Cache the Wave
In this edition of the Smart Investor newsletter, we spotlight the engine room of enterprise AI. But first, let’s review the latest news and developments.
1
Portfolio News and Updates
❖❖ Alphabet’s (GOOGL) Google chose not to participate in the U.S. government’s recent $2 billion quantum funding initiative. Google Quantum AI COO Charina Chou cited concerns that the program’s conditions could slow the technology’s development. The move reflects GOOGL’s confidence in its own quantum roadmap and its willingness to prioritize operational flexibility over government capital.
That confidence is well-founded. In a March 2026 paper – freshly back in the headlines this week as Ethereum continues to notch heavy losses – Google Quantum AI researchers slashed the estimated hardware needed to break the cryptography securing Bitcoin and Ethereum by roughly 20 times, to fewer than 500,000 physical qubits. For ETH specifically, the logical-qubit estimate fell from tens of thousands to around 1,200. Crucially, the elliptic-curve encryption at risk isn’t unique to crypto – it also underpins much of mainstream digital security, including Google’s own systems. The work was credible enough that Google set itself an internal 2029 deadline to migrate its own systems to quantum-safe encryption.
❖ In other news, Google announced a $1.5 billion investment across 2026 and 2027 to expand its data center campus in Jackson County, Alabama, reinforcing the scale of Big Tech’s AI infrastructure buildout. The site has operated since 2019 on a repurposed former coal-plant property, turning an old energy asset into part of Google’s cloud and digital-services backbone. GOOGL said it will fund 100% of its own power and infrastructure costs tied to the expansion, an important detail as data center growth faces rising scrutiny over electricity demand and grid pressure.
❖ In yet other news, GOOGL cut the price of its entry-level AI Plus plan from $7.99 to $4.99 per month, a small consumer move that may point to a much larger shift across the AI market. At the same time, reports indicate that OpenAI is considering steep token-price cuts as competition with Anthropic intensifies – just as both AI labs move toward the public markets.
This is another signal that foundational AI models are undergoing a fast commoditization process, and their providers are starting to face real price pressure. Frontier systems are still differentiated at the high end, but for many everyday business tasks, models are becoming increasingly interchangeable. That setup resembles a capital-intensive AI utility with uncertain pricing power more than a premium software model.
This is a much tougher setup for OpenAI and Anthropic than for a massive tech conglomerate. GOOGL can use AI as part of its sprawling ecosystem – Search, YouTube, Android, Workspace, and Cloud – while pure-play AI labs need model usage itself to become highly profitable. That matters as public investors start asking harder questions about gross margins, compute costs, customer lock-in, and cash burn.
The implication for stocks in the AI universe is that the winners may not be the model providers, but rather the companies that own the layers where pricing power can survive. GOOGL benefits from its distribution power across its many businesses, while Microsoft (MSFT) gains from enterprise workflow control through Microsoft 365, Azure, GitHub, and Copilot. Oracle (ORCL) profits through proprietary enterprise data sitting inside its database and applications stack, and Amazon (AMZN) benefits from cloud capacity through AWS. Alongside data and software, “the usual suspects” that gain the most are the hardware and infrastructure leaders. NVIDIA (NVDA) provides the GPUs powering the compute layer, Broadcom (AVGO) supplies networking gear and custom silicon to GOOGL and others, and Vertiv (VRT) provides the power and thermal infrastructure needed to keep the AI buildout running.
1
❖❖ SpaceX’s historic IPO has driven a lot of excitement and made many investors significantly richer. However, one of the biggest beneficiaries may ultimately be NVIDIA (NVDA) – especially after Elon Musk’s post-IPO message about “taking our exciting partnership with NVIDIA to the next level” put a fresh spotlight on that opportunity.
