Nvidia (NVDA), fresh off an AI fuelled surge of 25%, has announced a range of new AI products including Spectrum X, touted as the “world’s first high-performance ethernet for AI”. The ethernet switch is built on Taiwan Semi’s (TSM) 4nm technology and consists of 100 billion transistors, routing 64 800-gig ports. This is at parity with Broadcom’s Tomahawk 5. With this tech, Nvidia are also building a test-bed generative AI supercomputer called Israel-1 using Dell’s PowerEdge XE9680 (GPU server) composed of Nvidia H100 GPUs. In addition to this, Nvidia announced a new AI computer, the DGX GH200, a “new class of large-memory AI supercomputer for giant generative AI models”. The GH200 (across 24 racks, below) is the first product to ship with Nvidia’s “superchip” (of which there will be 256!). This superchip (the Grace Hopper board) incorporates a Hopper GPU and a Grace CPU (the latter being an Arm-based CPU to compete with Intel and AMD’s x86-based CPU). This hardware, along with Nvidia’s software stack and ecosystem (namely CUDA), put it at a distinct advantage to power the current wave of supercomputing (and it’s being priced as such!).
In the same announcement, Nvidia announced a partnership with global advertising giant WPP (WPP.LN). Leveraging Nvidia’s Omniverse Cloud, WPP will be able to connect with third-party platforms like Getty Images and Adobe (including its AI product Firefly) to render thousands of images and videos. For example, creatives (or non-creatives!) will be able to seamlessly sub in a backdrop of Sydney Harbour or the Singapore skyline to target consumers in specific demographics. Ultimately, this paves the way for advertising to be even more granular in its targeting. For example, if you have a young family and love surfing, AI would ultimately be able to have these renders in the cloud, ready to target you.
Foxconn (2354.TW) is also getting an AI infused boost, forecasting it’s AI server business at least doubling in the second half of the year. As well as being Apple’s largest manufacturing partner, Foxconn also produces a large portion of servers sold by the likes of Dell, as well as cloud operators Amazon and Google.
JPMorgan (JPM) has submitted a trademark application for IndexGPT, a chatbot to provide online financial information and advice to its customers. According to the application, IndexGPT will tap “cloud computing software using artificial intelligence” for “analysing and selecting securities tailored to customer needs”. Similar technologies are being developed at Goldman Sachs and Morgan Stanley, however, JPM, appear to have a lead amongst the big banks given a live product must be released within three years of trademark approval. I tested an ‘advisory’ chatbot called potato.trade, asking “I would like to capitalise on the AI age”. It responded (below) with an allocation of 40% to AI software, 25% to robotics, 15% to autonomous vehicles, 10% to machine learning and 5% to big data and cloud respectively; much of which was allocated to broader ETFs (as well as individual stocks like Nvidia). This is obviously a very premature product today, but it is certainly something that will revolutionise advice (and cause headaches for regulators!).
According to Fidelity, Twitter (private) is now worth one-third of what Elon Musk paid for it back in October (~$15bn). In March, Musk said employees would receive stock awards based on a $20bn valuation, according to the WSJ but was quick to highlight “a clear, but difficult, path to a >$250b valuation”!
Alphabet (GOOG) and HPE (HPE) have participated in a $154m Series C round in Lightmatter (private), bringing its total raise to $267m. Lightmatter develops ‘photonic’ processors that are “faster, more efficient and cooler” than any processors on the market. This is particularly critical for the AI end of the market where there is an increased need for compute speed and low energy density. As well as developing its own AI chips (Tensor Processing Units), Alphabet are active investors in emerging technologies like Lightmatter and SambaNova. Other emerging startups attracting considerable corporate investment include Mythic (HPE, LAM Research, Micron), Graphcore (Dell, Samsung, Bosch) and Vastai (Alibaba, Kuaishou).
Elon Musk’s brian implant company Neuralink (private) has received FDA approval to conduct its first clinical trials in humans. The implant (below) can decode brain activity and link it to a computer. Such technology has the potential to restore function to people with debilitating conditions such as paralysis. But beyond this, it is also touted as something which will surpass “able bodied human performance”.
The FT published a fascinating chart this week outlining the supply chain of EV battery components; or more accurately, China’s stranglehold on the supply chain. Interestingly, this was an element of a discussion I attended last night with Australia’s former Chief Scientist Alan Finkel who, in his new book Powering Up, discusses the need (amongst other things) for Australia to significantly mature from being a pure miner, to being a processor (and beyond). At present Australia mine ~50% of all lithium (spodumene), but export almost all of that to China who refine and process it into lithium sulfate and lithium hydroxide respectively. This is a key ingredient for EVs. Companies like Mineral Resources and Albermarle, along with IGO and Wesfarmers are developing on-shore processing capabilities which will help reduce the current dependency on China. The on-shore processing of other materials, however, is significantly more challenging.
Three Chinese astronauts, including a civilian, have successfully docked with the Tiangong space station, relieving crew (below) from a prior Shenzhou-15 mission. This is China’s fifth manned mission to the space station since 2021. China has plans to launch a crewed mission to the moon before 2030 after having successfully landed rovers and returned specimens from the moon’s surface. At present, the US are planning on returning astronauts to the moon by early 2026 as part of the Artemis program.
Thank you and have a great week and weekend ahead!
Charlie Nave
LinkedIn or E-Mail (cnave@granitebaycap.com)
Associate Professor (Practice) Monash Business School and Monash Center for Financial Studies
Granite Bay Capital is an innovation focussed consultancy with a deep focus on the companies at the leading edge of innovation across major themes such as AI, ubiquitous computing, sustainability, automation and longevity. Any views expressed in this article are those of the author and do not constitute financial advice.