a narrow point solution; it is a full private cloud operating system that orchestrates compute, storage and networking for AI and general workloads. It is designed to run seamlessly across heterogeneous hardware, including Nvidia GPUs, AMD GPUs and Broadcom’s own XPUs, and it is tuned specifically for high-throughput, low-latency AI environments. Version 9.0 added advanced NVMe memory tiering, so that fast NVMe devices extend the effective memory pool and allow more virtual machines or containers per physical host, which directly improves capital efficiency in AI clusters. Kernel-level optimizations delivered up to three times faster data throughput, lower latency and reduced CPU overhead, all of which matter when training and running large models. This is the deep plumbing that keeps multi-billion-dollar GPU clusters utilized; it is not a front-end application that generative AI can trivially replace. Independent assessments of VCF 9.0 describe it as a unified, cost-optimized private and hybrid cloud stack that directly challenges hyperscalers by offering a comparable experience on-premise at favorable economics. That is a classic wide-moat software position: it is deeply integrated into mission-critical infrastructure, tied to compliance and security models and extremely costly to rip and replace. Financially, a $27B license revenue base at around 93% gross margin provides a powerful buffer for the overall group. Even if certain legacy Symantec-type products see pressure from AI tools, the growth vector in VMware AI infrastructure software will dominate the segment mix and protect margins.

China exposure – from one-third of revenue down to the mid-teens share

China remains a material risk but the structure of that risk has changed compared with a few years ago. Public data suggests that in 2023 Broadcom generated over $11.5B of revenue from China, which represented approximately 32% of total sales at that time. By FY-25, China revenue was still around $11.1B, so the absolute level had barely changed, but this now accounted for only about 17% of the mix, because total company revenue had nearly doubled. That shift shows systematic de-risking: Broadcom did not abandon the Chinese market, but it diluted dependence by expanding faster elsewhere. Recent reports that Chinese authorities want domestic firms to reduce foreign cybersecurity software usage and reduce VMware reliance are clearly negative at the margin for Broadcom’s Infrastructure Software business in that region. However, in a world where AI-grade networking and accelerators remain supply constrained, it is unrealistic to assume that a given volume of Broadcom chips will sit unsold simply because one geography reduces direct purchasing. As long as global demand remains tight, the main question becomes which customers take which products and at what price, not whether the company can sell the hardware at all. A more serious downside scenario would involve hard restrictions on hardware imports and use, enforced tightly enough that volume cannot be rerouted through subsidiaries or third countries. That would take a significant bite out of the remaining ~17% revenue exposure. At this stage, though, the pattern of the last several years shows that Broadcom can grow strongly even while China’s relative contribution falls.

Margins, leverage and cash generation – deleveraging while spending for AI

Broadcom’s profitability and balance sheet look closer to a mature software and infrastructure compounder than to a typical cyclical chip stock. For the current year management has guided to EBITDA margin around 67%, an exceptionally high level given the scale of ongoing investment in AI platforms. Free cash flow margin is trending toward the high 40s percent, with projections near 48% as AI semis and high-margin software scale together. In FY-25 Broadcom reduced long-term debt by roughly $4.3B and cut total long-term liabilities by around $8.2B, demonstrating that it is investing in AI growth out of robust internal cash generation rather than through heavy incremental leverage. Dividend policy reinforces this picture. The company has maintained annual dividend increases near 10%, funded comfortably from free cash flow, and at a share price around $325 the yield of roughly 0.8% represents a growing income stream even if the percentage looks modest. The expected 100 basis point quarter-on-quarter hit to gross margin in Q1 FY-26 from higher pass-through costs on complete AI systems and racks, including third-party HBM memory, is manageable inside this framework. The key metric is that the EBITDA margin guidance remains intact, because the VMware software stack and higher-value AI semis revenue absorb gross-margin dilution while still expanding per-share earnings.

