Broadcom's AI revenue is on pace to reach $115 billion in fiscal 2027—a jaw-dropping projection that reflects an infrastructure race so intense it's rewriting the semiconductor playbook. What's striking is that this forecast is now being driven not just by NVIDIA's dominance, but by a proliferation of custom chips designed specifically for OpenAI, Google, and a handful of mega-cap AI players who refuse to be locked into single-vendor dependencies.
The Custom Chip Revolution: Why Broadcom's $115B AI Forecast Matters
For years, NVIDIA held an almost unshakeable grip on the AI infrastructure conversation. But by September 2026, that narrative has fundamentally shifted. Broadcom's projected $115B in AI revenue for FY2027 signals something far more sophisticated: the world's largest cloud platforms are now building their own silicon, and they're using Broadcom's design, packaging, and manufacturing expertise to do it. OpenAI has been collaborating with Broadcom on proprietary inference chips that optimize their language models, while Google's TPU (Tensor Processing Unit) evolution continues to lean on Broadcom's advanced interconnect technology and chiplet integration capabilities. This isn't competition replacing NVIDIA—it's complementary infrastructure multiplication.
The reality is that custom chips represent a margin expansion opportunity that traditional analysts often overlook. When Google or OpenAI design their own silicon with Broadcom's help, they're not just buying commodity processors. They're engaging in long-term design partnerships worth billions annually. Broadcom's custom silicon revenue stream is growing at a clip that makes their traditional networking and storage businesses look stagnant. In the second quarter of 2026 alone, Broadcom reported over $14 billion in quarterly revenue with AI-related segments growing 73% year-over-year. That trajectory puts the $115B annual AI revenue forecast well within striking distance by next fiscal year.
OpenAI and Google: The Demand Engine Behind September 2026's Momentum
OpenAI's partnership with Broadcom marks a watershed moment in AI infrastructure. After years of relying on NVIDIA GPUs for training and inference, OpenAI has committed to a multi-year custom chip roadmap designed specifically for their model architectures. The efficiency gains are material: internally benchmarked custom inference chips can deliver 2.3x better performance-per-watt compared to generic GPU solutions. That's not marginal improvement—that's the kind of operational leverage that justifies massive capital expenditures. For traders, this signals that OpenAI's capex spending, which exceeded $7 billion in 2025, is now flowing directly into silicon design partnerships. Broadcom wins both the design work and the manufacturing orders.
Google's position is even more strategic. With their TPU lineage already established, Google is now pushing Broadcom to co-design next-generation interconnect fabrics that can tie together 10,000+ AI accelerators in a single training cluster. Here's what most traders miss: this isn't just about Google's internal ML workloads anymore. Google Cloud is actively selling AI infrastructure services to enterprise customers, which means the custom chips Broadcom helps design become a competitive moat for Google Cloud's market share battles against AWS and Microsoft Azure. When you factor in that Google Cloud's AI and databases revenue segment grew 26% year-over-year in Q2 2026, you see how infrastructure investment flows directly to top-line growth. That visibility is exactly what's driving Broadcom's September 2026 momentum.
"The custom chip era isn't replacing GPU demand—it's creating a parallel infrastructure market where integration partners like Broadcom capture design fees, manufacturing contracts, and long-term support revenue streams that generate recurring 70%+ gross margins."
Trading Broadcom's September 2026 Setup: Levels and Entry Points
From a technical standpoint, Broadcom has built a compelling setup heading into late September 2026. The stock has consolidated between $175 and $198 through much of August, with volume drying up—a classic accumulation pattern ahead of major earnings cycles. The $115B AI revenue forecast, if confirmed in the company's September guidance update, would likely trigger a breakout above the $198 resistance level with $215-$225 as a realistic near-term target. Volume on rallies should exceed 25 million shares for the breakout to have conviction. What's critical for traders right now: this isn't a pure momentum play. It's a fundamental re-rating based on visibility into multi-year custom chip contracts with world-class counterparties. Compare this to September 2025, when similar infrastructure announcements barely moved the needle. The difference in 2026 is that institutional money now understands the margin and revenue visibility embedded in these partnerships.
Options traders should watch the September 27 expiration for call spreads targeting $210/$225. The implied volatility crush post-earnings typically creates opportunities to define risk tightly while maintaining 2:1 reward-to-risk ratios. Position sizing matters here because while the fundamental narrative is strong, Broadcom's stock can be range-bound for weeks between major announcements. Don't chase into the initial breakout—wait for a pullback to the $205-$210 zone with volume confirmation. The real money in Broadcom through 2027 won't come from chasing breakouts; it'll come from understanding which custom chip partnerships generate the highest gross margin revenue and building positions accordingly.
Risks, Reality Checks, and the Road to $115B
No forecast is risk-free, and the $115B AI revenue projection carries real execution risk. First: custom chip development timelines are notoriously difficult to predict. OpenAI or Google could face technical setbacks that delay next-generation chip rollouts by 6-12 months, pushing revenue recognition further out. Second, NVIDIA's product roadmap isn't standing still. Their Blackwell and next-generation architectures are increasingly customizable, which could erode some of Broadcom's competitive advantage in the "semi-custom" segment. Third, geopolitical risk around semiconductor export controls and Taiwan strait tensions could disrupt manufacturing capacity. If TSMC faces supply chain disruptions, Broadcom's ability to deliver custom chips to OpenAI or Google gets squeezed regardless of design partnership strength.
That said, the forward-looking reality is compelling. Broadcom's management has been consistently conservative with guidance, and they're now openly discussing AI custom silicon as a core strategic pillar—not a one-time windfall. The company is investing heavily in design centers and engineering talent specifically for AI partnerships. When management starts allocating capital this deliberately, it signals genuine long-term conviction. For traders with a 12-18 month horizon, the risk-reward in Broadcom at current levels (mid-$180s to low-$190s in September 2026) tilts meaningfully to the upside. The $115B AI revenue forecast represents 32% of total projected revenue by FY2027, which is exactly the kind of structural shift that justifies premium valuations. Watch for Broadcom's next quarterly earnings in late October 2026—that'll be your opportunity to see whether custom chip adoption is tracking to plan or facing headwinds. Until then, position sizing and patience matter more than speed.
Track Live Signals: Access real-time momentum scanner, unusual volume alerts, and AI-powered stock signals at TDVW Scanner. Track $AVGO on Yahoo Finance.
Not financial advice. Always do your own research.