AI has taken the world by storm in recent months as the technology, which was largely a work in progress, hit the masses thanks to the ChatGPT breakthrough.
If the AI frenzy is currently mainly about the first step in machine learning (training), the second step (inference), is slowly but surely becoming the key for AI’s mass adoption and is likely to become another massive growth driver for the Tech industry in the next couple of years as most organizations roll out AI services and apps.
AI is indeed the underlying and necessary base for many new technologies including autonomous driving, robotics or the Internet of Things and the increasing complexity of training deep learning models is having a profound impact on the cloud computing architecture.
Hyperscalers are notably redesigning their cloud infrastructures with data transfer speed and energy consumption optimization in mind.
Longer term, quantum computing emerges as the next step in high-performance computing.
As AI is expected to open up new revenue opportunities and bring major efficiencies, it should gradually become pervasive and provide a material boost to GDP over coming years.
A multi-year investment cycle is then in sight to train AI and deliver AI services to customers at scale, an obvious positive for semiconductors, software and data center providers.
Industrial and consumer goods companies should also have the opportunity to leverage their installed base of IoT products and devices (cars, robots, industrial machines…) to offer a wide range of connected digital services (self driving, remote monitoring and maintenance…) and over-the-air options through subscription-based models.
For years, AI infrastructure was simple: GPUs ruled, and everything else followed. CPUs, memory, and networking existed, but as background plumbing, not strategy. That perception is changing. A structural shift is bringing CPUs back to the center. Not as legacy hardware from a pre-AI era, but as the foundation for the next wave: reinforcement learning…
Memory stocks such as SK Hynix, Micron Technology and SanDisk—among the best performers in the AI space over the past 12 months—came under pressure yesterday following reports that Google is preparing to roll out its TurboQuant data compression algorithm for AI inference. TurboQuant enables more efficient storage of data in the key-value (KV) cache—the memory…
OpenClaw, an AI agent builder, has surged in popularity in China thanks to its ability to autonomously perform tasks on behalf of users—organizing and responding to emails, drafting work reports, and preparing slide decks. Its versatility has prompted many Chinese tech companies to offer OpenClaw services on their cloud platforms, while local governments and enterprises…
For decades, copper carried the internet on its back. It threaded through server racks, snapped into Ethernet ports, linked chips to switches, and did it all with quiet dependability. Electrons racing through metal powered backplanes, patch cords, and direct-attach cables. Copper was cheap, familiar, and fast enough. Then generative AI changed the definition of “fast…
In a report published 18 months ago, we warned that “AI challenges were emerging and posing fundamental risks to entire segments of the software industry, suggesting that one would need to be very careful when investing in the sector.” Since then, we have remained underweight enterprise software—and that stance has paid off, with our AI…
SpaceX has just announced its blockbuster merger with xAI. While Elon Musk publicly frames the move as part of his long-standing ambition to take data centers to space, there is also a clear and pragmatic financial rationale. xAI has been cash-hungry since inception and is likely to remain so for the foreseeable future, whereas SpaceX is reportedly…
As we expected, the concerns that emerged in the final months of 2025 and weighed on the AI trade are rapidly fading. Early earnings reports this season have shifted the focus back to what matters most: impressive growth and expanding margins. And worth noting, the equity rally is broadening. While our core AI portfolio is…
In recent weeks, Nvidia has announced a series of investments with a clear common thread: AI inference, i.e. the running of AI applications. The company signed a $20 billion technology licensing deal with Groq, a specialist in inference chips; took a stake in Baseten, a fast-growing startup providing inference infrastructure for AI applications; and, more surprisingly, acquired a 2.9%…
Bloomberg reported yesterday that Meta and EssilorLuxottica are considering doubling annual production capacity of their Meta Ray-Ban smart glasses to 20 million units by year-end, up from 10 million currently, with the option to scale further to 30 million units if demand warrants. This move, which points to strong demand for AI-powered wearables, follows Meta’s…
For the first time in years, memory has become interesting again (see here and here). Once the most predictable component in a computer’s bill of materials, it has quietly turned into a global bottleneck and artificial intelligence sits at the center of the shift. The disruption begins with high-bandwidth memory, or HBM, a premium version…
Marvell’s $3.25 billion buyout of Celestial AI isn’t just another semiconductor acquisition. It’s a wager—big, loud, and physics-driven—on the idea that the future of artificial intelligence won’t run on copper wires at all. To bankroll the bet, Marvell unloaded its automotive Ethernet division to Infineon for $2.5 billion, then pivoted hard into the fast-forming world…
