Tech Fund

Having done tons of coding, cloud engineering, as well as covering the software sector as an equity analyst over the last 15+ years, I'm really not seeing a software disruption here. Jensen is focused on boosting the productivity of his workforce, which can easily result into an economic value of >$100 billion for his company, as opposed to doing a set of highly risky and complicated software transitions that might save him $500 million. And other enterprises are doing the same going through recent earnings calls. All the sudden experts in software underestimate gravely how complicated it is to rip out and replace an ERP, database, CAD, supply chain management, or other core tool that runs the backbone of a business. These are complicated 2-4 year projects with mixed results of success, frequently resulting into earnings warnings and failed projects. The real value currently is to insert as much AI into your business as possible and boost productivity. And the easiest way to do is having AI leverage your core tools. We're currently using all our software tools more than ever before, as now our AIs are also using them as well. So, we're spending more on AI and software and this continues to accelerate. Where's the money coming from? We're spending less on human services as we can do more ourselves now. So, we'd be bearish on companies that provide services such as consulting, IT services, freelancing etc. However, for key software tools, revenues and EPS will accelerate from here in our view, as these names are typically building AI agents into their current tooling. From everything we're seeing, enterprises are very keen to lever these AI capabilities. It's not like enterprises - and especially SMBs - have a deep pool of AI talent to build this themselves. Free open source software also has been available for decades, illustrating that price is not the key factor in selecting a tool. It's about security, reliability, and network effects around software tools. techinvestments.io/p/the-bull-cas…

Merchant InP laser market shares (Cowen est) "Three of the four major suppliers have described strategies to increase laser manufacturing capacity, the outlier being Broadcom, which has not explicitly commented on capacity expansion in a public forum, but which we believe is likely expanding capacity as well. Sumitomo Electric, which we estimate had ~22% of the merchant InP market in 2024, plans to increase InP production capacity by 2.4x from 2023 to 2028. More qualitatively, Mitsubishi Electric (~18% share) has described "partially shifting" its ~$1.5B in annual capital investment from power devices to optical devices. Finally, Lumentum (~30% share) has described increasing capacity by 40% from September 2025 to June 2026."

As Moore’s Law has slowed, the obvious way to pack more transistors into a single computing system is to simply cobble more dies together. Therefore, it's easy to see how die bonding will remain one of the most interesting areas in semicap. Our outlook on the die bonding space and its key players: techinvestments.io/p/hbmdie-bondi…

$ASML EUV layers per node (Goldman)

Goldman - token usage is exploding due to AI agents

Short/avoid anything related to client compute

Our outlook on the HBM & DRAM market, we take a contrarian view: techinvestments.io/p/hbm-and-dram…

The market is currently assuming the worst of both worlds for software, i.e. revenue growth will be impacted, or decline, while software companies themselves will have no opportunity to cut costs. However, $SNOW already cut 200 employees last Q due to AI benefits. At the same time, revenue growth is accelerating. Our view continues to be that the revenue outlook in the coming years is strong for quality software names, and that AI productivity gains will cause margin expansion, accelerating EPS growth. We see the current market environment as a good old panic, largely driven by panic selling and with limited understanding just how resilient quality software is. Free open source alternatives have been available for decades for a lot these names. Our guess is that there will be plenty of AI winners in the software space, while some will be disrupted, and that there will be generational buying opportunities in this current market.

