Six customers underwrite Broadcom's $230bn AI chip target for 2028
Custom silicon made up 73% of the $16.7bn AI quarter. Hock Tan says that line doubles twice by fiscal 2028. He has already secured wafers and memory through 2028 to serve promises only six buyers have made.
The Investor · Invest desk
What happened
- Broadcom's AI semiconductor revenue reached $16.7bn in the fiscal third quarter, up 221% on the year and 54% on the prior quarter, on total revenue of $29.6bn for the period ended August 2.
- CFO Amie Thuener guided fourth-quarter revenue to about $34.8bn, with AI semiconductors at $21.7bn of it and non-GAAP operating margin held at 66%.
- Custom chips accounted for about 73% of the quarter's AI revenue, with the balance in the AI networking silicon that connects the clusters together.
- Total quarterly revenue slightly missed analyst expectations despite the AI figures, and the shares dipped in the aftermath.
Why it matters
- Who's exposed exposure Broadcom has locked wafer, high bandwidth memory and substrate supply out to 2028 against deployment promises, so a customer that slips a gigawatt programme by two quarters leaves that secured supply sitting on Broadcom's side of the table.
- What it limits constraint When roughly three quarters of the AI line is customer-specific silicon, Broadcom's growth rate is set by a handful of buyers' architecture roadmaps rather than by any product cycle it controls itself.
- What's now possible capability Converting 46% of revenue into free cash flow is what lets Broadcom carry a two-year supply position without raising outside capital, which is the mechanism that makes the pre-commitment strategy available at all.
- What doesn't add up contradiction Some analysts already mark 2028 nearer $200bn than Tan's $230bn, and against RBC's market forecast his number would be roughly 40% of all AI chip sales, so the argument is about achievable share rather than about demand.
Next story · Invest · 10h ago
Broadcom's Hock Tan likens AI investors to gamblers while forecasting $230 billion in AI chip salesConfirmed7 publishers
The ramp Hock Tan laid out on the September 2 call carries an arithmetic requirement that this quarter's guide does not yet meet [6]. His $115 billion of AI semiconductor revenue in fiscal 2027 averages $28.75 billion a quarter, about 32% above the $21.7 billion guided for the period ending November 1 [7][11][13][28], and the $230 billion he put on fiscal 2028 averages $57.5 billion a quarter, nearly twice the $29.6 billion the entire company booked in the three months to August 2 [6][10][29]. Getting from roughly $58 billion this fiscal year to $230 billion is a factor of about four, which is 99% compounded two years running [6][30]. (Cryptopolitan dates the year-ago comparison to 2025 while Crypto Briefing labels the same quarter fiscal Q3 2026, so read the fiscal labels loosely [1][6].)
Composition tells you more than the multiple does. Fourth-quarter guidance takes total revenue up about $5.2 billion sequentially and AI silicon up $5.0 billion of that, so 96 cents of every guided incremental dollar arrives on one line while everything else adds roughly $200 million on a $12.9 billion base [12][10][11][1][25][26][33][34]. Inside that line, custom accelerators were about 73% of the $16.7 billion, call it $12.2 billion, leaving some $4.5 billion of networking [16][1][27]. A year earlier the whole AI line was about $5.2 billion [1][18]. Infrastructure software grew a respectable 29% to $8.75 billion and is now under 30% of revenue, against AI's 56% [3][31][1][32].
The contracts underneath are long, which is the part that reads like an annuity even though the accelerators do not. Google's custom chip supply agreement, signed in April, runs through 2031 on top of a TPU relationship of roughly ten years [14], and the Jalapeño inference chip OpenAI unveiled in June with Broadcom is aimed eventually at about 10 gigawatts [4]. The trade is the ASIC trade: less flexible than a general-purpose GPU, cheaper, tuned to one job, and finding its market in inference rather than training [5].
