TipRanks Smart Growth Portfolio #55: Mind the Gap
Dear Investors,
Welcome to the 55th edition of the Smart Growth Portfolio and Newsletter, where we spotlight a company connecting the inner workings of modern chips. But first, here are some news and updates.
1
1
Portfolio News
❖ Telos (TLS) delivered a clear beat-and-raise quarter, capping off what management called a “transformational” year. Q4 revenue surged 77% year-over-year to $46.8 million, ahead of expectations, while both adjusted EPS of $0.06 and adjusted EBITDA of $7.3 million comfortably exceeded guidance and flew past analyst consensus. The performance was driven by continued strength in Security Solutions, particularly Telos ID, which remains the core growth engine.
Just as important, the company followed through with a higher 2026 outlook. Management now expects revenue of $187-200 million, up from prior indications of ~$180 million, alongside adjusted EBITDA of $20.6-28.0 million and expanding margins. Near-term momentum also looks solid, with Q1 revenue guidance above consensus. At the same time, Telos increased its share repurchase authorization to $75 million after already returning $13.6 million in 2025, reinforcing confidence in its improving cash flow profile.
Still, the story is not without friction. GAAP results remain negative, weighed down by a $14.9 million goodwill impairment tied to the declining Secure Networks segment, as well as restructuring costs. Margin dynamics also bear watching, with gross margins compressing due to revenue mix and expected to remain under pressure in 2026. These factors, alongside broader weakness across software stocks in recent months, likely explain why several analysts trimmed price targets despite maintaining generally constructive ratings.
The setup here is shifting. Telos has largely proven it can grow and generate cash, but the next phase depends on consistency. We will watch the stock closely to see whether the company can sustain margin expansion, convert its pipeline into durable revenue, and continue scaling its higher-quality Security Solutions business without further drag from legacy segments.
1
❖ Applied Digital (APLD) continues to strengthen its position as a critical AI infrastructure partner. The company is pioneering next-generation efficiency in its AI data centers through a groundbreaking collaboration with NVIDIA, CoreWeave, and Phaidra.
The companies released the joint white paper “AI Agents for Liquid-Cooled AI Factories,” introducing agentic liquid cooling management for NVIDIA Max-Q AI factories. At the heart of APLD’s Polaris Forge campuses, Phaidra’s self-learning reinforcement-learning AI Agent now integrates with the company’s proprietary liquid cooling systems and NVIDIA DSX Max-Q. Tested live on NVIDIA Grace Blackwell and GB200 NVL72 clusters (including CoreWeave environments), the AI agent uses real-time rack power data to predict and preempt thermal spikes, reducing overshoot magnitude by 75-80% versus traditional PID controls while cutting response time from 3-5 minutes to under 10 seconds.
By co-designing power, cooling, and workload management as one unified system, the technology safely raises coolant temperatures, slashes chiller load, and redirects stranded cooling power directly to revenue-generating AI compute. This directly supercharges Applied’s existing and upcoming facilities – including those leased to CoreWeave and Oracle – unlocking higher density, superior reliability, stronger SLAs, and billions in potential extra annual revenue at gigawatt scale. It transforms APLD from a pure capacity provider into a differentiated, efficiency-leading AI landlord with a clear operational edge in the power-constrained AI buildout.
1
❖ Micron (MU) delivered one of the strongest quarters in its history, posting a clear beat-and-raise as AI-driven demand continues to reshape the memory market. Fiscal Q2 revenue surged to $23.9 billion, nearly tripling year-over-year and well ahead of expectations, while adjusted EPS of $12.20 decisively beat consensus, surging by almost 700% year-over-year. Margins and cash generation reached exceptional levels, with gross margins at 75% and free cash flow hitting a record $6.9 billion, highlighting powerful pricing across DRAM and NAND.
The company followed through with decisive upward guidance. Q3 revenue is expected to reach roughly $33.5 billion with gross margins around 81% and EPS near $19.15 – far above Street estimates and implying another step-change in earnings power. Management also reinforced confidence through capital returns, raising the dividend by 30% to $0.15 per share while continuing opportunistic buybacks alongside a strengthened balance sheet.