Expert call with Software/Cloud consultant (via Mizuho) Key findings: $SNOW losing share vs Databricks and Fabric, Cybersecurity performing strongly - $PANW $CRWD $ZS, Cloud AI workloads accelerating - $AMZN, $MSFT "Our expert noted that cloud activity levels for general-purpose IaaS and PaaS consumption have been growing at a solid pace, while AI‐related workloads have expanded at a meaningfully faster rate. He highlighted that the rising on-prem infrastructure costs and extended hardware lead times are increasingly pushing customers toward cloud platforms as the fastest way to deploy new workloads. Our expert indicated that Azure and AWS (AMZN, Outperform, Covered by Lloyd Walmsley) have benefited most from this dynamic, with Google Cloud (GOOGL, Outperform, Covered by Lloyd Walmsley) also capturing a respectable share. From a data platform perspective, he cited a deceleration in SNOW, in favor of Databricks and MSFT Fabric. That said, he emphasized that SNOW remains the market leader in governed enterprise analytics, with a large installed base and continued relevance as AI adoption increases the importance of trusted data. Our expert once again noted that demand for his cybersecurity practice remains robust (with what he believes has "no end in sight"), and this in part is being driven by rapidly growing AI threats, heightened identity-related concerns, and broader data security concerns. And to this end, he highlighted strength in AI security, identity resilience and recovery, DLP and DSPM. He also noted good customer demand for cloud security (albeit to a somewhat lesser degree), along with API and code security. Meanwhile, some of the more traditional areas of cybersecurity, such as vulnerability management (VM), physical firewalls, and DDoS protection, continue to see steady demand, but they aren't commanding as much of the incremental budget that some of the aforementioned "emerging" areas are. Conversely, he was less bullish on hardware-centric firewalls, which he noted continue to face increasing pressures from other security areas, such as SSE and SD-WAN. On a related note, last week we published our CIO and CISO security survey report . Encouragingly, a strong 77% of cybersecurity buyers in our survey characterized their spending activity over the last 3 months as at least "somewhat" more than usual, and this includes 24% with spending activity that was "much more" than usual. Meanwhile, the security buyers interestingly labeled network security (received 55% of the votes) and AI security (53%) as the most important areas of cybersecurity. Cloud, Endpoint and Identity were also highly prioritized areas. Regarding anticipated growth in cybersecurity budgets for 2026, we calculated a blended average of 15%. This reflects essentially no expected deceleration, and is also in line with the Street’s mid-teens median revenue growth expectation for public cybersecurity vendors in CY26E. Our expert indicated that observability demand is healthy, and competitive dynamics amongst vendors remain largely similar quarter over quarter. He also stated that he is seeing a broader trend of ongoing consolidation. Further, he cited blurring lines between security and observability. On that note, he views PANW's recent acquisition of Chronosphere as a smart buy bolstered by growing interest in open-source data collection. He also cited the criticality of data collection in a competitive sense for observability, and we note this was also a significant focus during DDOG's investor day last Thursday ."

$AMZN - how we get to a 25% IRR in the coming years: techinvestments.io/p/the-bull-cas…

@above_consensus I think it's way oversold here, very impressed with all new AI capabilities they're adding. Can easily become a multibagger in the coming years.

@techfund1 Had the same view. You still seeing it this way?

$KVYO is a typical piece of business software that's just executing extremely well. It reminds me a lot of $HUBS in '16, which turned out to be a 10-bagger. And on a lot of metrics, the company is actually showing better potential than HUBS. One of the most attractive features is its pricing model, with revenues scaling with the amount of data a customer is processing over its platform. Attractive risk-reward imo. techfund.one/p/klaviyo-a-bi…

We're not far away from creating movies/series with a prompt and with consistent characters. This likely means that content on Youtube is going to explode, as independent creators can monetize their content via ads on Youtube. So, big positive for $GOOGL However, $NFLX will likely sell less subscriptions - if there's plenty of good and tailored niche content on Youtube, why buy a Netflix subscription? The scenes that Seedance 2.0 is generating are pretty impressive - imagine what it will be 1-2 years from now. youtube.com/watch?v=23qpkG…

Power Semis in the AI Data Center don't get much attention - however, growth is exploding here. We review the key players: techinvestments.io/p/power-semis-…

$MU: "HBM4 shipments have started. Capacity is ramping well. HBM4 delivers over 11 gigabits per second speeds."