Or rather, the more interesting version of the bear case treats the demand as borrowed rather than imaginary. Every name on the customer list still buys from Nvidia; Sam Altman still says Nvidia makes the best AI chips in the world [19], and Anthropic runs Claude on AWS Trainium and Google TPUs as well as Nvidia GPUs [20] while committing to 3.5 gigawatts of Google-powered capacity from 2027 that could approach 5 [15]. The stock dipped on a quarter whose total revenue slightly missed [36], so the price already contains some version of the ramp.
This is probably wrong, but the figure I would underwrite here is the customer count rather than the growth rate. Six [8]. With three quarters of AI revenue coming from customer-specific parts [16], Broadcom's fiscal 2028 is a function of architecture decisions made inside four or five buildings it does not own. What would prove it wrong is a seventh and eighth name on that list plus a January AI print near the $28.75 billion fiscal 2027 needs [28], at which point the concentration argument becomes an unearned discount.
What to watch
- Whether the six-customer list gains a seventh name, or quietly loses one, on the next call.
- Purchase commitment and inventory disclosure as the supply secured through 2028 is drawn down.
- Whether Tan repeats $230bn for fiscal 2028 unchanged, or trims it toward the analyst mark.
Information, not investment advice. This story is general information about companies and markets, drawn from cited reporting. It is not a recommendation to buy, sell or hold any security; check the cited sources before acting.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence74
- Adoption72
- Hype gap+22
- Incentives62
- Confidence70
Perspective Coverage
9 publishers- Builder
- Builder 14%
- Operator
- Operator 21%
- Investor
- Investor 65%
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
Broadcom's AI semiconductor revenue was $16.7 billion in the fiscal third quarter, a 221% increase from the same period in 2025 and up 54% from the prior quarter, according to Cryptopolitan.
- [2]
The semiconductor solutions segment grew 127% to $20.8 billion in the quarter.
- [3]
Infrastructure software revenue grew 29% to $8.75 billion in the quarter.
- [4]
In June, OpenAI unveiled Jalapeño, an inference chip designed and built with Broadcom, and the two companies set a target of eventually powering about 10 gigawatts of capacity.
- [5]
Broadcom's play is centered on the ASIC, less flexible than a general-purpose GPU but cheaper and tuned for specific AI jobs, a tradeoff that has found a market in inference.
- [6]
On Broadcom's fiscal Q3 2026 earnings call on September 2, Hock Tan forecast AI semiconductor revenue of roughly $230 billion by fiscal 2028, a fourfold increase from the $58 billion the company expects this fiscal year, according to Crypto Briefing.
- [7]
Tan projected AI semiconductor revenue would roughly double to $115 billion in fiscal 2027 before doubling again to about $230 billion in fiscal 2028.
- [8]
Six key customers are driving Broadcom's AI demand, including Google, OpenAI, Anthropic and Meta, making what Tan described as multi-gigawatt-scale deployment commitments.
- [9]
Broadcom says it has secured supplies of cutting-edge wafers, high bandwidth memory and substrates through 2028.
- [10]
Broadcom's revenue for the quarter ended August 2 was $29.6 billion, up 86% from a year earlier.
- [11]
Tan said he expects AI semiconductor revenue to climb to $21.7 billion in the fourth quarter, a 236% rise from the same period a year ago.
- [12]
CFO Amie Thuener guided fourth-quarter revenue to about $34.8 billion, up 93% from a year ago, with non-GAAP operating margin expected to remain at 66%.
- [13]
Broadcom's fiscal fourth quarter ends November 1, and the company cautioned that actual results may vary materially from the outlook.
- [14]
In April, Broadcom signed a long-term agreement with Alphabet's Google to supply custom AI chips through 2031, prolonging a TPU partnership of about ten years.
- [15]
Anthropic committed to 3.5 gigawatts of Google-powered computing capacity starting in 2027, a commitment that could approach 5 gigawatts over time.
- [16]
Custom chips accounted for about 73% of Broadcom's third-quarter AI revenue.