Musk’s AI lab, xAI – now merged with SpaceX – has been heavily reliant on NVDA’s GPUs for data center clusters such as Colossus. Terafab – a massive chip foundry in Texas with terawatt-scale ambitions, developed by Tesla, SpaceX, and xAI – is built on NVIDIA’s next-generation Rubin architecture. The IPO windfall gives SpaceX more firepower to finance the scaling of these businesses, alongside capex for Musk’s more futuristic ventures, including tighter space and AI synergies. Notably, SpaceX has plans for orbital data centers – and NVIDIA is developing space-optimized hardware.
Overall, SpaceX plans to expand its capex to $300 billion through the end of the decade, with ambitions to transform xAI into a $1.75 trillion business. This massive investment will benefit not only NVIDIA, but also the broader AI data center and rocket manufacturing supply chain – semiconductors, connectivity, cybersecurity, automation, electronics, power infrastructure, and more.
❖ In other news, NVIDIA raised $25 billion in high-grade bonds – more than the originally planned $20 billion, thanks to investor demand coming in at 4x the company’s target, reflecting intense appetite for AI-linked credit. This is NVIDIA’s first debt issuance since 2021, aimed at refinancing and other corporate purposes.
1
❖❖ Oracle (ORCL) dropped sharply after its fiscal Q4 2026 report – before recovering part of the decline – as a clean beat-and-raise was overshadowed by the scale of the company’s AI spending plans.
The quarter itself was strong. Revenue rose 21% year-over-year to a record $19.2 billion, while total cloud revenue jumped 47% to $9.9 billion. The core AI story was even stronger: Oracle Cloud Infrastructure (OCI) revenue surged 93% to $5.8 billion, with CPU and GPU infrastructure revenue up 119% to $4.8 billion. Cloud database revenue grew 29%, while multi-cloud database revenue soared 404%. This shows that ORCL’s database franchise is no longer just a legacy software business, but an integral part of its broader AI and cloud infrastructure growth base.
Earnings also cleared the bar. Non-GAAP EPS rose 24% year-over-year to $2.11, or 20% to $2.03 excluding one-time net investment gains. For the full fiscal year, Oracle generated $32 billion of operating cash flow, up 54% year-over-year, even as free cash flow turned deeply negative because of the company’s aggressive data center buildout.
The biggest number, however, was RPO. Remaining performance obligations rose 363% year-over-year to $638 billion, up $85 billion sequentially from Q3. Oracle also signed $67 billion of AI infrastructure contracts during the quarter, while prepaid and bring-your-own-hardware contracts reached $75 billion. That matters because these structures lower Oracle’s own funding burden, with customers either paying upfront or supplying GPUs while ORCL provides the infrastructure, networking, security, and cloud services around them.
Management also pointed to accelerating delivery. Oracle delivered more than 1.2 gigawatts of capacity in fiscal 2026, and Q1 FY27 deliveries are approaching 1 gigawatt – nearly equal to the prior four quarters combined. That pace helps explain the rising capex, while reinforcing the signal of exceptionally strong demand, as ORCL’s global GPU utilization was 97.5% in FQ4, while 98% of AI data center capacity is already contracted.
Guidance added to the constructive side of the story. Oracle maintained its aggressive FY27 revenue target of $90 billion, implying 34% constant-currency growth, and raised non-GAAP EPS guidance to $8.05, representing 18% growth after excluding fiscal 2026 one-time gains. For Q1, management expects revenue growth of 27-29%, cloud revenue growth of 58-64% in U.S. dollars, and non-GAAP EPS of $1.72-1.76. Management also said revenue and earnings should accelerate in the second half as more megawatts come online.
Many analysts called ORCL’s post-earnings stock drop overdone, but it wasn’t all irrational. Oracle expects around $70 billion of fiscal 2027 net cash outlay for capex, with reported capex even higher after including customer prepayments and timing impacts. The company also plans to raise about $40 billion through debt and equity, including a previously announced $20 billion at-the-market equity program. That raises fair concerns about leverage, dilution, gross-margin pressure, and the timing of free-cash-flow recovery.