Meta’s reported spending cuts in its Metaverse division—reportedly as high as 30%—combined with the high-profile hire of Apple’s former head of design, Alan Dye, and its recent push into AI glasses, suggest that the company is rapidly reallocating resources away from VR (Quest headsets) and virtual worlds (Horizon Worlds) toward AI-powered wearables. This pivot should…
The recent correction in AI equities increasingly looks like the two major sell-offs of the past 18 months (February–March this year and July-August 2024), suggesting that we are more likely in a healthy profit-taking phase than in the early stages of a bear cycle. The severity, speed, and magnitude are strikingly similar, with AI and…
Recent capex guidance from hyperscalers has once again been impressive, underscoring the sustained momentum in AI data center buildouts. However, the next phase of this investment cycle will hinge increasingly on monetization — turning AI capabilities into scalable, revenue-generating products. In the past few days, both OpenAI and Microsoft have taken significant steps in that…
Amazon is racing toward a future shaped by automation, but it has not forgotten the humans still on its payroll. Its latest experiment, a pair of smart glasses for delivery drivers, shows how the company aims to boost efficiency without replacing every worker with a machine. The glasses automatically scan packages and provide turn by…
Data centers are evolving at breakneck speed. Artificial intelligence (AI) and high-performance computing (HPC) are pushing power demands to new heights, with next generation AI clusters requiring nearly a megawatt per rack. Traditional alternating current (AC) distribution systems, built for lower loads, are buckling under these demands. Multiple power conversions, bulky cabling, and energy losses…
Not so long ago, making a satellite phone call from a remote corner of the planet was an ordeal. You had to lug around a brick-sized device, angle it precisely toward the sky, and pray for a faint signal. Even then, the connection was often scratchy, and the bills were exorbitant, turning a simple conversation…
Microsoft announced yesterday that it has been testing a new chip cooling technology based on microfluidics, which it says removes heat up to three times more effectively than today’s data center liquid cooling standard, cold plates that are placed on top of chips. The approach channels liquid coolant directly into the processor through microscopic, vein-like…
Yesterday, Nvidia announced a $5 billion investment in Intel, alongside a joint development agreement that will allow both companies to break into markets they previously couldn’t reach. In the PC segment, Nvidia dominates discrete GPUs for workstations and gaming, but laptops have always been out of reach. With roughly 150 million units sold annually, this…
Nebius, which rents out AI infrastructure, has announced a landmark $17.4 billion agreement with Microsoft over five years, potentially ramping to $19.4 billion. Under the deal, Nebius will deliver dedicated capacity to Microsoft from its new New Jersey data center. This represents a transformational win for Nebius, a significant position in both our AI and…
We have long argued that SpaceX is on track to disrupt the global telecom market. Yesterday, the company took a decisive step in that direction with the announced $17B acquisition of EchoStar’s 50 MHz S-band spectrum in the US along with its global spectrum licenses. SpaceX’s Starlink initially focused on providing broadband internet to underserved…
The software sector has remained a minefield for investors this year, in line with our scenario (see our June 2024 report). So far, only two clear winners of the AI revolution have emerged—Microsoft and Palantir, both core holdings in our portfolios—while many other software vendors face the risk of being disrupted by AI. That said,…
After the latest tariff twists and chips export restrictions to China, initial earnings in the AI space were widely expected. They didn’t disappoint, with both Taiwan Semi and Siemens Energy releasing strong quarterly figures and either maintaining or exceeding full-year expectations. Taiwan Semi, the leading maker of chips whose growth has been buoyed in recent…
It’s been a turbulent few weeks, with tariff uncertainty fueling broader concerns about the macroeconomic outlook and, specifically, about the AI spending environment. Even if semiconductors were initially expected to be spared from tariffs, the prospect of an economic downturn affecting the core businesses of major hyperscalers—such as declining advertising revenues at Google and Meta,…
At last week’s GTC conference, Nvidia provided a glimpse of its upcoming AI chips, including the Rubin Ultra GPU, set to launch in late 2026. This next-generation AI accelerator is expected to consume up to 1.8kW per unit and 600kW per rack, a significant jump from previous models. For comparison, the current GB200 draws 1.2kW…
In recent days, quantum computing has seen significant progress, with Microsoft and D-Wave announcing groundbreaking developments. As detailed in a previous report, a major hurdle in quantum computing has always been the decay of qubits, which leads to coherence loss and computational errors. Indeed, creating qubits involves producing particles that can be electronically controlled, which…