At the same time, MU is leaning aggressively into the opportunity. Fiscal 2026 capital expenditures are now expected to exceed $25 billion, with an even larger step-up planned for 2027 as the company expands DRAM, HBM, and advanced manufacturing capacity. This is not speculative spending – supply remains severely constrained, with management indicating it can currently meet only about 50-66% of key customer demand, reinforcing the structural AI-driven supply gap. Micron is also advancing its technology lead, having started volume shipments of next-generation HBM4 memory for NVIDIA’s upcoming Vera Rubin platform, further strengthening its position at the center of AI infrastructure demand.
Still, the stock dropped despite the blowout print, driven by a sharp risk-off shift across markets. Oil prices have surged amid escalating Middle East hostilities, while the Fed’s latest signals around “sticky inflation” have pushed expectations toward zero rate cuts this year, pressuring high-beta tech shares. Against that backdrop, MU became a natural candidate for profit-taking. The stock has surged more than 350% over the past year and just delivered near-peak-looking metrics, with margins approaching 80% and expectations reset sharply higher. At the same time, the step-up in capital spending introduces timing concerns around margin durability, even if strategically justified.
Wall Street, on the other hand, is looking past the sell-off. Analysts have broadly raised price targets and reiterated bullish ratings, pointing to structural supply shortages and sustained AI-driven demand as key supports. This is a positioning reset, not a fundamental shift – if anything, it may help to set up the next leg higher.
This reframes the setup. Micron has already proven the strength of its business, and the supply-demand imbalance remains firmly in its favor, with demand expected to outpace supply well into 2027. There is no clear near-term fundamental catalyst for a slowdown – current constraints point to continued pricing power and earnings strength. The real variable is external: rising macro uncertainty, where higher oil prices, persistent inflation, and a more restrictive Fed stance are driving risk-off behavior. After the surge, MU has become a source of liquidity for investors looking to lock in gains, especially as capital spending ramps and peak-cycle concerns resurface.
From here, the decision is less about Micron’s execution and more about the broader environment. We will watch the stock closely through that lens – whether macro conditions deteriorate into a recession or stagflation scenario that forces de-risking, or whether stability returns and allows Micron’s underlying strength to reassert itself. In the absence of a broader market breakdown, the current fundamentals offer little reason to step aside.
1
1
1
This Week’s Top Growth Pick: Arteris (AIP)
Arteris, Inc. develops system IP that helps semiconductor companies design and connect complex chips. Its technology manages how different parts of a system-on-chip communicate – linking processors, memory, and specialized accelerators into a cohesive architecture. As chip designs grow more complex, driven by artificial intelligence, automotive systems, and high-performance computing, efficient data movement inside the chip becomes a key constraint. Arteris operates at that internal connectivity layer – enabling more predictable performance, better scalability, and greater visibility across increasingly sophisticated silicon designs.
Source: Arteris, Inc. Corporate Overview, February 2026
1
Connecting Complexity
Arteris was founded in 2004 with a focus on improving how components inside a system-on-chip (SoC)1 communicate, pioneering network-on-chip (NoC)2 interconnect technology as an alternative to traditional bus-based3 designs. As chip complexity increased, this approach gained traction among semiconductor companies building more advanced architectures and managing rising design complexity.
The company’s transformation into a broader system-level platform has largely taken shape over the past several years. A key early step came in 2020 with the acquisition of Magillem assets, which expanded Arteris beyond interconnect into system integration and design automation. This shift accelerated following its 2021 IPO on Nasdaq – under the AIP ticker – which increased visibility and supported further investment in platform expansion.
Several strategic deals reinforced that trajectory. The 2023 acquisition of Semifore added hardware-software interface capabilities, while the 2026 acquisition of Cycuity introduced hardware security verification tools, extending the company into security assurance at the silicon level. Alongside that, deep partnerships with ARM, Synopsys, chip foundries, and ecosystem players have further embedded Arteris in modern design flows. Together, these moves positioned the company beyond connectivity – toward a more comprehensive role in chip design and validation.
Product innovation supported AIP’s evolution, with the company expanding from a specialized NoC IP provider into a broader system-level design partner. The company continued advancing its FlexNoC and Ncore platforms while introducing new capabilities such as CodaCache for last-level caching and FlexGen, a newer generation of interconnect technology designed to handle increasingly complex, AI-driven workloads and multi-die4 architectures.