- [17]
Broadcom's AI semiconductor division includes both custom accelerators and AI networking chips, with the networking fabric connecting chips as AI clusters grow.
- [18]
Broadcom's AI semiconductor revenue was about $5.2 billion in the year-ago quarter.
- [19]
None of Broadcom's customers is ditching Nvidia, and OpenAI's Sam Altman has publicly said Nvidia makes the best AI chips in the world.
- [20]
Anthropic runs Claude on AWS Trainium, Google TPUs and Nvidia GPUs to distribute its risk.
- [21]
Tan estimated adjusted earnings per share for fiscal 2028 would surpass $30.
- [22]
Some analyst estimates for Broadcom's 2028 AI revenue land closer to $200 billion than to Tan's $230 billion figure.
- [23]
RBC has projected the total AI semiconductor market could exceed $550 billion by 2028.
- [24]
If RBC's market figure holds, Broadcom's $230 billion target would represent roughly 40% of the entire AI chip market.
- [25]
The fourth-quarter AI guide implies a sequential increase of $5.0 billion in AI semiconductor revenue.
- [26]
The fourth-quarter total revenue guide implies a sequential increase of about $5.2 billion.
- [27]
Custom accelerators were about $12.2 billion of the quarter's AI revenue, leaving roughly $4.5 billion of networking and other AI silicon.
- [28]
Fiscal 2027 AI revenue of $115 billion averages $28.75 billion a quarter, about 32% above the $21.7 billion guided for the fourth quarter.
- [29]
Fiscal 2028 AI revenue of $230 billion averages $57.5 billion a quarter, about 1.94 times the $29.6 billion of total company revenue in the quarter ended August 2.
- [30]
Going from about $58 billion of AI revenue this fiscal year to $230 billion in fiscal 2028 is a factor of 3.97, equivalent to about 99% growth compounded for two years.
- [31]
Infrastructure software was about 29.6% of Broadcom's quarterly revenue.
- [32]
AI semiconductor revenue was about 56% of Broadcom's total quarterly revenue.
- [33]
AI silicon accounts for about 96% of Broadcom's guided sequential revenue growth.
- [34]
Non-AI revenue was $12.9 billion in the reported quarter and is implied at $13.1 billion in the guide, an increase of about $0.2 billion.
- [35]
Broadcom generated $13.7 billion in free cash flow, about 46% of revenue, and non-GAAP operating margin was 66%.
- [36]
Broadcom's overall third-quarter revenue slightly missed analyst expectations and the stock dipped in the aftermath.
Sources
9 independent publishers whose own reporting we read for this story.
- cnbc.comBroadcom's stock drops 5% as weak guidance overshadows earnings beat
1 article · September 2, 2026
- cryptobriefing.comBroadcom forecasts AI semiconductor revenue growth to $230B by 2028
2 articles · September 3, 2026
- cryptopolitan.comHock Tan guides Broadcom's AI revenue to $21.7 billion after a $16.7 billion quarter
1 article · September 2, 2026
- en.sedaily.comBroadcom Revenue Jumps 86% but Shares Slide on Q4 Outlook
2 articles · September 3, 2026
- finance.yahoo.comBroadcom stock sinks as chipmaker results 'not enough to keep investors happy'
2 articles · September 2, 2026
- fool.comWhy Broadcom Stock Dropped Today | The Motley Fool
1 article · September 3, 2026
- prnewswire.comBroadcom Inc. Announces Third Quarter Fiscal Year 2026 Financial Results and Quarterly Dividend
1 article · September 2, 2026
- proactiveinvestors.comBroadcom stock drops on Q4 guidance miss even as AI semiconductor revenue triples | NASDAQ:AVGO, XETRA:1YD
1 article · September 2, 2026
- techtimes.comBroadcom Custom AI Chip Revenue Surges 221% to $16.7B, Q4 Guidance Disappoints
1 article · September 3, 2026
Topics and entities
Topics
- Custom AI Accelerators
- AI Semiconductors
- AI inference