The Power Spread
In this edition of the Smart Investor newsletter, we spotlight a company laying the fiber rails of intelligence. We are not removing any holdings today, as the recent weakness has been concentrated in select tech and industrial infrastructure names, and we prefer to let sentiment settle before making portfolio changes. But first, let’s review the latest Smart Portfolio developments.
1
Portfolio News and Updates
❖❖ CrowdStrike (CRWD) performed its first-ever stock split after the close on July 1 – a 4-for-1 move announced alongside its earnings report in June. The stock edged up on its first day of split-adjusted trading despite a broad sell-off in tech shares. The split lowers the nominal per-share cost and is expected to broaden retail accessibility amid high demand for the shares of the leading cybersecurity specialist enjoying accelerating revenue growth and strengthening profitability.
1
❖❖ MasTec (MTZ) announced that it has agreed to acquire The Superior Group, a premier full-service electrical contractor focused on critical infrastructure, in a $1.65 billion cash-and-stock deal. The acquisition is expected to close in mid-to-late July 2026, after which Superior’s business will operate as a new MTZ division. MasTec expects the deal to be immediately accretive, adding $800-900 million in revenue and $100-115 million in adjusted EBITDA over the remainder of 2026. For full-year 2027, Superior is expected to generate revenue of $2.2-2.5 billion and adjusted EBITDA of $250-275 million. The acquisition reflects MTZ’s accelerating push into data center and mission-critical infrastructure buildout, adding further momentum to an already strong growth narrative.
1
❖❖ Palantir (PLTR) surged over the past week, bucking the recent sell-off in AI-related stocks, after DA Davidson upgraded the shares to Buy from Hold and raised its price target to $175. This implies over 30% upside from current levels. Analysts noted that Palantir “has grown into its valuation,” describing the current multiple as the most attractive it has been in a long time and addressing a key bearish concern.
The bull case rests on Palantir’s increasing indispensability to enterprises. Its positioning as a model-agnostic AI orchestration layer – often described as an “AI operating system” – provides significant competitive advantages in the AI era. This was underscored by Anthropic’s recent dispute with the U.S. government, leading to a temporary removal of its most advanced models. This clearly demonstrated the risks – including systemic disruption – associated with relying directly on a single frontier model provider rather than a neutral platform.
1
❖❖ U.S. technology leaders are working to replicate some of Palantir’s business methods, testifying to the strength of its model. The developments also highlight the fact that the barrier to increasing productivity with AI isn’t about token use or model strength, but the ability to integrate AI into the daily workflow.
Amazon’s (AMZN) AWS announced a $1 billion initiative to create a unit of forward deployed engineers (FDEs) – i.e., experts that work with customers on-site to co-develop, deploy, troubleshoot, and scale bespoke AI solutions. FDEs are expected to alleviate one of the key bottlenecks in enterprise AI adoption: building secure and governable agentic workflows on top of the customers’ data and processes. The massive scale of the initiative reflects how high the stakes in the AI race have risen, with AMZN, like other hyperscalers, striving to prove its massive capex can bring as massive an ROI.
In recent months, both leading AI labs announced new joint ventures that aim to broaden their reach by embedding engineers with customers to offer scalable and customized AI solutions. Anthropic has revealed a $1.5 billion JV with Blackstone, Goldman Sachs, and other Wall Street firms that will send teams of engineers to clients – including the investors’ portfolio companies – to incorporate AI across their operations. Soon after that, OpenAI went big with a $10 billion JV backed by 20 PE and consulting firms, pursuing the same model. On top of bringing customer onboarding costs to zero and speeding up the procurement time, this model also helps house the heavy upfront costs of AI implementation and customization directly on the joint venture’s books, optimizing the AI labs’ margins ahead of their anticipated IPOs.
Microsoft (MSFT) is the most recent firm to enter the field. The tech giant announced this week a launch of a new subsidiary – named Microsoft Frontier Company – at a cost of $2.5 billion. The new 6,000-employee subsidiary will dispatch engineering experts into client organizations to help them choose and deploy the most suitable AI tech – from MSFT itself or other providers – integrating it with the customers’ internal data.