End-market exposure also broadened, with NoC deployments in over four billion units across more than 200 customers. While automotive became a key driver – supported by industry-wide adoption of software-defined architectures and ADAS – Arteris increasingly gained traction in AI and machine learning, data center infrastructure, 5G systems, and chiplet-based designs. This expansion reflects growing adoption across multiple high-performance end markets, including advanced chiplet architectures and edge AI platforms.
—
1 – System-on-chip (SoC) – A semiconductor design that integrates multiple functions – such as processors, memory, and input/output – into a single chip, allowing devices to run complex workloads within a compact and power-efficient footprint.
2 – Network-on-chip (NoC) – An internal communication system within a chip that routes data between its components using a network-like structure, enabling faster, more scalable, and more efficient data movement as chip complexity increases.
3 – Bus-based architecture – A traditional method of connecting chip components through a shared communication pathway (“bus”). While simpler, it can become a bottleneck as more components compete for bandwidth. AIP’s NoC approach replaces this with parallel, routed connections, improving performance, scalability, and efficiency.
4 – Multi-die architectures – A design approach where multiple smaller chips (“dies”) are combined into a single package, working together as one system. This allows for greater flexibility, better yields, and higher performance compared to building everything on a single large chip.
—
Source: Arteris, Inc. Corporate Overview, February 2026
1
Fabric of Compute
Arteris operates at a critical layer of modern semiconductor design – the movement of data inside increasingly complex chips. As compute shifts toward AI-driven workloads, performance is no longer defined solely by processing power or memory capacity, but by how efficiently data flows between them. That shift has turned interconnect architecture into a core design constraint. Data movement is now a bottleneck, not a background function – which is precisely the layer AIP is designed to optimize.
At its foundation, the company develops and licenses NoC intellectual property – an interconnect architecture that acts as the internal communication fabric of a system-on-chip – delivered through its FlexNoC5 and Ncore6 platforms. Instead of building interconnect systems from scratch – a costly and high-risk process – chip designers rely on AIP’s pre-validated IP to shorten development cycles and avoid redesign cycles. As chip architectures evolve toward multi-die and chiplet-based7 designs, that role becomes more central, not less.
Over time, AIP has expanded beyond interconnect into a broader system-level platform. FlexGen8 introduces AI-driven automation to interconnect design, enabling faster development cycles, improved layout efficiency, and measurable gains in latency and power. CodaCache9 adds last-level cache functionality, while Magillem10 software supports full-chip integration and orchestration, reducing design time by roughly 30%. The 2026 acquisition of Cycuity extends the platform into hardware-level cybersecurity verification, addressing a fast-growing market as vulnerabilities increasingly shift from software into silicon itself. This move positions AIP not just as a performance enabler, but as a validation and assurance layer in chip design.
This expanding product stack is central to how AIP makes money and scales. The model is increasingly driven by selling multiple products into the same customer, raising content per design and enabling cross-sell opportunities. Full-suite deals already exceed $1 million, and the addition of security expands the addressable market beyond traditional interconnect buyers. As designs move into production, royalties tied to chip volumes create a second growth engine, turning design wins into recurring revenue streams.
Adoption reflects that positioning. AIP’s technology has been deployed in more than 4 billion chips and chiplets, with production volumes rising across automotive, data center, cloud, edge AI, industrial, and aerospace applications. Automotive remains the largest individual end market, supported by the industry’s shift toward ADAS, autonomy, and software-defined vehicles. At the same time, AI-driven workloads now represent more than half of the business, cutting across data centers, edge systems, and intelligent devices. Importantly, this is not just design activity – it is real silicon deployment at scale.
The customer base reinforces that credibility. AIP is used by 9 of the top 10 semiconductor companies, with more than 275 customers and over 900 SoC design starts, alongside retention above 90%. Partnerships and integrations with ARM, Cadence, Samsung Foundry, and other ecosystem players position the company as a platform-neutral layer that works across processors, IP blocks, EDA tools, and manufacturing flows. That neutrality lowers adoption friction and allows AIP to sit inside complex, multi-vendor design environments.