However, the comparison also needs context. ORCL has now firmly entered the hyperscaler category, and hyperscale AI has become a brutally capital-intensive business. Even the strongest players are tapping outside capital while spending heavily on AI infrastructure. The difference is that MSFT, AMZN, and GOOGL have larger internal cash engines, so investors give them more room for error. Oracle already has the hyperscaler demand and cloud growth profile – but investors still demand proof that it can convert its massive RPO into revenue and cash flow without overstretching the balance sheet.
Analyst moves reflected the split between recognition of ORCL’s extremely strong underlying business momentum and concerns about its execution timeline and ballooning financing needs – with each verdict depending on the weight attached to each variable in the equation. Evercore ISI, D.A. Davidson, Guggenheim, Bernstein, Goldman Sachs, BMO Capital, Barclays, and Piper Sandler raised their price targets, while Wedbush and Scotiabank cut theirs. Guggenheim went further, calling the selloff an opportunity to buy aggressively, while Wedbush flagged the debt/equity raise as a clear near-term overhang. Even after the cuts, however, the revised targets still implied 30%+ upside.
1
Portfolio Earnings and Dividend Calendar
❖ The Q1 2026 earnings season has ended, and Jabil’s (JBL) report today will wrap up the season for the Smart Investor Portfolio holdings.
❖ The ex-dividend date for Broadcom (AVGO) is June 22, while for Amphenol (APH) it is June 23.
w
1
New Buy: Hewlett Packard Enterprise (HPE)
Hewlett Packard Enterprise operates in one of the most demanding layers of modern computing – the infrastructure that allows enterprises, governments, and research institutions to run cloud, supercomputing, and artificial intelligence workloads at scale. HPE’s portfolio spans enterprise servers, hybrid cloud platforms, high-performance computing systems, AI-optimized architectures, networking, storage, and liquid cooling – areas where performance, reliability, energy efficiency, and system integration are becoming harder to separate. The company sits between traditional enterprise IT and the next generation of sovereign AI infrastructure, supplying the hardware, software, and services needed to build secure, high-density computing environments. As data centers move from general-purpose cloud expansion toward AI factories, national computing capacity, and liquid-cooled server architectures, HPE is positioning itself as a core infrastructure provider for the next phase of enterprise and sovereign compute.
1
Compute and Conquer
Hewlett Packard Enterprise was born from one of the most important corporate separations in technology. When the old Hewlett-Packard split in 2015, HP Inc. kept the PC and printer business, while HPE inherited the enterprise infrastructure side – servers, storage, networking, services, and the responsibility of proving that legacy hardware could still matter in a cloud-first world.
The early years were about sharpening that focus. HPE moved away from slower, less strategic activities and leaned into areas where enterprise customers needed specialized infrastructure: hybrid cloud, high-performance computing, storage, networking, and edge connectivity. The 2019 acquisition of Cray became one of the defining moves. At the time, it looked like a supercomputing deal. In hindsight, it gave HPE a head start and a deeper position in the engineering spheres that AI infrastructure now demands – dense systems, extreme performance, and liquid cooling.
That foundation has become even more valuable in recent years, with HPE building on its advantages to turn its infrastructure base into a more intelligent, software-defined platform for complex enterprise environments. Determined AI, acquired in 2021, brought ML training software into HPE’s HPC stack, linking its supercomputing strength more directly to enterprise AI development. The 2023 acquisitions of OpsRamp, Axis Security, and Athonet widened the platform around observability, cloud security, and private 5G – three areas tied to the way large organizations deploy hybrid infrastructure. In 2024, Morpheus Data added multicloud orchestration and self-service provisioning, giving GreenLake – HPE’s hybrid cloud platform – a stronger software layer for customers managing applications across public cloud, private cloud, and on-premises systems.
The company’s most transformative recent move came in 2025, when HPE closed its acquisition of Juniper Networks. That deal doubled HPE’s networking business and added a full AI-native networking stack at a time when AI clusters are making network performance a core infrastructure constraint.
