• Wayve’s deal with Nissan is a big shot in the arm for the “brute force” approach to autonomous driving, but Nissan has made no promises to use it beyond level 2 leading me to think that a somewhat reluctant Nissan has been coaxed into giving it a try by SoftBank.
• Wayve is a UK-based autonomous driving start-up that uses a single large end-to-end model to drive the vehicle.
• This means that sensor data goes in one end and driving instructions to the vehicle pop out the other.
• The advantage of this is that if one can get to work, then there is no need to limit where the vehicle can go which also means that no HD map will be needed.
• The dream of autonomous driving is to have software that can drive a vehicle more safely than humans under any conditions and be able to deal with situations for which it has not been explicitly trained.
• This is exactly how humans do it and as long as one is prepared to exchange a large neural network for a human brain, then all should be well.
• However, this is a bridge too far for me which brings us right back to RFM Research’s old chestnut of causality.
• Humans can drive a vehicle safely because they understand the cause and effect of the road, while the large model merely matches inputs to statistical characteristics and estimates what the output should be in the given situation.
• For example, any human would never mistake a large restaurant sign with red, yellow and green circles for a traffic light but unless the machine has been explicitly taught about that sign, it will.
• This means that for situations where the dataset is both stable and finite (i.e. all outcomes can be predicted and trained for), then a neural network can perform really well.
• However, the road is neither finite nor stable which makes a large neutral network a suboptimal choice to solve this problem.
• This is where opinion in artificial intelligence diverges.
• On the one hand, you have Elon Musk, OpenAI, SoftBank, Anthropic and so on who claim that with a big enough model, enough data and enough compute, magically, machine superintelligence will pop out at the other end.
• This is the argument that keeps the money pouring in and the valuations at very high levels.
• On the other hand, there are the sceptics and gadflies like Gary Marcus, RFM Research and many others who think that until a statistical-based system can truly reason, we will be as far away from superintelligent machines as we were 10 years ago.
• In my opinion, the “reasoning” models are not actually reasoning but simply offering up a very good simulation of it.
• This is because while the models can ace PhD level maths, they fail to reason that if A=B, then it follows that B=A.
• This is the classic paradox that has plagued AI for decades in that machines can be taught to do very difficult things but fall to bits when asked to do the simple stuff.
• It is not until this issue is beginning to be solved that I think the Wayve approach to autonomous driving has a chance of working in a truly commercial setting.
• One can see this in how Nissan will be using Wayve’s technology starting in 2027 where it will be used for level 2 only at the outset (see here).
• Level 2 is hands-on ADAS where the human is still piloting the vehicle and does not go much beyond staying in lane and adaptive cruise control.
• I take this to signal a “let’s see” approach and I suspect that as SoftBank is a major investor in Wayve and is championing a collaboration between OEMs to share data and resources to achieve full autonomy with an end-to-end system, it has had some influence on Nissan when it came to taking software from Wayve.
• Nissan has made no commitment that I can see to take this beyond level 2, and so I do not take this as a sign that the end-to-end large model approach is the right one.
• In fact, I think this approach will end up falling short and an approach that uses a combination of rules-based software and machine learning will be the one that wins out at the end of the day.
• This also means that autonomous driving components such as an HD map, lidar, radar and cameras will all be needed to help reduce the volatility of the dataset of the road as well as produce redundancy that can make the cars safer than humans.
• With all of the hype and excitement around LLMs, this approach is currently not in favour, and so I suspect that it will be later rather than sooner that we begin to see fully autonomous vehicles on the road.
• Hence, I think that Wayve and Deeproute.ai (which also uses this approach) will never become going concerns in their own right and will end up being acquired.

