.png)
March 6, 2025
The Zero Trust Path - Zscaler and the Evolution of Enterprise Security
By
Kristal Investment Desk
We are in extreme fear. Chart shows CNNยดs greed/fear vs SPX.
Source: Macromicro
Volatility has arrived with a vengeance on Wall Street. A potent cocktail of tariff concerns, tech sector corrections, and dismal manufacturing numbers has pushed investor sentiment into what JPMorgan delicately terms "cyclical lows". The CNN Fear & Greed index stands in "extreme fear" territory, with semiconductors down 20-30 per cent despite accelerating revenues.
Yet this market sell-off reads more like a standard correction than a harbinger of prolonged economic distress. The Atlanta Fed's GDP forecasts are indeed tanking, but this stems primarily from companies front-loading imports to beat expected tariffs โ creating inventory buildups that obscure underlying economic resilience.
A closer look at corporate behavior suggests rational stockpiling rather than panic. Companies are filling warehouses now to avoid paying more later, effectively borrowing from future imports. This creates a temporary GDP headwind through lower net exports that should reverse once import patterns normalize.
The yield curve, traditionally a reliable recession indicator, offers a mixed signal. The 10-year Treasury yield has declined throughout 2025, reflecting expectations that slower deficit growth will reduce government debt issuance. This downward pressure comes alongside disinflationary forces from fiscal discipline, with TIPS pricing suggesting limited inflation concerns despite tariff headlines.
Meanwhile, tech valuations have contracted substantially. Semiconductor stocks crucial to the AI revolution are trading at multi-year valuation lows despite growing revenues. This sector rotation appears more reflexive than fundamental, with defensive sectors capturing flows as investors execute the classic flight-to-safety playbook.
Federal Reserve positioning is increasingly supportive. While markets anticipate 2-3 rate cuts, signals suggest a more aggressive cutting cycle is plausible. This should occur alongside the expected end of quantitative tightening this year, potentially providing a double liquidity boost.
For investors with longer time horizons, today's fears may represent tomorrow's opportunity. Similar market dislocations in 2015, 2018, 2020, and 2022 rewarded those who purchased quality assets during periods of heightened fear.
Current market conditions have pushed the S&P 500 less than 10% off all-time highs, with another 5-10% downside possible. This hardly constitutes a market catastrophe, particularly given the index's performance over recent years.
The AI narrative cycle offers a telling example of market myopia. In merely twelve months, conventional wisdom has swung wildly from "AI will commoditise software" to "hardware makers will capture all value" to "hardware will become commoditised" and now to "recession fears and geographic rotations".
Tariff concerns that dominate today's headlines will likely resolve more quickly than pessimists expect. The asymmetric impact of trade restrictions creates powerful incentives for diplomatic solutions, particularly with North American partners.
The wise investor neither throws in the towel nor goes all-in during such periods. They simply continue executing their process with discipline โ the financial equivalent of keeping calm and carrying on.
Chip designer's first-quarter profits jump as it rides AI wave with hyperscaler partnerships
Broadcom's strategic pivot towards artificial intelligence is bearing fruit, with the semiconductor giant reporting a 77 per cent year-on-year increase in AI revenue to $4.1bn in its first fiscal quarter of 2025.
Chief executive Hock Tan, whose acquisition strategy was once viewed primarily as financial engineering, is now seeing his seemingly disparate collection of assets coalesce into a powerful infrastructure layer for the AI era.
The company's approach diverges markedly from industry standard. While Nvidia pursues general-purpose AI chips, Broadcom has doubled down on custom silicon designed specifically for hyperscale customers and their frontier AI models.
"New frontier models and techniques put unexpected pressures on AI systems. It's difficult to serve all classes of models with a single system design," Mr Tan told analysts, explaining the company's tailored approach.
This strategy creates deep competitive moats. Once a hyperscaler integrates Broadcom's custom accelerators into their AI infrastructure, switching costs become prohibitive, transforming the relationship from transactional to symbiotic.
Broadcom currently has three hyperscale customers deploying its custom accelerators at scale, with four additional hyperscalers in "advanced development." The company projects this will translate to a serviceable addressable market of $60bn-$90bn by fiscal 2027 from the three existing customers alone.
The VMware acquisition, completed in early 2024, has undergone a remarkable transformation under Broadcom's stewardship. Operating margins have jumped to 70 per cent, up from less than 30 per cent pre-acquisition, while quarterly spending has been halved from $2.4bn to $1.2bn.