At the same time, HPE’s relationship with NVIDIA moved from partnership to platform strategy. HPE Private Cloud AI and new AI factory systems built with NVIDIA brought Blackwell-based infrastructure, AI-ready storage, observability, and deployment services into one enterprise-focused architecture. Combined with Cray’s supercomputing DNA, Juniper’s networking depth, and the GreenLake hybrid-cloud model, HPE has turned itself from a post-split enterprise hardware company into a serious platform supplier for AI, sovereign cloud, and liquid-cooled high-performance computing.
1
Liquid Assets
Hewlett Packard Enterprise now sits where several infrastructure shifts are converging at once: AI, hybrid cloud, high-performance computing, networking, security, storage, and data control. Increasingly, the company is building the systems layer that lets enterprises, governments, telecom operators, and research institutions run demanding workloads across private data centers, edge sites, sovereign environments, and AI factories.
The business runs on two main engines: Cloud & AI and Networking. Cloud & AI spans servers, storage, HPE Cray supercomputing, AI systems, data analytics, GreenLake software and services, and HPE Financial Services. Networking, magnified by the Juniper acquisition, has grown into a larger strategic pillar reaching across campus and branch connectivity, data-center switching, routing, security, WAN automation, and AI networking. Together, the two segments broaden and deepen HPE’s role across the infrastructure layer: compute to run the workload, storage to feed it, networking to move data reliably, software to orchestrate the environment, and security to govern access across distributed sites.
As AI demand moves beyond the first wave of hyperscale training clusters, that integrated position shapes HPE’s outlook. Enterprises and governments now want to deploy AI closer to their own data, users, rules, and latency requirements – a shift that favors infrastructure capable of running in private and sovereign clouds, hybrid environments, regional sites, and edge locations. HPE’s AI backlog leans toward sovereign and enterprise customers, which plays directly to the company’s long-standing strengths in regulated, complex, and mission-critical IT.
Juniper also reshaped the AI opportunity itself. AI systems do not scale on servers alone – they need fast, secure, low-latency networks that move data across racks, data centers, regional sites, and edge locations without letting connectivity become the bottleneck. Juniper deepens HPE’s role in that layer through data-center switching, routing, coherent optics, WAN automation, security, and telco-grade networking. That makes Networks for AI more than an adjacent product line; it turns networking into part of the AI system itself.
This matters more as inference moves closer to users and data. HPE AI Grid with NVIDIA extends the company’s reach from centralized infrastructure into distributed AI environments, where telecom operators, service providers, retailers, manufacturers, hospitals, and public-sector customers need real-time performance across many locations. Here, HPE is not merely providing servers, but the fabric that ties AI factories, regional clusters, edge systems, and secure private environments into one operating model.
The same logic runs inside enterprise networks. HPE Mist and HPE Aruba Central give the company an “AI inside the product” angle, using agentic automation to detect, diagnose, and resolve network issues before they turn into downtime or help-desk volume. That is a practical growth driver, because customers do not upgrade networks for speed alone, but also to cut operating burden, improve reliability, enforce security policies, and support more distributed digital operations.
GreenLake and HPE Private Cloud carry the business from infrastructure supply into infrastructure control. GreenLake gives HPE a platform for managing hybrid environments at scale, while Morpheus adds orchestration, automation, migration, and self-service provisioning. GreenLake’s surrounding software layer adds migration, recovery, backup, governance, and data-management capabilities around HPE’s private-cloud platform. That is increasingly important as enterprises reassess virtualization costs, move workloads across private and hybrid environments, and prepare AI pipelines without giving up control over resilience or compliance.
Storage is where the AI story becomes more durable. Once models move into production, enterprises need to store, retrieve, secure, replicate, and govern the data feeding those systems across private-cloud, inference, and mission-critical environments. Alletra pushes HPE into that layer, with file-and-object storage for unstructured data and workloads where availability cannot be an afterthought. AI may start with compute, but it scales through data architecture.
Security and financing make the platform more usable for the customers HPE knows best. AI governance, Zero Trust controls, hybrid mesh firewalls, confidential computing, post-quantum readiness, and threat intelligence all matter for sovereign and regulated buyers, while HPE Financial Services helps reduce the upfront friction of large AI, networking, and private-cloud projects.