This week's earnings calls show huge productivity improvements & margin expansion with AI: $HOOD: Robinhood's internal AI deployment is generating substantial operating leverage. The company now resolves over 75% of customer support cases with AI, including complex cases previously requiring licensed brokerage professionals. In software engineering, where Robinhood aims to be "the best in the world," AI optimization across the entire development pipeline delivered an estimated nine-figure cost savings in 2025 alone. "This is a big reason why we've been able to drive such high product velocity, while keeping our costs down," Tenev explained. The company is now pushing every function to integrate AI with the goal of achieving best-in-class performance across all operations. For 2026, management expects even larger benefits as automation extends deeper into code review, deployment, and testing. $SPOT: Spotify is experiencing dramatic productivity improvements from AI coding tools. Söderström revealed a system called "Honk" where engineers can ask Claude on Slack during their morning commute to fix bugs or add features to the iOS app, receive a testable version, and merge to production before arriving at the office. "We've been told by key AI partners that our work here is industry-leading," he said. The pace of change accelerated dramatically in December 2025. "Christmas this year was an event, a singular event in terms of AI productivity," Söderström noted. "When I speak to my most senior engineers, the best developers we had, they actually say that they haven't written a single line of code since December. They actually only generate code and supervise it." The implication is that software companies like Spotify will produce "enormously more amount of software" as friction decreases. $KVYO: For the full year, operating margin hit 14%, up 170 basis points. Free cash flow surged 61% year-over-year to $87 million in Q4, with the company surpassing $1 billion in cash on hand for the first time. The margin expansion isn't just scale—it's AI-driven internal efficiency. Bialecki described building features 5-10x faster than in Klaviyo's early days, with entire prototypes now built "in an evening that used to take us weeks." More significantly, non-engineers are now building internal systems using AI coding agents, effectively expanding the engineering workforce without headcount growth. R&D expense growth slowed notably, prompting analyst questions. Bialecki characterized the company as benchmarking against "what the best AI companies are doing" in terms of efficiency. "We're able to go much faster and not 20% faster. I'm talking 5, 10, 20x faster than the early days," he said.

@komsit37 Yes, but AI agents using the tools can be monetized

@techfund1 > We're spending less on human services as we can do more ourselves now. Isn’t this the bearish argument for SASS? At least for companies with per seat pricing model. Less seats.

Having done tons of coding, cloud engineering, as well as covering the software sector as an equity analyst over the last 15+ years, I'm really not seeing a software disruption here. Jensen is focused on boosting the productivity of his workforce, which can easily result into an economic value of >$100 billion for his company, as opposed to doing a set of highly risky and complicated software transitions that might save him $500 million. And other enterprises are doing the same going through recent earnings calls. All the sudden experts in software underestimate gravely how complicated it is to rip out and replace an ERP, database, CAD, supply chain management, or other core tool that runs the backbone of a business. These are complicated 2-4 year projects with mixed results of success, frequently resulting into earnings warnings and failed projects. The real value currently is to insert as much AI into your business as possible and boost productivity. And the easiest way to do is having AI leverage your core tools. We're currently using all our software tools more than ever before, as now our AIs are also using them as well. So, we're spending more on AI and software and this continues to accelerate. Where's the money coming from? We're spending less on human services as we can do more ourselves now. So, we'd be bearish on companies that provide services such as consulting, IT services, freelancing etc. However, for key software tools, revenues and EPS will accelerate from here in our view, as these names are typically building AI agents into their current tooling. From everything we're seeing, enterprises are very keen to lever these AI capabilities. It's not like enterprises - and especially SMBs - have a deep pool of AI talent to build this themselves. Free open source software also has been available for decades, illustrating that price is not the key factor in selecting a tool. It's about security, reliability, and network effects around software tools. techinvestments.io/p/the-bull-cas…

@jweil130 Yes, I think new customer growth will be slower, as every company is focused on levering AI right now. But, I do think that the good software tools will see a tailwind from building AI automation in their platforms.

@techfund1 Ok fair enough but what about new business customers and startups etc. that’s the growth part. No one saying software cos are going away. Just paying a lower multiple for them. Make sense?