Oct 27, 2026 earnings call — AI infrastructure ramps and capacity investments
Q3 revenue $5.83 billion; company raised 2026 revenue outlook to $21 billion and guided Q4 revenue to $6.275–$6.575 billion.
Revenue $5.7B (est $5.4B) · EPS $3.30 (est $2.93) · Next-quarter revenue outlook $6.4B
Key quotes
"Our design-led HPS programs require customers to pursue early alignment on product development with us and our silicon ecosystem partners up to 24 months in advance of mass production."
"Capital expenditures are expected to be in the range of $1.5 billion-$2 billion, which directly supports the programs we have won to date."
"Earlier this month at the OCP Global Summit, we introduced the DS6005, our first co-packaged optics design built on Broadcom's Tomahawk 6 Davisson chip, integrating high-density silicon photonics engines, liquid cooling, and field-replaceable external laser modules in a 1.5OU form factor, delivering maximum power efficiency for dense clusters."
Analyst Q&A 16
$40 billion outlook vs supply constraints
Q (David Vogt · UBS): Analyst asked how to reconcile the $40 billion outlook with supply chain constraints (silicon, optics, substrates, manufacturing) and how confident management is to navigate those challenges.
A: They built the $40 billion guide factoring in ramp risks, component availability, and execution, and they feel high confidence at 90% year-to-year growth. They expect second-half weighted revenue and see upside if constraints ease, while noting land, power, and geopolitical issues as potential detractors.
"Our $40 billion guide took into consideration the risks around the ramps, the component availability, and the operational execution, David." — Todd Ankenmann (CFO)
R&D hiring and program timing
Q (Kevin Krishnaratne · Scotiabank): Where are you in hiring to grow from 1,600 to 2,000 engineers, where talent comes from, and whether R&D supports programs into 2028 and beyond and other system areas.
A: They are adding roughly 500-600 engineers per year across nine global sites, with Austin, Taiwan, and Silicon Valley ramping strongly and a U.S. focus. They are wrapping many 2027 programs (in test) and are actively engaged in 2028 programs, now about 60-70 active programs and scaling investment.
"In terms of engineering, we continue to build. We've been adding, in general, about 500, 600 engineers a year." — Gavin Cato
Cash needs and free cash flow timing
Q (Joseph Cardoso · JPMorgan): Asked for a range for 2027 net cash use and how cash cycle days compose it, and what needs to happen to return to positive free cash flow and whether 2027 exit could be positive.
A: They reset 2026 FCF given higher demand and view 2027 as a year of investment to support the incremental $20 billion in revenue. They raised $3.5 billion in August, have $4 billion cash plus liquidity, and are confident they can target a return to positive cash flow in 2028.
"When you look at our 2027 view of the $40 billion and looking at the free cash flow there, it is going to be a year of investment to support that incremental $20 billion of revenue." — Todd Ankenmann (CFO)
Margins and product complexity
Q (Karl Ackerman · BNP Paribas): How much rising switch and server complexity (liquid cooling, integrated optics) influences improved margin outlook and margin improvement in server business and full rack offerings.
A: They are selective about products and deals to favor better margins. They expect gross margins to be in the low 11s in 2027, dependent on networking vs compute mix, and see networking as accretive to margins supporting expansion of operating margin to 9% in 2027.
"When you think of it, looking at our gross margins exiting 2026, we feel pretty strongly that our margins will continue in 2027 in the low 11s." — Todd Ankenmann (CFO)
Pluggable vs co-packaged optics at 1.6T
Q (Mehdi Hosseini · Susquehanna International Group): Will 1.6T transceivers be majority pluggable or a mix with CPO/NPO?
A: They expect a lot of pluggables to remain at 1.6T, with some efforts to prove CPO/NPO capabilities; CPO will likely play a larger role at 3.2T and with 400 gig SerDes. They also noted compute and storage growth drivers and specific OpenAI and TPU opportunities ramping in H2 2027.
"I think what we are going to do in 1.6T is certainly see a lot of the pluggable remain." — Gavin Cato
Half-to-half gross margin changes in 2027
Q (John Shao · TD Cowen): With OpenAI ramping in H2 and other networking-heavy programs, will growth and gross margin metrics differ between H1 and H2 2027?
A: They don't expect different gross margin dynamics between the two halves of 2027; quarter-to-quarter mix will matter but overall they expect strong margins in 2027 at 11%.
"No, I don't think there'll be different gross margin dynamics in 2027 between the two halves." — Todd Ankenmann (CFO)
OpenAI proof point winning other deals
Q (Daniel Zhu · Bernstein): How has the Jalapeño work with OpenAI been received and helped win other programs solving similar large-scale design problems?
A: The Jalapeño program serves as a strong proof point and has led to other opportunities; they have a strong funnel but are selective about engagements. Their end-to-end rack design, testing, validation, and manufacturing translation is differentiated and helping in engagements.
"I think that the work that we have done with OpenAI and the Jalapeño program is really serving as a strong proof point of our capabilities." — Steven Dorwart (PRESIDENT)
2027 CapEx and capacity planning
Q (Michael Ng · Goldman Sachs): Given strong 2028 pipeline and current ROIC, how do you balance more capital investment vs ROIC to ensure enough capacity for demand?