Competition remains intense, particularly from Dell, while HPE also faces focused pressure in individual layers of the stack. HPE is not the only winner in AI infrastructure, and order conversion will remain uneven because large AI projects are lumpy and supply-constrained. Still, its core customers – enterprises, governments, research institutions, and regulated industries – tend to value integration, control, support, and long-term operating consistency. That gives HPE a credible path to benefit from AI infrastructure growth without relying solely on commodity server share.
All in all, HPE’s opportunity extends well beyond selling boxes into an AI cycle: it is building toward an integrated infrastructure model for a world where compute, networking, storage, security, software, financing, and services increasingly operate as one system.
1
Cache Flow
HPE’s financial profile increasingly reflects a shift from recovery to strong expansion. Q2 FY26 highlighted that clearly, as the company achieved record revenue, gross margin, and adjusted EPS, as well as its highest-ever free cash flow generation for a second quarter, reflecting strong execution and healthy demand across the business.
HPE’s total FQ2 revenue jumped 40% year-over-year to $10.7 billion, above the high end of guidance and well ahead of consensus. But top-line growth doesn’t tell the whole story, as profitability soared even more. Non-GAAP diluted EPS more than doubled year-over-year to $0.79, surging past analyst estimates and extending its long track record of beating Wall Street consensus. Non-GAAP gross margin expanded 7.5% to 36.9%, and operating profit rose 132% to $1.4 billion. That margin expansion is important because AI server growth can easily become a low-margin volume story if pricing, mix, and component costs move the wrong way. In HPE’s case, Q2 showed stronger demand flowing through to profitability.
Cloud & AI is now the main revenue engine. The segment generated $7.7 billion of revenue, up 23% year-over-year, with operating profit reaching $954 million and operating margin rising to 12.4%, helped by stronger server demand, higher-value configurations, and improving scale. AI Systems orders reached $1.8 billion, cumulative AI Systems bookings rose to $16.4 billion, and AI Systems backlog stood at $5.9 billion. Including Networks for AI, broader AI backlog exceeded $6.3 billion.
The Storage results were more mixed: total segment revenue rose only 2%, but Alletra MP orders and revenue grew triple digits, signaling that the go-forward AI/private-cloud storage platform is stronger than the reported segment growth rate implies, while adding a strategically important growth line.
Networking has become the second major profit engine, with the segment’s revenue rising 148.2% from the prior-year period – although year-over-year comps are less clean post Juniper. Thus, Data Center Networking – primarily a Juniper-added capability – expanded sharply, while Routing grew from a marginal category to a $775 million revenue line. Juniper has also materially expanded HPE’s security portfolio and campus networking capabilities.
Cash flow shows that the business reset for the AI era is being matched by a financial reset. HPE generated $1.4 billion of operating cash flow and $915 million of free cash flow in FQ2, bringing first-half free cash flow to $1.6 billion, around 75% above the prior-year period. Free cash flow was below $1 billion in FY25, but now HPE guides for at least $3.5 billion in FY26 and at least $4.5 billion in FY27, having raised the outlook again after FQ2 results.
Guidance for other metrics was also significantly lifted. For Q3, HPE guides revenue to $11.5-12.1 billion, implying roughly 29% year-over-year growth at the midpoint, and non-GAAP EPS of $0.88-0.93, reflecting adjusted earnings growth of about 106%. For the full fiscal year, total revenue is expected to rise 29-33% on a reported basis – or at a high-teens rate on a normalized basis – while non-GAAP operating profit growth is seen at 80-85%, and non-GAAP EPS is expected to expand roughly 75% to $3.35-3.45. Notably, the updated FY26 outlook ranges for non-GAAP diluted EPS and free cash flow are higher than targets outlined for fiscal 2028 in October 2025.