Earnings setup – probability of a Q1 FY-26 double beat is still high

The next catalyst is Broadcom’s Q1 FY-26 report, with the market currently expecting revenue around $19.1B and EPS near $2.02 on a post-split basis. Using the company’s guidance and its historical pattern of conservative estimates, a reasonable scenario has revenue landing closer to $19.3B, which is about a 2% beat, in line with typical performance against consensus. If net profit margin runs close to 52%, which is consistent with a 67% EBITDA margin after depreciation, interest and tax, and if the diluted share count edges up modestly toward 4.93B equivalent shares, then quarterly EPS calculates in the $2.03–$2.04 range. That would be a modest beat on the current consensus and can easily be exceeded if AI semiconductor revenue grows faster than projected or if VMware integration synergies drop through cleanly. The stock has already fallen roughly 18% since the last earnings print, primarily on China and gross-margin worries that are known and quantified. A result that delivers both a top-line beat and a margin or EPS beat, combined with any upward revision to the $73B backlog, would force a repricing of the near-term earnings path and likely re-anchor sentiment toward the medium-term AI growth story.

Valuation – forward P/E compressed while two-year EPS growth stays above 40%

The core of the Broadcom stock thesis is the relationship between its current valuation and its earnings trajectory. Late in 2025 AVGO traded near 50x forward EPS, reflecting the early phase of AI enthusiasm. After the correction from the December high around $414.61 down into the high-$280s to low-$300s and the current stabilization near $325, the forward multiple has compressed to around 30–32x. Over the same period, analyst models have moved EPS growth expectations to roughly 50.5% for FY-26 and about 40.5% for FY-27. If Broadcom hits those numbers and delivers FY-27 EPS around $14.4 on a post-split basis, and if the market assigns even a mid-range 30–35x P/E to that earnings power, then a reasonable target band in the next two years is $432 at 30x, $504 at 35x, with a mid-point scenario around $492 once modest upside to estimates is included. From the current $325 area, that implies in the region of 50% upside before any contribution from dividends. Relative to peers, this profile is compelling. If AI semis indeed move toward three-quarters of the semiconductor mix by late 2026, if Infrastructure Software continues to print >90% gross margins, and if hyperscaler capex plans hold near the $650B mark, then Broadcom justifies a structural AI premium closer to other leading compute and networking platforms. The present ~32x forward multiple therefore prices in fear more than it prices in the documented growth and margin expansion.

Technical picture and positioning – deep retracement into a support cluster

Price action over the last few months has already de-risked much of the valuation without any fundamental collapse. From the December peak near $414.61, AVGO slid almost 30% at the intraday low, briefly probing the $285–$300 range before stabilizing. Recent sessions show the stock churning in the $300–$335 band, with the latest close around $325.17, down $6 or about 1.8% on the day and a small additional move after hours. This region coincides with a visible medium-term support zone around $300–$310 where buyers previously stepped in when the stock was undergoing an earlier correction. Within the broader $138–$415 52-week range, AVGO now trades in the lower half of its recent distribution while its AI exposure, backlog, margins and balance sheet strength have all improved relative to the last time it was priced in the low $300s. For capital allocation, that combination of technical support and compressed valuation argues for accumulating positions in the $300–$320 zone when volatility allows, with current levels still acceptable for building or maintaining a core stake as long as the AI capex data, backlog commentary, VMware metrics and margin guidance remain consistent with the present story.

Risk map and decision – where the Buy case can fail and current stance

The bullish case for Broadcom stock relies on several assumptions that need to be monitored. A sharp reversal or delay in AI infrastructure capex, whether due to macro shock, political intervention or power-supply constraints, would undermine the $73B backlog and drag on revenue growth. A more aggressive and tightly enforced Chinese policy that moves beyond software subscriptions to hard restrictions on hardware imports and usage could put a larger portion of the ~17% China revenue share at risk and reduce flexibility to reallocate supply. Missteps in VMware integration, pricing or customer handling could slow the $27B ARR software franchise and erode the 93% gross margin that underpins group profitability. More severe and sustained gross-margin compression than the guided 100 basis point hit could also force a re-rating if EBITDA margin begins to fall materially below the 67% level, particularly if AI semis fail to compensate with enough volume and mix. Finally, regulatory pressure on networking or virtualization dominance could require changes in packaging or pricing that affect economics. At this point, none of these downside scenarios is visible in the reported numbers or in the capex and backlog data. With the share price near $325, a forward multiple around 32x, projected EPS growth above 40% per year for two years, and AI and software franchises both expanding, Broadcom (NASDAQ:AVGO) still justifies a Buy / bullish stance. The stock is best treated as a core AI infrastructure holding, accumulated on weakness in the $300–$320 range and monitored closely for any break in AI capex trends, VMware growth or margin discipline that would warrant shifting to a more neutral view.

That's TradingNEWS