1
❖ ❖ DA Davidson views Microsoft’s (MSFT) Copilot as another kind of AI orchestration layer, serving different corporate needs. While Copilot serves as a productivity-focused front-end – like an “executive assistant” handling everyday work across the company – Palantir functions as a deeper “senior project/operations manager,” excelling at complex, high-impact initiatives with sophisticated data integration.
Corporations are increasingly wary of depending solely on individual frontier model providers. A model-agnostic approach offers better cost control, performance optimization, resilience, and broader adoption. Although Copilot began as an AI assistant, it now routes queries to the most suitable models (OpenAI, Anthropic, and potentially others). Microsoft’s global dominance in enterprise software gives it a powerful distribution advantage, with millions of companies already inside its ecosystem. DA Davidson has reconfirmed its Buy rating on MSFT with a $550 price target, implying an upside of about 40% from current levels.
1
❖ ❖ Amazon’s (AMZN) latest batch of 29 satellites lifted off last week on a United Launch Alliance (a Lockheed Marting and Boeing space launch JV) Atlas V rocket, bringing the total number of deployed satellites to 396. The constellation, named Amazon Leo and managed by AMZN’s subsidiary Kuiper, targets the eventual deployment of over 3,600 satellites. While in its the current state, Leo remains far from that target, management believes that with the latest launch, the company has enough satellites to provide initial broadband service this year, with future missions to add coverage. Leo is planned as a Starlink competitor, although Elon Musk’s low-orbit internet service, operating thousands of satellites and providing services worldwide, remains well ahead. Interestingly, besides ULA’s Atlas V and Jeff Bezos’s Blue Origin, Leo satellites are also launched to space using SpaceX rockets, highlighting an odd setup where Amazon is building a rival to Starlink while also relying on the rocket that helped SpaceX deploy its own network.
❖ In other news, Panos Panay, the head of devices and services at Amazon, confirmed that the company is developing custom chips for its key consumer devices to create a seamless hardware-and-software connection. The move, expected to begin in 2027, also aims to lower costs in Amazon’s non-AI businesses as it ramps up AI spending. AMZN is already a major player in custom silicon, with its cloud chips Trainium and Inferentia used in AWS data centers and their compute capacity rented to enterprise customers. Now, the retail and cloud leader is ramping up its efforts to build a custom-chip empire, doubling down on consumer AZ Series chips used in physical gadgets such as Echo, Fire TV, or Kindle, processing data at the edge.
❖ According to media reports, AMZN is tapping the U.S. investment-grade bond market to fund its AI buildout. Amazon is looking to raise at least $25 billion through an eight-part offering of floating- and fixed-rate notes with maturities ranging from three to 40 years. Fitch Ratings assigned the proposed notes a rating of AA–. The bond sale marks Amazon’s fourth major debt-market foray in less than a year after it resumed issuance last November following a three-year pause. In March 2026, AMZN completed a record-setting $37 billion U.S. dollar sale that formed part of a broader $53.8 billion multi-currency deal – the largest corporate bond offering in capital markets history.
Hyperscalers – Meta Platforms, Alphabet, Amazon, and Oracle, with Microsoft a frugal outlier – collectively issued around $121 billion in U.S. corporate bonds in 2025 alone, more than 4x their prior five-year annual average. The borrowing spree continued into 2026, as the tech giants are collectively projected to spend roughly $725 billion on capital expenditures this year, up more than 60% from 2025 levels, with the vast majority directed toward AI infrastructure.
1
❖❖ Broadcom (AVGO) rallied on news that it had expanded its partnership with Apple through 2031 to develop and supply custom silicon across multiple generations of Apple devices. This is a significant revenue visibility boost for the chipmaker, as Apple remains one of its largest customers. AVGO’s chipmaking deals have been on fire in recent months: since the start of April, the company expanded its partnership with Meta, announced multi-year deals with Google and Anthropic, and helped OpenAI build Jalapeño, its first-ever custom AI chip. These ASICs partnerships massively expand Broadcom’s role as a major supplier for custom AI chips to tech firms seeking alternatives to general-purpose GPUs.