This fast progress is not risk-free, as competition remains intense and margin expansion is not linear due to inventory and receivables timing, as well as R&D and other expenses necessary to support scale and market share. At the same time, HPE is witnessing the same memory cost inflation as all hardware providers, while supply constraints in DRAM, NAND, wafer capacity, and key networking components – also shared across the industry – can delay backlog conversion, especially for large AI systems that depend on coordinated component availability and deployment timing.
Still, HPE appears to be firmly on an upward trajectory. The company is growing faster, earning more on each dollar of revenue, converting more of that growth into cash, and pulling forward financial targets that were supposed to belong to a later phase of the plan. Just as important, the stronger numbers are not coming from one isolated pocket of demand. Cloud & AI, Networking, AI Systems, Alletra, GreenLake, and Juniper-related synergies are all contributing to a broader step-up in HPE’s earnings power.
1
Untapped Compute
HPE now sits across several infrastructure categories at once, which makes peer selection more about strategic overlap than a perfect match. Dell is the closest broad comparison, with similar exposure to enterprise servers, storage, AI infrastructure demand, and the shift from legacy hardware toward higher-value systems. Cisco provides the mature networking benchmark, especially after Juniper made networking a much larger part of HPE’s business, while Arista Networks frames the premium end of AI networking, with high exposure to data-center performance and low-latency connectivity. Despite the similarities, Cisco and Arista – both Smart Investor Portfolio holdings – are more focused networking leaders, while HPE combines networking with servers, storage, GreenLake, AI systems, services, and financing.
The breadth of these companies’ offerings along the AI value chain has played a major role in their stock performance over the past year, while starting valuation, growth rates, and backlog momentum help explain the gaps between them. Both Cisco and Arista delivered strong gains of more than 75%, but those returns still look more measured next to Dell’s 250%+ surge, driven by its close alignment with the AI infrastructure narrative and supported by orders and backlog expanding at an exceptional pace. Within this peer group, HPE came a strong second, rising more than 165% despite its smaller scale than Dell and its more diversified business mix, which generally points to broader, but less concentrated, growth drivers.
Yet despite that triple-digit advance, the market may still be treating HPE more like a traditional infrastructure vendor than a company whose earnings, cash flow, and AI exposure have changed materially over the last year. That distinction becomes important when looking at valuation.
The numbers show the gap clearly. HPE trades at about 14.4x forward earnings and 9.5x on forward EV/EBITDA, well below all peers in the group. Sales-based multiples tell a similar story, with HPE at 1.8x forward EV/Sales, only slightly above Dell’s 1.7x and far below Cisco’s and Arista’s multiples. That spread partly reflects HPE’s lower margin profile, but it also suggests the market is still assigning limited credit to the company’s faster growth, stronger cash generation, and larger post-Juniper networking exposure.
That is where the setup becomes even more interesting. HPE’s forward revenue growth of 18.3% and forward EBITDA growth of 20.6% are well above Cisco’s and not far below the higher-growth peer set, while its expected non-GAAP earnings growth is second only to Dell’s. Meanwhile, its forward non-GAAP PEG ratio of 0.49x is extremely low in absolute terms and much lower than peers, sending the clearest valuation signal and hinting at a “growth at a reasonable price” setup. In today’s AI-aligned infrastructure market, that is rare – growth is visible, but the multiple still looks restrained. That’s part of the reason why Wall Street sees an additional upside of nearly 40% for the stock.
Moreover, HPE’s stated financial framework commits to returning a significant portion of its free cash flow to shareholders over time via consistent dividend growth and opportunistic buybacks – with recent free cash flow growth giving the company ample liquidity to do so. With the quarterly dividend having been raised at a strong clip over the past decade, HPE’s dividend yield now stands at a solid 1.15%.
At the same time, HPE has a long-standing share repurchase program, expanded by $3 billion in fresh authorization in October 2025. Repurchases have slowed post-Juniper as HPE has prioritized reducing its net leverage back down to target, with total FY25 repurchases at just over $200 million, but this fiscal year saw a gradual return to its previous pace, with roughly $310 million in buybacks over the two latest quarters. Moreover, management outlined a clear capital-return path forward: once leverage reaches the 2x target, the company expects to return at least 75% of free cash flow to shareholders through dividends and repurchases.
