An Update on Two Holdings with Potential to Benefit from AI

September 2026 

Performance data quoted represents past performance; past performance does not guarantee future results. The investment return and principal value of an investment will fluctuate so that an investor’s shares, when redeemed, may be worth more or less than their original cost. Current performance of the fund may be lower or higher than the performance quoted. Performance data current to the most recent month end, and standardized performance can be obtained by viewing the fact sheet or by clicking here. Neither forward earnings nor earnings growth is a measure of a fund’s future performance.

This was a difficult quarter for the Fund. Fortunately, in nearly all cases, poor performance was driven by sector rotation and basket trades rather than adverse fundamental company developments. We profile below our two largest detractors – AST Space Mobile and Applied Materials – which accounted for about 67% of the decline in the quarter (10.7 percentage points). We also exited two positions and reduced the size of a third during the quarter; we explain our rationale for these decisions at the end of the letter.

Update: AST SpaceMobile

AST SpaceMobile (ASTS) was our largest detractor   this quarter, declining 34% from $89 to $59. The third quarter was challenging for most high beta, long duration companies, including early-stage space stocks. In the quarter, the Procure Space ETF dropped 16.5%, Rocket Lab 31%, and Planet Labs 51%. These moves were driven more by thematic trades in broker-constructed baskets than changes in company specific outlooks.

Additional selling pressure on AST shares stemmed from its $1.15 billion convertible bond issuance in July. While this built cash to $3.7 billion – de-risking the constellation’s funding requirements – it introduced technical selling pressure from convertible arbitrage hedging. The market was also spooked by minor operational delays, including a 30-day unfurling delay for BlueBird 11, and a 45-day shipping delay for BlueBirds 14-16, both of which have since been favorably resolved.

In the short term, AST’s price remains tethered to thematic baskets and technical trading mechanics.

Over the long term, execution and fundamentals will decide performance. On that front, the business is advancing nicely. Some notable developments since Q2 include:

•    Constellation Deployment. Following a Falcon 9 launch in early August, AST successfully deployed the phased arrays on BlueBirds 11, 12, and 13 (with BB 11 unfurling after an initial delay). All three are operating as planned, bringing the constellation to six Block 1 and seven Block 2 full-sized satellites in orbit (accounting for the loss of B2BB 7 earlier this year due to a Blue Origin orbital insertion failure).

•    Launch Pipeline. BlueBirds 14, 15, and 16 have shipped from Midland, Texas, to Cape Canaveral ahead of an anticipated Falcon 9 launch in late October. We expect a steep ramp in manufacturing and deployments through 2026 and 2027. Once the constellation reaches 20-25 satellites, AST will initiate non-continuous beta testing (expected late 2026). A critical mass of 45-60 satellites (projected for the second half of 2027) will enable continuous service in key initial markets (U.S., Canada, parts of Europe, the Middle East, and Asia), acting as the primary catalyst for commercial revenue.

•    Manufacturing Scale. Integration is currently underway on over 20 spacecraft structures, with line-of-sight progress through BlueBird 50. The manufacturing footprint has reached approximately 500,000 sq. ft., with plans to double capacity to 1 million sq. ft. To support this, headcount has expanded significantly, growing from 578 employees and consultants at year-end 2024 to 2,250+ employees as of disclosure on October 6.

•    Government & Defense Pipeline. AST recently disclosed three new government contracts representing over $100 million in funded near-term value for 2026-2027. This augments a rapidly expanding defense portfolio that includes the Space Development Agency’s $30 million HALO Europa Track 2 tactical satcom prototype and a prime IDIQ position on the Missile Defense Agency’s SHIELD program. Federal filings indicate a deepening strategic reliance by the U.S. and its allies on AST’s infrastructure for Space-Based Radar (Golden Dome), Electronic Warfare, Alternative PNT, Blue Force Tracking, Drone Control, and Signals Intelligence. During the Q2 earnings call, management outlined the scale of this demand, projecting that government revenue could account for up to half of the company’s targeted $1 billion total revenue in 2027.

•    Global MNO Integration & Forward Demand. Ahead of commercial launch, AST SpaceMobile’s coalition of Mobile Network Operators (MNOs) has grown to over 60 partners representing over 3 billion subscribers. Signaling strong forward demand and operational readiness, these MNOs are making tangible infrastructure commitments, with more than 60 global gateway sites in various stages of development and planning as of October 2026. This ground infrastructure buildout is accompanied by major technical milestones, including routing the first space-based VoLTE calls through AT&T and Verizon’s core networks in July 2026. Recent network integration tests in Canada enabled the country’s first direct-to-device broadband connections and video streaming with Bell, alongside a critical in-orbit satellite-to-smartphone test with TELUS.