Beyond financial engineering, VMware is becoming an unexpected AI enabler for enterprises, addressing growing concerns about data sovereignty by allowing companies to run AI workloads on-premises while virtualizing GPUs.
Broadcom's networking expertise may prove its most decisive advantage as AI clusters grow exponentially. The company has already taped out its next-generation 100 terabit Tomahawk six switch, positioning it to capture an increasing share of infrastructure spending as AI workloads scale to million-node clusters.
For investors separating hype from sustainable advantage in the AI sector, Broadcom's approach offers an intriguing counterpoint to conventional wisdom. The company is betting that AI's future belongs not to general-purpose computing but to deeply customized, tightly integrated systems designed for specific architectures.
Whether this approach ultimately dominates remains uncertain, but Broadcom's journey from industry consolidator to AI enabler represents one of the most remarkable strategic pivots in semiconductor history.
Semiconductor maker's Q4 revenue beats forecasts yet shares slide nearly 20%. Marvell Technology has joined the exclusive club of chipmakers whose fortunes are being transformed by artificial intelligence. Yet a sharp post-earnings share price drop suggests investor expectations are running even higher than the company's surging sales.
The semiconductor group beat analysts' forecasts in its fourth-quarter results, with revenue rising 27 per cent year-on-year to $1.82bn. Non-GAAP earnings per share jumped 40 per cent sequentially to $0.60, exceeding estimates of $0.59.
Despite these robust numbers, Marvell's stock plunged nearly 20 per cent. The contrast with rival Broadcom, which reported just a day later, could not be starker. Broadcom shares rose 5 per cent after posting a 77 per cent increase in AI-related revenue to $4.1bn and a more confident outlook.
The divergent market reactions highlight the companies' fundamentally different strategic positions. Broadcom offers a comprehensive AI infrastructure stack โ custom silicon, networking solutions, and VMware's software layer โ creating an ecosystem that increases customer switching costs. Marvell, lacking this full-suite approach, appears more vulnerable to competition despite its technological prowess.
Chief executive Matt Murphy confirmed AI now constitutes more than half of Marvell's data center revenue, which itself represents 75 per cent of total sales. Yet without Broadcom's integrated software assets or Nvidia's dominant AI ecosystem, Marvell must compete primarily on silicon innovation alone.
While Broadcom emphasized its symbiotic partnerships with hyperscalers, Marvell faced pointed analyst questions about customer stickiness. Broadcom's vertically integrated approach offers clients a more complete solution, from chip design through to infrastructure software.
The VMware acquisition has proven particularly prescient for Broadcom, with operating margins reaching 70 per cent, up from less than 30 per cent pre-acquisition. This software component gives Broadcom recurring revenue and deeper customer relationships that Marvell cannot match.
For investors, the lesson appears clear: in the AI gold rush, those selling complete solutions are being valued more highly than those supplying individual components, however sophisticated they may be.
The cybersecurity darling's recovery from last summer's global outage remains a work in progress. Revenue beat estimates by 2.5 per cent last quarter, but forecast earnings of $3.39 per share for the coming year fell well short of $4.43 consensus. A higher tax rate explains $0.98 of this gap.
George Kurtz, founder-chief executive, deployed "Customer Commitment Packages" to retain clients after the July software error that paralysed airlines and hospitals. These concessions cleverly funnel customers towards broader platform adoption rather than mere discounting.
The strategy complements "Falcon Flex", CrowdStrike's frictionless module purchasing programme that has seen deal value surge 80 per cent quarter-on-quarter to $2.5bn. Flex customers use an average of nine modules versus just 21 per cent of the overall customer base using eight or more.
Meanwhile, cloud security grew 45 per cent year-on-year to $600m in annual recurring revenue, while security information management surged 115 per cent to $330m. State-sponsored cyber threats are rising sharply as generative AI democratises sophisticated attacks.
CrowdStrike's rehabilitation will likely not produce financial results resembling pre-outage patterns until the year's final quarter. Still, the company has reiterated ambitious targets: 30 per cent operating margins by fiscal 2029 and $10bn in annual recurring revenue by fiscal 2031.
For investors, the question is whether CrowdStrike's premium valuation reflects the lingering costs of rebuilding trust in a domain where your own reliability is as critical as the threats you defend against.