HPE offers a rare mix in today’s AI infrastructure market: strong recent performance, visible growth, a still-discounted multiple, and a shareholder-return path that should strengthen as leverage normalizes.
1
Investing Takeaway
HPE is turning from a post-split enterprise infrastructure company into a broader platform for AI-era computing. Its advantage sits in the combination of servers, storage, networking, GreenLake, supercomputing, security, financing, and services – pieces that become more valuable as enterprises and governments move AI from pilots into controlled production environments. Juniper deepens the networking layer, Cray strengthens high-performance computing and liquid-cooled systems, and GreenLake gives HPE a management model for hybrid infrastructure. Risks remain, especially around competition, supply constraints, integration execution, and uneven AI order timing. Still, HPE now offers a stronger mix of growth, cash generation, backlog visibility, and valuation upside than the market seems to fully recognize.
1
New Sell: IBM (IBM)
We are selling IBM because the stock’s strongest upside drivers now look more likely to play out over a longer timeline than the Smart Investor Portfolio’s review window. The thesis remains intact: IBM is a high-quality enterprise tech giant with real AI and quantum optionality and a strong moat. However, portfolio-timing and opportunity-cost considerations have moved the needle for now.
The recent news flow has been supportive. IBM delivered a strong Q1 beat, with revenue rising 9% year-over-year to $15.9 billion and operating EPS increasing 19% to $1.91. Software guidance was raised, Red Hat continues to anchor the hybrid-cloud story, IBM Z posted strong growth, and the z17 mainframe gives the company a credible AI-inference angle close to mission-critical enterprise data. IBM also secured a major government quantum award, committed heavily to its quantum roadmap, expanded AI partnerships with Google Cloud and ServiceNow, and launched Project Lightwell to strengthen open-source security for enterprise AI systems.
This underscores the fact that the exit is not caused by any strategic or fundamental weakness. In fact, IBM may eventually become one of the most important long-term winners in enterprise AI and, further out, in quantum computing as well. The issue is that IBM is a sprawling company in the middle of a long restructuring around AI. Its software, consulting, infrastructure, mainframe, and Red Hat strategies are all connected and well-structured, but they will not turn the company into an agile AI winner overnight. We also believe that IBM’s quantum initiatives will become a major long-term path to value, but they are not likely to become a revenue growth catalyst for at least a couple of years.
The market already gave IBM a sharp AI/quantum re-rating before the proof fully arrived in the numbers. Since then, the stock has lagged as investors realized that the transformation will take time. Consulting remains a drag on sentiment, even though IBM’s consulting business has much stronger strategic ties to enterprise AI implementation than traditional consulting peers.
There is also a portfolio dimension. IBM now competes for the same enterprise AI slot as Oracle: databases, cloud infrastructure, hybrid systems, and large-organization modernization. ORCL is more volatile, but it currently offers a much clearer near-term AI growth signal through OCI, massive RPO, hyperscaler demand, and accelerating infrastructure buildout.
We are locking in gains and will continue watching the stock closely from the sidelines. IBM’s modest near-term growth expectations are reflected in its somewhat high forward PEG ratio. Meanwhile, the stock is relatively inexpensive on most other metrics – both versus tech sector medians and peers with comparable moats. This setup gives us enough latitude to step back in as soon as we see stronger evidence of the mainframe AI cycle, consulting acceleration, Red Hat/RHEL recovery, or the emergence of other catalysts.
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, EME, APH, IBKR, VRT, STRL, ASX, MTZ, PH, ORCL, GOOGL, CRWD, CSCO, KEYS, ATI, JBL, PANW, BNY, MS, NVT, CRDO, C, RTX, and JPM.
The first runner-up is now PM with a 18.10% gain since purchase. Will it break into the winners’ circle, or will another stock outrun it to the finish line?
1
|
