Rolf Bulk

Head of Semiconductors & Infrastructure @FuturumEquities | No investment advice.

[Rolf Bulk @rolfbulk]

Memory, especially DRAM, remains in structural shortage.

The industry needs to add roughly 300k DRAM WSPM between now and 2030. Getting there requires perfect execution across new cleanrooms, tool procurement/installation, and node migrations at SK hynix, Samsung, and Micron.

Very challenging. Difficult to see shortages abate any time soon.

[Daniel Newman @danielnewmanUV]

$MU and the broader AI infrastructure is taking a breather today, but $10 Trillion+ in Capex between now and end of 2030 makes the prospects for Micron to continue to rise very high. 💪🏻🚀

[Rolf Bulk @rolfbulk]

@StockSavvyShay Congrats!!

[Rolf Bulk @rolfbulk]

I think Meta will continue to find profitable internal use for its compute.

There are so many behind-the-scenes inference use-cases: ad targeting, content ranking, search, recommendations, etc.

Understandable Meta wants to keep its options open here, but I can't see it happen for now.

[Rolf Bulk @rolfbulk]

Very unlikely we see anything resembling supply/demand balance in DRAM before 2029, potentially later.

More cautious on NAND: adding bit supply is easier, and Lam’s commentary on the upgrade cycle suggests manufacturers are moving rapidly to migrate to higher layer-count devices.

For DRAM, there is no magic formula; you need to add wafer capacity, and current shell space + shells coming online are simply not enough to meet increasing AI-driven demand.

[Rolf Bulk @rolfbulk]

Curious to see how rapidly China can ramp domestic hybrid bonding and die-attach tools.

Besi’s dominance extends beyond hybrid bonding. It has ~75% share across advanced assembly tools requiring accuracy of 7 microns or below, including TCB and flip-chip, and has held that position for years.

Besi’s 60–65% GM is testament to its capabilities: accuracy plus throughput (you need both to be competitive), versus Western competitors like KLIC and ASMPT, Korean players, and likely domestic Chinese vendors.

I would not discount Besi here.

[Daniel Newman @danielnewmanUV]

Wall Street already runs on ETR's signal. Today, Futurum signed a definitive agreement to acquire the dataset institutional investors trust to read enterprise tech spend.

Here is why it matters.

→ 9,000+ enterprise IT decision-makers feeding real spending intent, not vendor-sourced guesses

→ $2 Trillion in tracked technology spend that hedge funds and institutions use to call the print before the print

→ The earliest read on who is winning and losing budget across every major name in tech

→ Now sitting on the Futurum Intelligence Platform, paired with our analysts, our coverage, and our reach

The legacy research firms tell you what happened last quarter. This data tells you what is happening right now.

Research built on evidence. This is what AI-native looks like when it has the data to back it up.

Futurum and ETR. Subject to customary closing conditions.

[Rolf Bulk @rolfbulk]

@paurooteri Sensible to me. Hynix doesn’t want to be beholden to any single large customer, which is the strongest position to be in. DRAM capacity will remain tight for years, so there’s little need to hedge its position.

[Rolf Bulk @rolfbulk]

AMD’s Venice CPU ramp later this year, on TSMC’s N2 node, is underappreciated.

It should clearly lead on specs, with 256 cores and multithreading. But the more important point is capacity: N2 is not as tight as N3, which is the main node for most AI GPU/ASIC ramps. AMD should be able to secure the N2 capacity it needs.

Undoubtedly more on this front next week at Computex in Taiwan. We’ll be there.

[Daniel Newman @danielnewmanUV]

$AMD crosses $500 today. The @FuturumEquities head of semis and infrastructure @rolfbulk put a 12 month target of $750 on it just last week. 👀

[Rolf Bulk @rolfbulk]

Taiwan and Korea are major beneficiaries of the AI infrastructure cycle. Happy to contribute to CNBC's explainer on shifting global market dynamics.

[The Futurum Group @TheFuturumGroup]

AI infrastructure demand is redefining leadership across global equity markets.

In a discussion on how AI is influencing global stock market rankings, @TheFuturumGroup’s @rolfbulk explained that Taiwan and South Korea are benefiting from exceptionally strong industry fundamentals, with AI demand driving rapid profit growth across key parts of the semiconductor supply chain.

Rolf pointed to a demand and supply imbalance at the center of the story. The shift toward agentic AI is creating a major increase in compute requirements, while supply for the chips, memory, and infrastructure supporting that demand remains constrained. That dynamic is giving manufacturers significant pricing power and helping lift markets with deep exposure to AI hardware.

He also noted that South Korea’s strength extends beyond memory leaders, with profit growth supported by sectors such as defense, shipbuilding, power equipment, K-culture, and broader corporate reform.

▶️ Watch the segment for Rolf’s analysis of how AI infrastructure demand is influencing global market rankings.

[Rolf Bulk @rolfbulk]

Great to discuss Nvidia's print at Bloomberg following our formal coverage initiation on global Semiconductors and AI infrastructure for @FuturumEquities yesterday. A broadening demand base, $20B in CPU revenues this year, 75% gross margins - strong quarter on all fronts as Jensen + team continue to execute.

Counter to what AH market moves suggests, I think the read-across for AMD here is very positive. A $200B CPU TAM is well above expectations; Arm guided $100B a few months ago, AMD $120B. AMD/Intel will capture their fair share of this market, despite Nvidia/Arm emerging as a real challenger.