The broader market backdrop provides a clear runway. System IP and chip design infrastructure represent a multi-billion-dollar market growing at a double-digit pace, driven by rising complexity, chiplet adoption, and the need for power-efficient, high-performance architectures. Despite strong adoption, AIP’s share of that market remains modest, leaving room for expansion as its platform deepens and standardization efforts in areas like automotive chiplets and AI accelerators take shape.
Risks remain, however. The business still requires heavy R&D investment, and profitability is emerging rather than established. Large customers retain the option to develop internal solutions, and execution on platform expansion is critical. Still, the direction is clear: as chips become more distributed, more interconnected, and more security-sensitive, the infrastructure that ties them together becomes indispensable – and AIP is positioning itself directly at that junction.
—
5 – FlexNoC – Non-coherent interconnect IP for general data movement between components inside a chip, optimized for performance, lower power, and scalability when full data synchronization is not required.
6 – Ncore – Cache-coherent interconnect ensuring all processors and accelerators share consistent data across the system, unlike non-coherent interconnects that do not automatically synchronize shared data.
7 – Chiplet-based designs – Chips built from multiple smaller components (chiplets) connected together, often across separate dies (a subset of multi-die architectures).
8 – FlexGen – AI-assisted tool that automatically generates and optimizes interconnect architectures for complex chip designs.
9 – CodaCache – Last-level cache IP that improves data access efficiency and reduces memory latency.
10 – Magillem software – Design automation tool that integrates and validates all components of a system-on-chip.
—
Source: Arteris, Inc. Corporate Overview, February 2026
1
License to Scale
Arteris is still in the early innings of scaling, but the financial trajectory is starting to reflect a business moving from pure development toward commercialization. The most recent quarter reinforced that shift. Q4 2025 revenue reached $20.1 million, up 30% year-over-year and above the high end of management’s guidance, extending a streak of revenue beats that now spans five consecutive quarters. Full-year revenue came in at $70.6 million, up 22% year-over-year, with growth supported by both licensing activity and a rapidly expanding royalty base.
That royalty component is becoming increasingly important. Variable royalties rose 50% year-over-year to $6.6 million, and the number of meaningful royalty-paying customers has expanded from one five years ago to nine today. This reflects a transition from design wins to real production scale, with more than 4 billion chips shipped using AIP technology. As those volumes increase, royalties begin to compound, creating a more durable and recurring revenue layer on top of the licensing business.
Forward indicators point to continued momentum. Annual contract value (ACV) plus royalties reached $83.6 million, up 28% year-over-year, while remaining performance obligations (RPO) climbed to roughly $117 million, up 32%. About half of that backlog is expected to convert into revenue within the next twelve months, providing improved visibility – though not perfectly linear due to timing variability. Management’s guidance reflects that trajectory, calling for 2026 revenue of $89-93 million, implying roughly 25-32% year-over-year growth, alongside ACV plus royalties exceeding $100 million.
Profitability is not yet established, but the direction is becoming clearer. Gross margins remain exceptionally high at around 90-92%, typical of an IP-driven model where incremental revenue carries significant drop-through potential. Early operating leverage is beginning to emerge, with revenue growing 22% year-over-year compared to operating expense growth of roughly 14%, signaling that the cost base is scaling more slowly than the top line.
That said, the business remains intentionally investment-heavy. R&D spending – roughly $50 million annually – continues to fund platform expansion and multi-product development, keeping operating margins negative for now. Profitability, therefore, is less a question of gross economics and more a function of scale, with the company targeting non-GAAP operating profitability by the fourth quarter of 2026. On a GAAP basis, losses remain significant, with a $34.7 million net loss for the year, reflecting the ongoing cost of building out the platform.
Cash generation is a notable bright spot. Arteris delivered $5.3 million in free cash flow for the full year, or about 8% of revenue, marking a clear inflection toward a self-funding model. The balance sheet remains clean, with approximately $59.5 million in cash and no debt, providing flexibility to continue investing without immediate reliance on external capital.
There are still moving parts. The Cycuity acquisition is expected to contribute roughly $7 million in revenue in 2026, but will also introduce modest near-term margin pressure and about $1 million in operating loss. Royalty revenue can be uneven due to audit timing, and roughly half of the contracted backlog sits beyond the current year. Still, the broader picture is constructive: accelerating revenue, a scaling royalty engine, improving cash generation, and a visible path toward profitability. For a company positioned at the center of rising semiconductor complexity, the financial model is beginning to align with the opportunity.