•    European MNO Consortium & Spectrum Initiative. In October, Deutsche Telekom, Orange, Telefónica, and Vodafone signed a Letter of Intent to explore a joint bid for Mobile Satellite Services spectrum. While the consortium’s satellite partner remains officially unselected, all signs point to AST. On August 6, AST confirmed integration testing was underway with these exact operators. Following the October announcement, AST issued a supportive press release highlighting its readiness to support the consortium via its Satellite Connect Europe joint venture. This momentum is further bolstered by Vodafone’s existing equity stake in AST and widespread, confirming LinkedIn engagements from employees across AST, the joint venture, and the MNOs.

•    Japan. A joint venture between AST and Rakuten was selected for Japan’s J-LEO sovereign infrastructure program. This is expected to unlock up to $1 billion in non-dilutive, non-debt government capital to subsidize Japan-flagged satellites. In October, the AST-Rakuten partnership was explicitly highlighted in a joint statement by the U.S. and Japanese governments.

As we discussed in our Q2 letter, Starlink (SpaceX’s satellite communications subsidiary) entered the direct-to-cell market a few years ago with a wholesale business model in partnership with legacy carriers. Over the last year, however, it has begun to reposition its business to bypass incumbent carriers and sell direct to consumers.

On October 8 it was announced that SpaceX had agreed to purchase (for a reported $8 billion) 14 MHz of paired nationwide low-band 800 MHz spectrum from Grain Management. With this action, Starlink takes another big step toward becoming a nationwide U.S. wireless carrier. The arrival of a new, aggressive, and well-funded carrier is a clear negative for AT&T, Verizon, and T-Mobile. However, as we discuss below, we believe this is a strategic windfall for AST. The market does not yet see it our way; AST’s stock declined from $57 to $42 in the days following the SpaceX-Grain transaction announcement.

Starlink plans to come to market with an integrated terrestrial and space-based wireless offering. The incumbent carriers today only have terrestrial networks, and they will need space-based capability in order to match one of Starlink’s key selling propositions. AST’s carrier-friendly wholesale business model and multi-year head start on the competition, make it a “must have” partner for the incumbent U.S. carriers (and put the company in pole position with international carriers similarly concerned about the Starlink threat).

The direct-to-device joint venture that AT&T, Verizon, and T-Mobile announced in May and launched on October 1 is evidence of satellite’s now critical strategic importance. AST is already partnered with AT&T and Verizon. This joint venture now opens the door for an AST / T-Mobile partnership, which would give AST near 100% U.S. market share with incumbent carriers. T-Mobile recently scrubbed Starlink’s name from its website; we think an arrangement with AST is in the works.

Further, we think Starlink’s emergence makes it much more likely that basic AST service will be bundled with most wireless plans resulting in 50-100% subscriber penetration (with data caps and upsells for higher capacity). This compares to our original assumption that AST would achieve about 20% subscriber penetration with its U.S. carrier partner(s). While AST’s ARPU will need to be lower to enable this higher volume / higher penetration model, we expect it to yield a materially higher total revenue opportunity.

AST is very well funded and making steady progress toward commercialization. Its strategic value has materially increased. Our outlook for a $300 to $600 per share future value is unchanged. Yet, at today’s price of $41 ($17 billion EV) investors do not need to pay up for this potential. AST has control of roughly $17 billion of wireless spectrum secured through the Ligado spectrum lease in 2025. This spectrum is helpful to AST’s business plan, but not essential. By virtue of its partnership model, AST can rely on spectrum sharing with carriers. There is a robust market for this type of spectrum today. While we think it highly unlikely the company will part with spectrum rights, the math highlights the very attractive value present at today’s share price.

Update: Applied Materials

Applied Materials (AMAT) was our second-largest detractor in the quarter, down roughly 29% (from $723 to $511) after more than doubling in Q2. The pullback was triggered by worries that AI infrastructure spending is outpacing actual AI demand. These concerns ranged from fears over memory pricing and circular financing in July to calls from Anthropic’s Dario Amodei and OpenAI’s Sam Altman to “pace the frontier” for safety reasons in September. Investors have also questioned whether cheaper open-weight models, and the neoclouds hosting them, might strip revenue from the frontier labs that serve as the industry’s largest buyers of compute.