[Rolf Bulk @rolfbulk]

SoftBank Group ($SFTB.NE), which owns ~90% of Arm and 13% of OpenAI, has had a strong rally on the back of Nvidia’s $200bn CPU TAM target, Arm’s share-price reaction to it, and newsflow around a potential OpenAI listing.

Happy to discuss with CNBC: https://t.co/gNbrX42uKB

[Rolf Bulk @rolfbulk]

Great to discuss Nvidia's print at Bloomberg following our formal coverage initiation on global Semiconductors and AI infrastructure for @FuturumEquities yesterday. A broadening demand base, $20B in CPU revenues this year, 75% gross margins - strong quarter on all fronts as Jensen + team continue to execute.

Counter to what AH market moves suggests, I think the read-across for AMD here is very positive. A $200B CPU TAM is well above expectations; Arm guided $100B a few months ago, AMD $120B. AMD/Intel will capture their fair share of this market, despite Nvidia/Arm emerging as a real challenger.

[tae kim @firstadopter]

OMG. It's almost as if single digit P/E with triple digit growth rate = undervalued or something.

[Rolf Bulk @rolfbulk]

Sensible move by OpenAI; it meets a real need (100% reliable access for agentic workflows), and it's difficult for Anthropic to replicate given their constraints on compute.

[OpenAI @OpenAI]

Introducing OpenAI Guaranteed Capacity: a new offering that enables customers to guarantee long-term access to OpenAI compute.

We’ve made long-term investments in infrastructure, partnerships, and capacity planning to help customers scale reliably.

Now, Guaranteed Capacity helps customers plan ahead for critical workloads in a compute-constrained world.

https://t.co/TN4OkZr2Uo

[Rolf Bulk @rolfbulk]

@chrisbarber Great interview, good insights

[Rolf Bulk @rolfbulk]

SoftBank Group ($SFTB.NE) is a great name to track. It owns 13% of OpenAI, but more importantly for now, still holds 90% of Arm, one of the main beneficiaries of the data center CPU bottleneck.

Masa and Rene Haas’s call to go all-in on CPUs, rather than focus on merchant GPU/ASICs, now looks very smart.

Happy to discuss SBG’s recent rally with CNBC: https://t.co/JCvN6IaH8C

[Rolf Bulk @rolfbulk]

Great print across the board. AMD is uniquely exposed to both the AI data center GPU opportunity, with MI450 ramping in 2H, and the CPU cycle driven by agentic AI, where it continues to take share as Intel works to regain its footing. Further upside ahead, even after the AH move.

[Futurum Equities @FuturumEquities]

$AMD Q1 results showed no surprises, with the company tracking toward full realization of its data center AI opportunity.

Server CPU TAM expands to $120B+ by 2030 as agentic AI workloads demand new compute patterns, to address the ongoing CPU bottleneck

[Rolf Bulk @rolfbulk]

$KLAC earnings:

2026: >$140bn WFE spend. No real surprise after the $LRCX print earlier this quarter.
2027: growth accelerates again, implying WFE of ~$165–175bn.
2030: $195–235bn WFE outlook maintained from the March CMD.

The key debate on semicap is whether high-teens to low-20% growth of 2026 and 2027 can extend into 2028 and 2029, or whether we're in for a slowdown, as the pace of capacity additions related to the AI build-out moderates.

KLA's commentary on 2027 and guide on 2030 imply only ~5–10% WFE spending CAGR from 2027 to 2030.

[Rolf Bulk @rolfbulk]

The upward 2026 capex revisions we are seeing today make sense: higher component pricing is forcing hyperscalers to revise budgets. Microsoft is putting some numbers to it, disclosing a $25bn impact from higher component costs within its $190bn total capex budget.

Importantly, the higher capex is being supported by accelerating AI-related revenue growth: Azure +40%, Google Cloud +63%, AWS +28%, Meta +33%.

With revenues accelerting, gard to see spending slowdown anytime soon. I would not be surprised to see top-4 capex trend toward $825–850bn next year; consensus still needs to move up by 10–15%.

[Rolf Bulk @rolfbulk]

Big week for tech, with all the hyperscalers set to report. Hard to see how 2026 capex guidance cannot be revised up further, given what we’ve seen in memory and CPU pricing over the past few months.

[Daniel Newman @danielnewmanUV]

The bull thesis just got validated. In a single afternoon.

Meta. Microsoft. Amazon. Alphabet. All four reported. All four delivered.

The numbers tell the story:

$MSFT Azure +40% — beat the high end of guidance. AI business now a $37B run rate, +123% YoY. Copilot crossed 20 million paid seats.

$GOOGL Cloud +63% to $20B. Backlog of $460 billion. Pichai called enterprise AI "the primary growth driver of cloud for the first time."

$AMZN AWS +28% to $37.6B — the fastest growth in 15 quarters. Amazon reaffirmed $200B in capex for the year.

$META +33% revenue growth — the fastest since 2021. And they raised full-year capex guidance to $125–$145B.

The deceleration narrative is dead.

The "AI capex is speculative" narrative is dead.

The "where's the AI revenue" narrative is dead.

This was the prove it quarter. They proved it.

What we saw tonight is durable, compounding cloud demand, accelerating AI monetization, and a capex cycle being underwritten by signed customer commitments, not optimism.

Sorry bubble bears. 🐻
This isn't 1999. Real customers. Real revenue. Real cycle.

The companies investing in AI infrastructure today are buying the most valuable real estate of the next decade and tonight they showed exactly why.

Buckle up. We're just getting started.