Never trust, always verify. That is the cyber security mantra propelling Zscaler's growth as it supplants legacy firewalls with its cloud-based zero trust approach. The company's latest quarterly results suggest its strategy is paying off โ both technically and financially.
Zscaler's approach is as merciless as it sounds. Rather than granting users blanket access once they enter a network, its Zero Trust Everywhere platform verifies every connection attempt individually. This architecture shrinks the attack surface and limits the lateral movement of threats within networks. The concept is not new, but Zscaler's execution has been exemplary.
The company's numbers tell a compelling story. Revenue rose 23 per cent year-on-year, beating estimates by a comfortable margin. Billings โ a key indicator of future revenue โ also outperformed, despite macroeconomic headwinds that have squeezed IT budgets. Perhaps most encouraging was the first quarter-on-quarter rise in net revenue retention in over two years, suggesting existing customers are spending more.
Valued at about 45 times free cash flow and 10 times EV to Sales, Zscaler trades at a premium to the broader software sector. Yet the multiple looks less extravagant when considering its operating leverage and growth trajectory. Operating expenses rose just 19 per cent year-on-year โ slower than revenue growth โ despite heavy investment in product development.
The question is whether the company can maintain momentum in an increasingly crowded security market. Palo Alto Networks, Cloudflare and others are pushing similar zero trust architectures. But Zscaler has been methodically expanding beyond its core products. Its branch security and cloud workload offerings are gaining traction, as demonstrated by a recent Fortune 50 energy company win.
Artificial intelligence presents both challenge and opportunity. While AI could eventually reduce the number of human users requiring licenses, it will also generate more network traffic to secure. Already, the company has seen its AI analytics suite double year-on-year and has raised pricing on its AI assistant.
With $2.9bn in cash against $1.15bn in debt, Zscaler has ample resources to continue innovating and acquiring. Its overhauled go-to-market strategy is bearing fruit, with improving sales productivity and stronger partnerships with global system integrators.
For investors, the message is clear: in cyber security, zero trust does not mean zero confidence. On the contrary โ trust in Zscaler appears well-placed.
Swiss shoemaker stretches its stride but valuation demands marathon-level performance. When Roger Federer invested in On in 2019, the Swiss running shoe start-up was little known outside its home market. Today, the Zurich-based group's cloud-cushioned footwear has propelled it into the fast lane of the global sportswear market.
On's fourth-quarter results sprinted past expectations, with revenue beating estimates by 2 per cent and earnings per share of $0.33 eclipsing the expected $0.18. The performance validates its premium positioning strategy and suggests it is taking meaningful strides against industry heavyweights Nike and Adidas.
The direct-to-consumer channel now represents nearly 49 per cent of sales, up 3 percentage points year-on-year. This pivot towards higher-margin channels helped lift gross margins to 61.6 per cent, 50 basis points above estimates. The figures demonstrate that On's premium positioning resonates with consumers willing to pay CHF 170 for its distinctive cloud-soled Cloudmonster shoes.
On's growth story extends beyond footwear. Apparel crossed the CHF 100m annualised revenue threshold, with 67 per cent year-on-year growth in direct channels. The company's physical retail expansion โ 19 new stores in 2024, bringing the total to 50 โ serves as brand billboards and supports this diversification.
Geographic expansion provides another lane for growth. Brazil revenue doubled year-on-year, reflecting the potential in emerging markets. Meanwhile, the company expects China to reach 10 per cent of sales by 2026. On's focus on market-specific products and strategic partnerships exemplifies its localised approach to globalisation.
For 2025, management projects 27 per cent foreign exchange-neutral growth and a 17.3 per cent EBITDA margin. The company is running ahead of targets set at its 2023 investor day.
Yet On's forward price-earnings ratio of 39 times demands Olympic-level performance. This valuation positions it more as a luxury goods purveyor than a traditional athletic footwear manufacturer. Investors are betting that On's premium positioning โ "at the intersection of performance and cultural relevance," as co-founder David Allemann puts it โ will continue generating outperformance.
Distribution challenges at its Atlanta warehouse, though being addressed, highlight potential growing pains. The company expects normalisation by mid-2025, but such operational issues could trip up rapid expansion plans.
For investors considering the stock, the question is whether On can maintain its premium positioning while achieving mass market scale. The company's focus on brand-building through high-profile partnerships and cultural moments shows promise, but sustaining this momentum requires continued innovation and marketing prowess.