Source: Arteris, Inc. 4Q 2025 Earnings Presentation
1
Flow to Yield
Arteris operates within a narrow but increasingly relevant segment of the semiconductor IP landscape, where a handful of public peers offer useful benchmarks across its growth trajectory. CEVA represents the closest direct comparison today, with a similar licensing-and-royalty model at a comparable scale. Rambus reflects the more advanced version of that model, where royalties and margins have fully scaled, while Adeia anchors the mature, highly profitable end of pure IP monetization. QuickLogic adds a small-cap growth-stage reference. Together, this spectrum frames AIP’s positioning – between early-stage execution and long-term royalty-driven scale.
Over the past year, the divergence in stock performance within that peer group has been largely driven by where each company sits on the AI monetization curve. AIP’s roughly 96% gain reflects a rapid re-rating as it moved from design-win visibility to tangible royalty scaling, with AI now driving a majority of its business and reinforcing its role at the center of chiplet-based architectures. That momentum – supported by consistent beats and forward guidance – contrasts sharply with CEVA, whose decline reflects slower growth, weaker royalty trends tied to consumer and IoT markets, and dilution that together have pressured sentiment despite a similar model on paper. Rambus and Adeia advanced on clearer monetization, with AI-driven memory demand and stable licensing cash flows, respectively, while QuickLogic benefited from smaller-scale AI and defense catalysts. The pattern is consistent – the market is rewarding direct AI exposure with visible revenue conversion, not just participation in the theme. That logic also underpins Wall Street’s optimistic forward view of the Strong Buy-rated Arteris, with the average price target implying over 35% upside from current levels, even after the past year’s rally.
Despite leading the group in stock performance, AIP still trades more like a scaling story than a fully priced asset, at roughly 9.3x trailing EV/Sales – a large and justified premium to CEVA, but still well below Rambus at 13.2x. That positioning reflects a balance between growth and execution risk. AIP’s revenue growth of over 22% year-over-year already approaches Rambus, while its forward growth is expected to be the fastest in the group, reinforcing its premium relative to slower-growing peers like CEVA and Adeia. On a forward basis, the valuation begins to compress more meaningfully, with AIP trading at roughly 7.2x EV/Sales – moving closer to mid-tier peers despite superior growth, while still leaving a gap to higher-multiple names like Rambus. At the same time, improving fundamentals support the valuation: gross margins remain exceptionally high at over 90%, and the company has reached positive free cash flow, signaling early operating leverage. In short, the market is starting to price in AIP’s growth trajectory, but not yet its full royalty-scale potential. This leaves ample room for further multiple expansion if execution continues.
1
To Sum It All Up
Arteris operates at a critical layer of modern computing – the movement of data inside increasingly complex chips. As AI workloads scale and architectures shift toward chiplets and multi-die systems, efficient interconnect design is becoming a foundational requirement, not an optimization. Arteris sits directly in that transition, providing the infrastructure that allows processors, accelerators, and memory to function as a unified system. Its expanding platform, growing adoption across leading semiconductor players, and increasing exposure to AI-driven designs position it within one of the most important shifts in the industry. While still scaling, the business is beginning to translate design wins into broader deployment and recurring royalties. If execution continues and the industry’s move toward more complex architectures accelerates, Arteris has a clear path to evolve from a niche IP provider into a core enabler of next-generation compute.
1
1
Smart Growth Portfolio
|
1
Disclaimer
The information contained in this article represents the views and opinions of the writer only, and not the views or opinions of TipRanks or its affiliates and should be considered for informational purposes only. TipRanks makes no warranties about the completeness, accuracy, or reliability of such information. Nothing in this article should be taken as a recommendation or solicitation to purchase or sell securities. Nothing in the article constitutes legal, professional, investment and/or financial advice and/or takes into account the specific needs and/or requirements of an individual, nor does any information in the article constitute a comprehensive or complete statement of the matters or subject discussed therein. TipRanks and its affiliates disclaim all liability or responsibility with respect to the content of the article, and any action taken upon the information in the article is at your own and sole risk. The link to this article does not constitute an endorsement or recommendation by TipRanks or its affiliates. Past performance is not indicative of future results, prices, or performance.