In recent days, SpaceX’s Starlink announced partnerships with two of India’s leading telecom giants, Airtel and Jio, to provide broadband internet services to even the most remote and rural regions of the country. This marks a significant step as Starlink seems poised to finally break into the Indian internet market, the second largest in the…
Alibaba, China’s leading cloud computing platform, impressed investors with a sharp acceleration in cloud revenue growth to +13% year-on-year, up from 7% in the previous quarter and 3% above expectations, driven by triple-digit AI growth. The real standout, however, was Alibaba’s $4.3 billion quarterly capex, marking a 300% year-on-year surge and 80% quarter-on-quarter growth. The…
Palantir, the largest position in both our AI and Cybersecurity portfolios, impressed again last night with revenue growth accelerating for the sixth quarter in a row to 36% (from 30%) and a FY25 revenue growth guidance of 31%, both figures coming 5-10% ahead of expectations. The strong top-line performance sparked another significant margin leverage, with…
The initial panic triggered by DeepShitSeek is starting to subside as more information comes to light. Recent reports indicate that the true cost of their training process was closer to $500 million, not the $6 million initially claimed. Additionally, it appears that DeepSeek’s LLM performance isn’t as groundbreaking as first suggested, with revelations that the…
The rise of the Chinese AI assistant DeepSeek to the top of the iPhone download charts has sent shockwaves through global markets—and, more specifically, the entire AI value chain. The app’s AI model is widely regarded as highly competitive with OpenAI’s ChatGPT and other leading models, all while being significantly less expensive to train and…
Bloomberg reported earlier this week that Meta, which successfully launched at the end of 2023 the Meta Ray-Ban smart glasses, is working on a new generation of wearable devices that are scheduled to ship this year. That includes smart Oakley glasses aimed at athletes and high-end glasses – code-named Hypernova – whose design is closer to the…
As promised during his campaign, President Donald Trump declared a national energy emergency following his inauguration and signed several executive orders aimed at boosting domestic fossil fuel production while reversing many of the clean energy policies from the Biden era. Trump, along with key cabinet members such as Doug Burgum, the newly appointed Secretary of…
Synapse Invest was founded in 2019 with the mission of offering a diverse range of thematic investment products centered around innovation. Our journey began with the launch of our first Actively Managed Certificates (AMCs) in the closing weeks of 2019, and we are now proud to celebrate more than five years of consistent growth and…
Debates about whether we are the peak of the AI infrastructure cycle or close to the peak are popping up on a regular basis as investors are trying to figure out when the inevitable growth slowdown for companies exposed to it (chips, networking) will happen. Broadcom’s latest earnings and recent press reports about xAI spending…
In contrast to classical computers that rely on binary bits and which can have only two states (zero or one), quantum computers rely on the ability of quantum bits, or qubits, to exist in a superposition of both states at the same time. This oddity of quantum physics allows quantum computers to perform certain calculations…
While most software companies, from the small ones (C3.ai, MongoDB…) to the large ones (Adobe, ServiceNow, Salesforce…), are now wannabe AI companies, Palantir is the real deal in our view. The company, known for its software products used by governments/military around the world, has developed over the years an AI platform (AIP) that is quickly…
Even if there’s been a lot of excitement around AI in the last two years, there’s been a very limited number of pure play IPOs given that the market is dominated by a handful of players in computing (Microsoft, Amazon…), models/software (OpenAI, Anthropic…) and computing hardware (Nvidia, AMD). News that Cerebras, a US-based AI chip…
On Friday, US electric utility Constellation announced the restart of its Crane nuclear plant, that had been shut down in 2019 for economic reasons, along with a 20-year power supply agreement with Microsoft. The Tech giant will then secure 835 MW of power over 20 years to supply its data centers in the PJM Interconnection,…
Earlier this year, Google unveiled an eye-popping text-to-video games model called Genie that allowed to build basic playable games in 2D based on a short description. Six months on, AI in video gaming has made significant advances with gaming giants Roblox and Tencent presenting in recent days their own Generative AI technology that now allows to…
Despite recent concerns around a softening macro, AI spending fatigue and Nvidia’s Blackwell slight delay, the second part of the year and early 2025 are actually shaping up well for the Tech industry. Oracle’s just released FQ1 numbers were highly encouraging and suggested that AI training in data centers remains buoyant, with the company claiming…
Despite all the recent tumult in Tech equities recently, optical networking stocks such as Fabrinet and Coherent are quietly nearing or breaking their highs as earnings confirm that a new spending cycle on optical gear from data centers is on track. Last night, Fabrinet reported Q4 revenue growth of 15%, 3% above expectations. If these…