1
Portfolio Earnings and Dividend Calendar
❖ The Q2 2026 earnings season will officially begin next week with earnings reports from the largest U.S. financial institutions. JPMorgan Chase (JPM) and Citigroup (C) will reveal their results on July 14, while Bank of New York Mellon (BNY), Morgan Stanley (MS), and PNC Financial (PNC) are scheduled to report on July 15.
❖ The ex-dividend date for Oracle (ORCL) is July 10, while for EMCOR Group (EME) it is July 15.
w
1
New Buy: Dycom (DY)
Dycom Industries operates at the physical foundation of America’s digital infrastructure – designing, building, upgrading, and maintaining the fiber, wireless, and utility networks that keep data, communications, and power flowing. The company delivers engineering, construction, and technical services to telecommunications providers and utilities, while increasingly expanding into the electrical and building systems that support modern data centers. A key partner to the nation’s largest telecom carriers, electric and gas utilities, and digital infrastructure operators, Dycom has become critical to projects that demand scale, specialized expertise, and reliable delivery. As fiber networks extend deeper into communities, hyperscale data centers multiply, and connectivity becomes an essential utility for consumers and businesses alike, Dycom plays a vital role in the infrastructure ecosystem – transforming long-term investment in digital capacity into the physical networks that enable the modern economy.
1
Intelligent Groundwork
Dycom’s story began in 1969, long before fiber broadband, 5G, cloud computing, and AI data centers became national infrastructure priorities. The company started as a specialty contractor and spent decades expanding through organic growth and acquisitions, gradually building a network of operating companies able to handle the physical work behind communications infrastructure – engineering, construction, maintenance, restoration, and field services across local markets.
That long buildout became DY’s real advantage. Telecom networks are local, labor-intensive, highly regulated, and difficult to scale without deep field presence. Over time, Dycom became the partner large carriers could rely on when network upgrades moved from planning documents to trenches, poles, homes, businesses, and rights-of-way. Its role widened as broadband became a competitive necessity, utilities needed locating and maintenance support, and wireless networks required denser, more complex deployment.
The past several years pushed that model into a new growth phase. As fiber deployments accelerated across the U.S., Dycom moved to strengthen both scale and geography. In 2023, it acquired Bigham, adding southeastern telecom construction capacity and expanding its ability to serve rural broadband programs. In 2024, Dycom bought Black & Veatch’s public carrier wireless telecom infrastructure business, extending its footprint in wireless modernization across several regions and adding capabilities relevant to 5G densification and Open RAN deployments.
At the same time, the company’s opportunity set began to stretch beyond consumer broadband and carrier upgrades. AI infrastructure depends on far more than chips and servers. Massive data centers need high-capacity fiber routes, reliable power infrastructure, structured cabling, wireless connectivity, and contractors capable of coordinating complex field work under tight timelines. Dycom was already positioned in several of those layers through its telecom and utility work, giving it a natural path into the physical buildout behind AI compute.
That broadening became clear in late 2025, when Dycom completed the acquisition of Power Solutions, a major Mid-Atlantic electrical contractor focused on data centers and other critical facilities. The deal moved DY deeper into the physical infrastructure layer of AI – where power, cabling, and execution capacity are as important as fiber routes. In 2026, it added another step in that direction, agreeing to acquire National Technology Integrators (NTI). The planned deal expands Dycom’s low-voltage, structured cabling, AV, and security capabilities for digital infrastructure customers.
Dycom did not suddenly become strategic because AI arrived. It spent decades building the workforce, field network, and customer trust needed to turn infrastructure demand into installed capacity. AI data centers simply pulled that capability into a larger arena.