On's current pace suggests it has legs. But at its premium valuation, investors should recognise they're paying marathon prices for a company that must avoid even the slightest stumble.
The Silicon Valley orthodoxy that businesses should "do one thing and do it well" has encountered a formidable challenge from Southeast Asia. Sea Limited, once a gaming company with fledgling e-commerce and financial ambitions, has proved the sceptics wrong.
Its Q4 results demonstrated not just growth โ revenue up 36.9 per cent year-on-year to $4.95bn โ but a decisive shift to profitability. Adjusted ebitda reached $591m, up 365 per cent, with margins expanding from 3.5 per cent to 11.9 per cent. The market has rewarded this performance: Sea's shares have risen 46 per cent this year, giving it a market capitalisation of $55bn.
What makes Sea noteworthy is its vertically integrated approach. While western tech groups tend to specialise within the value chain, Sea controls the entire customer experience across gaming, e-commerce and finance. This full-stack model creates structural advantages that specialised competitors struggle to replicate.
Consider SeaMoney's surprisingly robust loan book. Its 64 per cent year-on-year expansion would typically ring alarm bells, yet non-performing loans remain at just 1.2 per cent. The secret? Sea's ability to draw on customer data from Shopee purchases and Garena gaming patterns for superior risk assessment.
Cross-promotion provides another edge. When customer acquisition costs consume significant portions of revenue for standalone digital businesses, Sea amortises these expenses across multiple divisions. Garena games promote Shopee sales; Shopee offers incentives for SeaMoney payments.
The greatest transformation is Shopee itself. Once a cash-consuming growth experiment, it now generates $3.66bn quarterly โ seven times Garena's revenue โ and $152m in adjusted ebitda, a stark reversal from its $225m loss a year ago. Even in Brazil, where incumbent MercadoLibre dominates, Shopee has achieved profitability just five years after launch.
Gaming remains Sea's most profitable segment with 56 per cent ebitda margins, providing stable cash flow despite modest 1.6 per cent growth. This highlights another advantage of diversification: while standalone gaming companies live and die by hit titles, Sea's model provides resilience against the industry's inherent cyclicality.
Questions remain about competitive pressures, particularly from Alibaba-backed Lazada and GoTo in Southeast Asia. But Sea's Q4 results suggest its integrated model, once dismissed as unfocused, has created a formidable moat. The company that began as a gaming outfit with ambitious side projects has become something much more interesting: a blueprint for digital ecosystem building that challenges conventional wisdom about corporate focus.
The lesson? In emerging digital economies, controlling the full stack may be more valuable than perfecting a single layer. Investors who dismissed Sea's multi-pronged ambitions may wish they had taken a more integrated view.
ยท Alphabet: CFO Ashkenazi defends $75bn CapEx as meeting immediate demand, not speculation. AI features monetise at legacy-search rates, with "AI Mode" leveraging Gemini 2's capabilities against rivals while exploiting Google's multi-billion user advantage.
ยท PayPal: CEO Chriss claims checkout modernisation has finally caught rivals, lifting conversion 1-4 points. New Commerce API exposes 80m customer profiles to merchants for personalisation. Legacy system overhaul progresses slowly but steadily toward 80% global penetration by 2027.
ยท Meta: Meta AI reaches 700m monthly users, strongest on WhatsApp. Llama 4 will compress larger models' capabilities into smaller packages with native voice interaction. Cox celebrates Meta's open-source approach, allowing it to incorporate industry advances without relying on training monetisation.
ยท Uber: CEO Khosrowshahi reveals growing confidence in a fragmented autonomous vehicle market with "dozens of vendors and thousands of fleet operators" thanks to AI cost deflation and synthetic data generation. This scenario benefits Uber's marketplace model by increasing network dependency. The company plans to invest "a small percentage" of free cash flow to accelerate AV adoption through direct vehicle ownership and utilization guarantees to partners. Dara summarized business performance as exceeding investor day targets "across the whole thing," citing outperformance in membership, non-tier 1 cities, teen usage, and European markets.
Disclaimer: This musing is for informational purposes only and should not be considered as investment advice.
โ
By
Kristal Advisors
March 8, 2025
Liked it?
Share it with your friends & colleagues!
March 6, 2025
By
Kristal Investment Desk
A fully digital onboarding process that can be completed within 15 minutes.
No more voluminous paperwork and queuing!
I understand the financial products and would want to proceed with investing without a financial guide
Proceed