With AI electricity consumption surging and utilities starting to have difficulties meeting demand, both cloud giants and utilities are actively looking at solutions to tackle the power challenge. Deploying new solar/wind capacity and building nuclear reactors (or reopening idle ones) are clearly part of the agenda. But these initiatives also represent a multi-year effort, suggesting…
Moore’s law states that the density of transistors in an integrated circuit (IC) doubles about every two years. This empirical law has allowed the manufacturing of cheaper, faster, smaller and more efficient electronic devices for decades. But the pace of this trend is clearly slowing down and the techniques to miniaturize transistors are getting exponentially…
Nokia just announced the acquisition of Infinera (a Big Data portfolio company) for $2.3 billion, representing a 28% premium. Infinera, which is a supplier of optical semiconductors and networking solutions, will allow Nokia to increase the scale of its optical networks business and to raise its exposure to Internet/data center customers that account for roughly…
It’s been a tough H1 for software stocks, which have lagged hardware/semiconductors names by 20-30% on average. While expectations built up last year about the revenue opportunity for many software companies shipping new products with built-in AI, the reality caught up during the latest earnings season with the likes of Salesforce, MongoDB, Workday, Cloudflare and…
The US gave a major boost to its nuclear industry yesterday as the Senate passed the ADVANCED (Accelerating Deployment of Versatile, Advanced Nuclear for Clean Energy) Act with overwhelming bipartisan support, and as the DoE announced it would fund up to $900 million to support the deployment of small modular reactors (SMRs). The ADVANCED Act is a…
The 2024 edition of the Computex – a Taipei-based electronics show where many Asian companies are usually showcasing their latest motherboards, PC power supplies or laptops – turned into an AI festival last week with a flurry of non-Asian companies taking the stage to pitch their roadmaps, new chips and systems not only for AI…
It’s been an eventful week for the high performance computing and data center industry as many chip makers, including Nvidia and AMD, and data center suppliers announced new products and architectures during the Computex show in Taiwan. Among the flurry of announcements, we view Super Micro’s introduction of liquid-cooled AI racks and bullish expectations about…
There has been a lot of talk lately about the implications of surging AI usage. While the focus has been on electric power consumption, the water intensity of data centers should also emerge as a major challenge going forward. Indeed, AI is catalyzing a shift towards power-hungry “monster chips” that are becoming increasingly difficult to maintain…
Even if it’s a decades-old industry, video games remain an incredibly young business that stands at the forefront of new technologies including, obviously, AI. Since the beginning of the year, AI has been all about multimodal models and notably text-to-video, with OpenAI revealing its generative model Sora. But Google pushed the envelope of what’s possible,…
After some panic selling in recent days sparked notably by the absence of a Super Micro preannouncement, it looks like investor worries could start to ease following a set of strong quarterly figures and encouraging comments from Tesla, Vertiv and Texas Instruments. One of the highlights in the Tesla report was the major ramp-up of…
The disruption caused by Generative AI is not only deeply impacting the Tech industry and (soon) almost any economic activity but it will also have geopolitical, environmental, and energetic consequences. Since the early 2000s, cloud computing has rapidly grown to become the dominant IT platform, delivering any type of digital and/or communication service around the…
Reports have surfaced in the last couple of days about the concerted attempts of large Tech companies including Intel, Qualcomm and Google to break Nvidia’s dominance in data center AI chips by creating an open-source software platform on which developers can build AI models and apps. Indeed, Nvidia’s monopoly on the booming AI training segment…
A lot has been said in the last couple of days about the impressive performances of Nvidia’s latest family of chips, Blackwell. But one thing that has gone largely unnoticed is Nvidia’s shift towards liquid cooling for the first time for the high-end version of its next-generation DGX server subsystem, SuperPOD, that is built on…
At its semi-annual GPU Technology Conference (GTC), Nvidia showcased its latest server-class GPU family, called Blackwell, offering a jump in both computing power and bandwidth and, most importantly, a significant increase in integrated memory. As mentioned in our July 2023 article, memory is the main bottleneck in the training of large GenAI models. As data…
Even if data is widely considered as the oil of the 21st century, it has been largely under-monetized until now. As an illustration, OpenAI developed ChatGPT with massive amounts of data, free of charge… But data-rich companies, such as social media, have taken notice and the situation is rapidly evolving with companies including X and Reddit…
The stellar rise of Generative Artificial Intelligence continues to shake up the whole IT industry. At the end of last week, the world’s largest semiconductor foundry, Taiwan Semiconductor (TSMC) and the world’s second largest memory maker, SK Hynix, announced a new alliance around the development of AI chips. This news, which has gone almost unnoticed,…