1
The Fiber Frontier
Dycom sits at the field level of America’s digital-infrastructure buildout – where national broadband plans, telecom capital budgets, utility work, and AI data-center demand become permits, trenches, poles, fiber routes, electrical systems, and skilled crews. Its role is practical, physical, and increasingly strategic: turning long-term infrastructure demand into installed capacity.
That work now spans two distinct but increasingly connected layers. The Communications segment is DY’s connectivity layer – handling fiber routes, long-haul and middle-mile networks, aerial and underground construction, wireless work, engineering, maintenance, and fulfillment services. This is the business that helps telecom carriers bring fiber-to-the-home (FTTH) into more communities, while also building the high-capacity routes that connect data centers, cloud regions, businesses, homes, and end users.
The Building Systems segment is the facility layer. Created after the Power Solutions acquisition, it brings Dycom deeper inside data centers and other critical facilities through electrical systems, energy management, fire safety, security, and related infrastructure work. The planned acquisition of NTI adds another piece: low-voltage engineering, structured cabling, audio-visual systems, and security – the connective tissue inside the facility itself.
DY now reaches data-center infrastructure from both directions: Building Systems helps build and equip the facility itself, while Communications connects that site to the broader fiber network. In an AI world where computing campuses require more power, more cabling, and far larger data routes, that combined position is becoming more valuable.
Dycom’s core fiber business also has several growth curves running at once. FTTH remains the most immediate driver, with Q1 activity ramping faster than expected across multiple customers and markets. But the bigger strategic shift may be in long-haul and middle-mile fiber – the routes that connect data centers, cloud regions, carriers, and end users. As AI workloads grow, those routes need to carry far more traffic than traditional networks were built for. Management has pointed to hyperscaler-related routes that could eventually require 7,500-10,000 fiber strands, compared with current examples of 864 or 1,728. That kind of capacity does not get built in one spending cycle. It points to a multi-year, possibly decade-plus, expansion of the fiber architecture needed to support AI and cloud demand, with a more meaningful ramp likely in calendar 2027 and 2028.
Several broader trends reinforce the setup. Broadband expansion, AI data centers, wireless modernization, cloud growth, reshoring, and the re-industrialization of U.S. infrastructure all require skilled field execution. Federal and state programs such as BEAD1 add another potential source of demand as underserved areas move toward higher-speed connectivity. Meanwhile, customers are extending contract durations to secure DY’s workforce, underscoring how scarce skilled execution capacity has become.
These massive opportunities come with real risks. Dycom’s customer concentration remains high, telecom capex can move in cycles, projects can slip, weather and seasonality can affect margins and timing, and fuel, labor, and input costs can pressure execution. Still, DY’s position has broadened meaningfully. It is no longer just a fiber contractor riding one spending wave; it is becoming a scaled physical-infrastructure partner for the networks, data centers, and critical systems behind the next phase of digital growth.
—
1 – BEAD, or the Broadband Equity, Access, and Deployment program, is a U.S. federal broadband funding program designed to expand high-speed internet access, especially in unserved and underserved areas.
—
1
The Capacity Crew
Dycom’s financial profile is beginning to reflect the same shift visible in its operations: a larger company, a broader mix, and a backlog that gives management far more visibility than the old telecom-contractor label would suggest.
FQ1 2027 was the clearest evidence yet. Reported contract revenue reached $1.965 billion, up 56.1% year-over-year, with organic growth of 24.7%. That split tells the story: Dycom’s core Communications business is still growing quickly on its own, while the Power Solutions acquisition added a new Building Systems revenue stream tied to data centers and critical facilities. The resulting top-line number was far above analyst expectations, extending DY’s pattern of revenue outperformance.
The quality of the quarter was stronger than the headline growth rate alone. Adjusted EBITDA rose 74.6% year-over-year to $262.5 million, while adjusted EBITDA margin expanded to 13.4% from 11.9%. Profitability grew faster than revenue as DY benefited from scale, stronger utilization, better project execution, and a richer mix after adding Building Systems. Communications still carried the bulk of the business, generating $1.569 billion of revenue and $192.4 million of adjusted EBITDA, with margin rising to 12.3%. Building Systems contributed $395.4 million of revenue – already about 20% of the total – with a much higher 17.7% adjusted EBITDA margin.