A: The 2027 CapEx forecast of $1.5 billion-$2 billion is for awarded and won programs and supports 2027-2028 capacity. They evaluate scenarios continuously, and approved investments show strong ROIs sustaining ROICs above historical levels; they will re-evaluate as new wins arrive.
"When you look at the CapEx forecast for 2027 of $1.5 billion-$2 billion, those approved investments are for awarded programs and already won and support the capacity needs for 2027 and 2028." — Todd Ankenmann (CFO)
Confidence in 2028 demand visibility
Q (Ruben Roy · Stifel): How are you viewing 2028 dynamics relative to follow-on generations and unawarded opportunities; is confidence driven by multi-generation wins at Google/OpenAI or by potential awards?
A: They are very confident in 2028 with lengthening demand visibility; capacity build-out and data center lead times extend visibility into 2029 and 2030, and they are winning programs even before silicon exists. Their current capacity view covers already-won programs, while the funnel provides meaningful upside they are pursuing.
"We're very confident right now in 2028. The demand visibility is getting longer and longer." — Mandeep Chawla (PRESIDENT)
Which scale-up programs will drive 2027
Q (Thanos Moschopoulos · BMO): Will 2027 scale-up revenue mainly come from Helios and Jalapeño or other programs, and what is the 2028 scale-up dynamic?
A: Predominant revenue will come from the named wins like Helios and Jalapeño, and scale-up is an attractive space where they can provide broader solutions. They have additional opportunities in the funnel and noted they won 10 1.6T programs (four scale up, six scale out) that will drive networking growth into 2027.
"I think, as you've described, that's where the predominant revenue will come from, are those programs that are wins for us." — Steven Dorwart (PRESIDENT)
Nature of large-scale pursuit funnel
Q (Taran Katta · Wolfe): Asked about the nature of their significant funnel and the new large-scale engagements they are pursuing.
A: They target AI compute and hyperscale markets but are selective to where they bring differentiated value, growing and diversifying customers and offerings. The funnel demand is at a historic high and largely comes from core customers who are directing more requirements to them; new customer additions are considered carefully.
"I think when we look at the funnel, we have got a strong demand from our current customers, and I would say it is at a historic high level." — Steven Dorwart (PRESIDENT)
Manufacturing investments and margin timing
Q (Steven Fox · Fox Advisors): Asked how manufacturing advantages and investments may hurt margins early but help later, and how they balance regional investments to win business.
A: They are making targeted CapEx (Thailand, Malaysia, Japan, Texas) to provide regional end-to-end solutions, driving productivity and operating leverage. Investments in square footage and power enable wins competitors can't secure, leading to higher-value HPS engagements with better pricing and margins while being disciplined in pursuing only selected opportunities.
"What this is really doing is supporting our customers from a regionalization perspective, and really giving them that global end-to-end solution." — Mandeep Chawla (PRESIDENT)
NCIB after $3.5B equity raise and capacity limits
Q (Ruplu Bhattacharya · BofA): How to reconcile the $3.5 billion equity raise with the NCIB up to 2.5% of float; is buyback to offset stock comp or meaningful return, how much revenue can current footprint support, and flexibility for acquisitions?
A: They renewed the NCIB to purchase up to 2.5% and will be opportunistic, but capital priority is organic growth to support $20 billion growth into 2027 and beyond. On M&A they have an active, capability-driven funnel and have the ability to pursue acquisitions; they won't give a single revenue figure tied to capacity but feel investments cover 2027-2028 needs.
"The priority and focus of our capital is on the organic growth to support the $20 billion growth we see into 2027 and the substantial growth we see beyond that into 2028." — Todd Ankenmann (CFO)
Customers' ability to fund planned spend
Q (Todd Coupland · CIBC): With labs ramping more in 2027-2028, do customers have financing in place to support the spend you built into forecasts?
A: They see no financial distress and believe customers are well-funded with adequate liquidity; they review customer funding and use third-party funding or contractual backstops when needed. They also plan inventory purchases based on customers' near-term demand and credit assessments and currently have no concerns.
"When we're bringing that inventory in, it's assessed on the credit that they have in the moment." — Mandeep Chawla (PRESIDENT)
Keeping factory utilization high
Q (Robert Young · Canaccord Genuity): With high utilization historically, how confident are you to keep utilization high after the large 2027 revenue adds, and how important is utilization to reach 9% operating margin?
A: High utilization is deliberate and tied to business cases; they add capacity only when they expect to fill it quickly. Asia can bring capacity online in as little as 12 months, U.S. takes longer, and they expect high utilization in 2027 and are working to keep it high in 2028.
"We are running at a high level of utilization, specifically in the factories that I mentioned." — Mandeep Chawla (PRESIDENT)
Competition: supply vs design and few competitors left
Q (Paul Treiber · RBC): Is supply availability now more important than design for next-gen wins, and are there fewer competitors able to pursue complex full-rack solutions?
A: Design and supporting leading-edge technologies plus ecosystem partnerships remain critical, and capacity timing is also important; these are bespoke solutions with co-investment. Competition is increasing where they succeed, but fewer competitors exist on 1.6 and 800G; their vertical integration and proof points (compute+network+liquid cooling+services) make it hard for some competitors to stay.
"Certainly the design piece and the ability to support leading-edge technologies, making the investments as we've outlined in R&D and building these strong ecosystem partnerships with key technology providers is really critical to our ongoing success, we feel." — Steven Dorwart (PRESIDENT)