Two days ago, AMD unveiled its long-awaited server-class GPU chip called the Instinct MI300X. This chip is positioned to be a direct competitor to the successful H100, Nvidia’s current Generative AI workhorse. As GenAI models keep on growing in size, memory and the way to access it represent one of the key elements of focus…
The Who’s Who of High-Performance Computing (HPC) is currently gathering in Colorado, US for the annual SuperComputing conference (SC23). Halfway through the conference, one trend is pretty clear: the whole IT industry is all in on Generative AI, starting from the chip designers up to the large system builders. Nvidia teased its upcoming server-class GPU…
During its recent earnings conference, security and edge computing infrastructure provider Cloudflare highlighted the AI inference opportunity, stating that it has a pipeline of customers (large and small) “interested in putting hundreds of billions of inference tasks on its infrastructure each month” and that it will have rolled out inference optimized GPUs in 100+ of…
Chip equipment makers have been hit in recent days and weeks by the intensification of export controls to China and by order delays as manufacturers are experiencing troubles in building their new fabs. Most notably, Taiwan Semiconductor (TSMC)’s $40 billion Arizona plant is running behind schedule. Accordingly, Dutch equipment giant ASML delivered this morning a…
There’s been a lot of attention in recent months about the computing power requirements of Artificial Intelligence and the rise of various pieces of hardware (GPUs, ASICs, TPUs…) used to train massive data models. Surprisingly, there’s been much less hype around software and, specifically, around data science and analytics software that appears as important as…
Electronic Design Automation (EDA) is the bedrock of modern digital transformation, encompassing software-based tools used by all semiconductor companies to design cutting-edge chips and complex integrated circuits that power our technological landscape. EDA has, accordingly, accompanied the growth of the semiconductor industry over the years. Looking forward, the growth pace of the main EDA players,…
As we commented a couple of months ago, generative AI has been emerging as a weapon for hackers, making phishing attempts (emails or texts inviting a user to open a malicious link to steal his/her credentials) bespoke and then much harder to detect, and helping to reduce malware software development cycles and/or rewrite code to…
If all investors’ eyes are currently focused on increasing microprocessors’ performances and speeds, a bottleneck is nevertheless emerging in AI training. This is called the memory wall, or the growing speed difference between microprocessors and computer memory, hence leaving the processing units idle until the requested/needed data becomes available for computing. The recent explosion of…
Even if data is widely considered as the oil of the 21st century, it has been largely under-monetized until now. As an illustration, OpenAI developed ChatGPT with massive amounts of data, free of charge… But data-rich companies, such as social media, have taken notice and the situation is rapidly evolving with both Twitter and Reddit announcing…
As the growth in data centers’ construction continues at an unabated pace around the world – hence sustaining investments in semis, software, networking, and storage solutions – the rise of Generative AI services (e.g., ChatGPT) is literally triggering a little revolution inside them. The workloads (training and inference) of these Deep Learning models necessitate an…
Nvidia’s last night earnings report was impressive, to say the least. The company notably guided to a 64% top-line growth in Q2 (53% above consensus!), including c.100% year-on-year growth in its data center segment. This strength should not come as a total surprise given that most companies are currently racing to deploy AI applications and…
The phenomenal success of Generative AI (GAI) services like ChatGPT or DALL-E has literally triggered an arms race amongst AI’s world leaders as the accuracy level coupled to the creative skills of GAI models are disrupting the $120bn annual revenue generating Web search industry (still heavily dominated by Google) as well as the nascent AI-based…
The so-called Moore’s law, promulgated by Intel’s co-founder Gordon Moore in the mid-60s, states that the number of transistors in an integrated circuit (or IC) doubles every two years. Contrary to what many media said and are still saying nowadays, the doubling of IC’s density doesn’t mean that semiconductor chips are twice as powerful every…
ChatGPT literally took the world by storm since its release in November 2022. The chatbot is so successful that it added 1 million users in only 5 days, slashing the iPhone’s (74 days) and Instagram’s (2.5 months) previous adoption speed records. OpenAI, the company behind ChatGPT, was founded in 2015 and backed by prominent figures…
Semiconductor stocks have suffered an across-the-board selling in recent months thanks to the rapid deterioration of several end-markets (PC, smartphones…) and new export controls imposed by the US government on chips and equipment shipped to China. But this “implosion” could also create buying opportunities as some end-markets and, more specifically, the data center one enjoys secular…
NAV & Performance
ISIN
CH1263269362
Currency
USD
Type
Active Equity
Issuer / Rating
UBS / S&P A+
Inception Date
10.05.2023
Liquidity
Daily