Earnings followed the same pattern. Net income rose to $91.3 million, while adjusted net income reached $134.3 million and adjusted diluted EPS climbed to $4.42, flying past the consensus of $2.72. Although FQ1 included $0.41 per share of income-tax benefits from share-based awards, compared with $0.08 per share a year earlier, the outcome would have smashed analyst forecasts even with no tax benefits at all. The quarter continued a long record of bottom-line execution: DY has surpassed adjusted EPS expectations in all but one quarter since FQ3 2021, including 9 consecutive recent beats.
Backlog is the bridge between the quarter and the outlook. Total backlog reached $11.906 billion, up 46.5% year-over-year, with a 2.2x book-to-bill ratio. Communications backlog stood at $10.8 billion, while Building Systems backlog reached $1.1 billion. About $6.4 billion of total backlog is expected over the next 12 months, covering roughly 85% of the raised fiscal 2027 revenue guidance.
Management raised full-year fiscal 2027 revenue guidance from $6.85-7.15 billion to $7.38-7.65 billion, above analyst expectations of roughly $7.07 billion. The new range implies Communications revenue of $6.03-6.20 billion and Building Systems revenue of $1.35-1.45 billion. At the midpoint, management expects about 38% total revenue growth and 14% organic growth, excluding last year’s extra week.
FQ2 guidance also points to continued strength: revenue of $1.94-2.01 billion versus consensus of about $1.77 billion, adjusted EBITDA of $284-303 million, and adjusted diluted EPS of $4.40-4.82 versus consensus of about $4.06. Importantly, both FQ2 and full-year guidance exclude the pending NTI acquisition, which is expected to add about $175 million of initial annual revenue run rate at mid-to-high-teens adjusted EBITDA margins.
The balance sheet is carrying more debt after the Power Solutions acquisition, but not at a stressed level. Dycom ended FQ1 with $538.8 million in cash and equivalents, about $1.29 billion of liquidity, $2.30 billion of notional net debt, and no revolver borrowings. Long-term debt stood at $2.81 billion, while quarterly adjusted EBITDA covered net interest expense by roughly 7.4x. Cash flow remains the main watchpoint: rapid project growth ties up cash in working capital before customer payments arrive, so operating cash flow was still negative $24.6 million and free cash flow was negative $92.1 million, though both improved year-over-year. Moreover, DSO2 improved to 96 days from 111, showing better collection discipline in a business where working capital can move heavily with project timing.
Dycom is scaling quickly, but it is still an execution-heavy contractor. Margins depend on crew utilization, project mix, customer timing, labor availability, and cost control. The financial story is strong because growth, backlog, margins, and guidance are now moving in the same direction. The challenge is keeping them aligned as the company becomes larger, more acquisitive, and more exposed to data-center infrastructure.
—
2 – DSO stands for Days Sales Outstanding, which measures how many days, on average, it takes a company to collect payment after recognizing revenue – an important measure for companies performing large project-based work.
—
1
Current Formation
Dycom sits in the same infrastructure-investment orbit as MasTec and Sterling Infrastructure – both Smart Portfolio holdings – as well as MYR Group, but it is not a perfect substitute for any of them. MasTec provides the closest operating comparison through its large communications and utility-infrastructure businesses, but DY is more concentrated in telecom contracting, especially fiber, network deployment, and field services for major carriers. Sterling is more tilted toward e-infrastructure, data-center site development, and the physical campus work supporting hyperscalers. DY touches the same AI infrastructure theme from a different angle: the fiber routes, connectivity, and network capacity feeding those data centers. MYR Group adds a useful benchmark for specialized electrical and utility contracting, helping frame Dycom as a more focused connectivity-infrastructure play.