While these are important issues for many AI businesses, the economy, and broader society, AMAT and other semi-cap equipment suppliers are not directly impacted. Whether the end product is from Nvidia or AMD, or custom silicon like tensor processing unit (TPU) and Trainium, all require the same core deposition, etch, and replace with chemical mechanical polishing (CMP) manufacturing steps. The rise of open-weight models simply shifts who is buying the chips, not whether they are bought. Cheaper tokens will ultimately drive better end user economics, broader AI adoption, and expand overall compute consumption, as evidenced by Google’s monthly token volume rising roughly sevenfold in a single year.

Meanwhile, AMAT’s underlying business has never been stronger. Fiscal Q3 delivered $9.1 billion in revenue (+25%), $3.50 EPS (+41%), and a 50.4% gross margin, marking the 13th straight quarter of year-over-year margin expansion. Management raised its outlook, giving Q4 guidance of $10.2 billion in revenue and $4.02 in EPS, beating consensus by $630 million (7%) and $0.30 (8%), respectively. Calendar 2026 Semiconductor Systems growth is now expected to exceed 30%. And wafer fab equipment (WFE) demand is accelerating. Since our Q2 letter, where we cited a $235 billion estimate, some Street models now project 2028 WFE at ~$275 billion, and buy-side estimates have reached $300 billion. Top customers are sharing rolling eight-quarter tool forecasts, and AMAT is securing cleanroom capacity that would enable it to double systems output by 2028.

This demand is driven by actual usage rather than subsidized capacity. Recent year-over-year quarterly revenue growth from hyperscalers reflects this: Google Cloud Platform +82%, Amazon Web Services +37%, and Microsoft’s Azure +43%, and all three remain capacity-constrained. Further illustrating this surging demand, AMAT’s own internal token consumption rose roughly 45-fold over the past year.

We believe the next major catalyst is agentic AI. Gartner predicts autonomous agents will be embedded in 40% of enterprise applications by late 2026, up from less than 5% in 2025. Agentic consumer agents such as Meta’s Muse are surging in use as well. Because agentic workflows require 5 to 30 times more tokens than traditional chatbots, enterprise AI spending continues to climb despite a 90% drop in per token pricing since 2023.

Increasing demand for tokens ultimately translates to increasing demand for silicon wafers. AMAT models that an agentic AI pod built on Nvidia’s next-generation Rubin platform (2027) will consume roughly 1,400 to 2,200 wafers, compared to just 600 for today’s Blackwell-based pods. With more than 85% of this silicon concentrated in leading-edge logic, replace with dynamic random-access memory (DRAM), and advanced packaging, AMAT projects these three segments will drive approximately 75% of all WFE growth through 2030.

AMAT currently trades at roughly 13x our base-case calendar 2028 earnings estimate of low-to-mid $30s per share. Today’s price seems to assume the AI buildout is a standard cyclical peak, rather than a secular transformation. The market is discounting a downturn that customers’ own forecasts simply do not show. We trimmed our position modestly early in the third quarter at much higher prices, but find today’s investment proposition compelling.

Exited / Reduced Positions: CDW, RELX plc & Shenandoah Telecom

During the quarter, we sold two positions and reduce the size of a third.

We exited CDW, which was about a 3% position at the beginning of the quarter. The company has begun pivoting to higher value services, which makes some strategic sense, but is a departure from its core competency as a world-class sales organization. We are reluctant to go along on this journey, and have also come to view CDW as a good but not great business. So, we exited the position and will await better opportunity elsewhere.

We also exited RELX plc., which was just a 0.4% position. We have concerns about several of its’ core businesses facing potential erosion from the emergence of AI based capabilities and tools. It was always a lower conviction starter position, and never exceeded 1% of assets.

Finally, we reduced our position in Shenandoah Telecom. Shenandoah faces ongoing competition from fiber overbuilders and emerging competition from satellite-based broadband connectivity providers (Starlink and Blue Origin). The competitive environment has been more difficult than we expected, and the future looks like it will be even more challenging with exponential growth in launch capacity over the next several years.

Conclusion

We thank you for entrusting your capital to us. We will continue to do our best to protect and grow your investment over time.

Sincerely,
Broad Run Investment Management, LLC

Click here for a full listing of Holdings.

Click here for full, standardized Fund performance.