The entire peer group has benefited from the AI infrastructure investment cycle over the past year, while also absorbing the recent selling pressure across AI hardware and infrastructure names, triggered primarily by investor worries over elevated valuations. Despite the recent volatility, Sterling still outperformed by far over the past 12 months, rising roughly 200% as one of the clearest public-market proxies for AI-related site development and e-infrastructure. MYR was a strong second, driven by its exposure to the massive power demand surge tied to AI data centers. MasTec also delivered a strong, though slightly lower, gain, as its diversified business benefited from momentum across telecom, energy, and AI-related infrastructure.
Meanwhile, Dycom entered the market’s AI-infrastructure narrative later than its peers and was long perceived as a more telecom-exposed contractor. The company made its strategic move into data centers in late 2025, missing much of the prior rally and posting a relatively modest gain of about 70% over the past year. However, the performance gap has narrowed after DY proved its data-center and fiber relevance in the latest quarter – while still leaving the stock less exposed to the most extreme valuation risks seen across some AI-infrastructure winners.
Relative to peers, DY’s valuation looks unusually reasonable for a company now delivering this level of growth. It trades at roughly 26x FY1 non-GAAP earnings and 22x FY2 – below MasTec, Sterling, and MYR Group on both measures, while its 14.3x forward EV/EBITDA is also by far the lowest in the group. Forward EV/Sales tells a more nuanced story: at 2.0x, DY is trading slightly above MasTec and MYR, but far below Sterling’s 5.5x, reflecting the premium investors assign to STRL’s data-center and e-infrastructure exposure. The strongest valuation point, however, may be growth-adjusted: DY’s forward PEG of just 0.82 is low both in absolute terms and compared to peers. That discount looks notable because DY’s expected revenue growth and EBITDA margin are second only to Sterling.
In short, DY is no longer a hidden fiber contractor, but it is still priced below peers despite above-average growth, improving mix, and rising AI-infrastructure relevance. These factors support the Wall Street consensus price target, implying more than 50% upside from current levels for the Strong Buy-rated stock. The stock offers a rare combination in the current infrastructure trade: a business whose AI relevance is becoming clearer, but whose valuation has not yet fully caught up to that shift.
Dycom has maintained an active stock repurchase program for years, with opportunistic buybacks ranking behind organic investment and M&A in the capital-allocation hierarchy. In February 2025, Dycom’s board authorized a new $150 million buyback plan, replacing the prior authorization. The company repurchased $65.6 million worth of stock during fiscal 2026, which ended in January 2026, and added roughly $36.0 million in buybacks in fiscal Q1 2027, signaling continued commitment to returning capital to shareholders when valuation and cash flow support it.
1
Investing Takeaway
Dycom is moving closer to the center of America’s digital-infrastructure buildout. Fiber expansion remains the foundation, but the company’s opportunity is now widening across data-center power, structured cabling, long-haul connectivity, wireless modernization, and federally supported broadband deployment. That gives DY more than one path to growth, while its skilled workforce and long-standing carrier relationships remain difficult advantages to replicate. The risks are real – customer concentration, project timing, labor availability, and cost pressure can all affect results – but they look more tied to execution than to demand. As the AI buildout’s bottleneck shifts from servers and chips to the physical networks and facilities around them, Dycom is increasingly positioned where the next layer of investment is slated to land.
1
Smart Investor’s Winners Club
The Winners Club represents stocks from the Smart Investor Portfolio that have risen at least 30% since their purchase dates.
The markets were extremely volatile, but the Club member count remained steady with 28: GE, AVGO, TSM, ANET, HWM, APH, EME, VRT, IBKR, ASX, MTZ, STRL, PH, CRWD, PANW, GOOGL, CSCO, KEYS, ATI, JBL, ORCL, BNY, MS, CRDO, JPM, RTX, C, and NVT.
The first runner-up is still LLY with a 24.95% gain since its purchase two months ago. Will it break into the winners’ circle, or will another stock outrun it to the finish line?
1
|