• Google has come out swinging against DeepSeek claiming that Gemini is more efficient in both training and inference, but its comments are even more nebulous than DeepSeek’s leading me to think that there is something in what DeepSeek has been claiming.
• In an interview with Bloomberg (see here), Desmis Hassabis makes a series of observations and claims with regards to Gemini vs. DeepSeek some of which are credible and some of which need to be substantiated before they can be believed.
  • First, Gemini is more efficient: Mr Hassabis claimed that Gemini is more efficient than DeepSeek R1 both in terms of “training to performance and its cost to performance” but failed to offer any evidence to support this claim.
  • Given that the performance of the top models these days are all within spitting distance of one another, Mr Hassabis is basically saying that Gemini is cheaper to train and cheaper to run than DeepSeek R1.
  • Unfortunately, this was substantiated by the statement “We don’t talk about that very much”, which, in my opinion, is worth nothing.
  • DeepSeek R1 is the first model I have come across where the fine-tuning step is fully automated which, combined with lower costs in general in China, and Mr Hassabis’ other comments, leads me to think that DeepSeek has achieved more than is being admitted to here.
  • Second, no silver bullets: While Mr Hassabis admits that DeepSeek is the “best team that I have seen come out of China” and that the model is “very impressive” he is correct to point out that there is nothing particularly new about what DeepSeek has done.
  • This concurs with RFM Research which has concluded that DeepSeek has not used any novel techniques, but it has done so in a way that has produced potentially very interesting results.
  • Across the technology industry, there are many examples of existing techniques being used in novel ways resulting in a large step forward in terms of performance and efficiency.
  • Consequently, the fact that DeepSeek has not invented a new technique does not detract from the possibility that the claims made have some basis.
  • In fact, the defensive nature of Google’s statements which had clearly been prepared in advance leads me to think that behind the scenes, DeepSeek has rattled Google and it is working to verify precisely what DeepSeek has achieved.
  • Third, DeepSeek is exaggerating: which is almost certainly true.
  • DeepSeek very cleverly implanted a $6m training figure in its press release (dressed up to look like a scientific paper) which everyone immediately latched onto and assumed that this was the total cost.
  • This was then compared with OpenAI’s rumoured $500m cost to develop GPT-4 resulting in the misconception that DeepSeek is 100x cheaper to train.
  • This is not a fair comparison as the $6m refers to the last training run that DeepSeek carried out and does not include all of the other training runs, design and build costs and company overheads that were incurred to get it to its published state.
  • RFM’s back-of-the-envelope calculation reveals that if all of DeepSeek’s claims are true, then it could be something like 7x cheaper than OpenAI.
  • This makes sense to me as OpenAI is a voracious consumer of compute as it believes that bigger models and bigger compute will make its models better.
  • With an almost endless supply of money, it has never had to worry about efficiency whereas China has been cut off from cutting-edge compute for a few years.
  • Back in October 2024, I postulated that a silver lining to the resource constraints that were aimed at containing China’s AI development would force it to innovate in this area (see here) which appears to be exactly what it has done.
  • Hence, the most likely source of innovations around efficiency was always going to be China as opposed to the West where the magic money tree (see here) is still showering its bounty upon anyone working on AI.
• The net result is that I think that Google is making some valid claims with regard to DeepSeek, but is underplaying the advances that DeepSeek has made when it comes to efficiency.
• What Google (and everyone else) will now do is attempt to replicate DeepSeek’s techniques and if they work, incorporate them into their workflows.
• DeepSeek is at a minimum working with the explicit approval of the Chinese state, and I suspect that there is more state involvement than DeepSeek is letting on.
• Hence, if the advances are as good as DeepSeek says it is, it makes no sense for China to give the innovations away and receive nothing in return for them.
• This is why I think that  DeepSeek innovations are in how it has modified and implemented existing techniques to train its models rather than in the model itself.
• These, it can easily keep to itself despite releasing the model to open source and China can then benefit from all of the PR buzz without having to give the real intellectual property away.
• When the magic money tree runs out of leaves and the correction comes, everyone will be forced to do more with less and in this circumstance, China may find itself ahead in this particular area.
• RFM and Alavan Independent have always rated China as a force to be reckoned with in AI where it will put up much more of a fight than it has in semiconductors.
• This looks certain to set the tone of the ideological struggle for 2025.

• Stargate is an ambitious project to build up to $500bn of AI Cloud capacity but does not take into account the fact none of the current players except SoftBank and MGX have the money to complete even the first, $100bn phase.
• Stargate is a joint venture formed between OpenAI, SoftBank and Oracle that intends to build a series of data centres as well as the supporting infrastructure to enable the anticipated mass adoption of AI.
• The initial phase is $100bn and it seems that OpenAI and SoftBank will each put in $19bn with Oracle and MGX putting in $7bn each.
• This leaves $48bn left to find for the first phase but given the glitz and sparkle that is being attached to this venture, I suspect the rest will be forthcoming without too much difficulty.
• The best way to think of Stargate is as a private equity fund with OpenAI and SoftBank as the main partners which will invest in projects that build and operate AI infrastructure.
• The remit is wider than just data centres and I suspect that there will also be money for investing in electricity generation which is rapidly becoming a key bottleneck for AI.
• Arm, Nvidia and Microsoft are also involved as technology providers but there is no doubt that this signals the end of OpenAI’s exclusive relationship with Microsoft.
• Oracle’s involvement makes it very clear that OpenAI will be running on Oracle infrastructure in some of the financed projects in a sign that Microsoft has had enough of coughing up vast amounts of cash for OpenAI.
• Microsoft also made a statement relating to this project where it confirmed its continued partnership but also stated that its relationship with OpenAI had moved from an exclusive partnership to one where it has the right of first refusal (see here).
• Consequently, I suspect it was asked if it wanted to part with another $20bn+ to which it replied, “No thanks”.
• Microsoft has also added support for Mistral and Anthropic in its data centres, but I think this is more about giving clients choice as I don’t think that its own offering has followed suit.
• The biggest question I have is where the money is going to come from as SoftBank has just $25bn of cash and cash equivalents, OpenAI has less than nothing given that it is burning cash like there is no tomorrow and Oracle has just over $10bn of cash on its balance sheet.
• Where the other $400bn (roughly 3x total hyper scaler 2025 capex) will come from also remains a mystery.
• The other question is demand as CES 2025 clearly demonstrated that generative AI is not ready for the consumer and all of the inference that was supposed to materialise in 2024 failed to do so.
• Furthermore, I think that in the long run, most inference will be conducted at the edge of the network and not in the cloud meaning that most of the built capacity will be for training rather than inference.
• Hence, there is a good chance that there is a correction in expectations and valuations and the $400bn takes far longer to materialise than anyone currently expects.
• In the meantime, this makes great geopolitical theatre cementing the USA’s leadership in AI and putting China and DeepSeek back in their box.
• Hence, I expect to see a lot more talk as well as some initial investments but once reality reasserts itself, Stargate will be a smaller, but probably better and more efficient investor.

• There are more signs that the “scaling laws” that have underpinned the AI explosion (and all of the hype attached to super-intelligent machines) are coming to an end meaning that the real potential of LLMs is now visible and is falling way short of the craziest of forecasts.
• It is important to note that these new indications are anecdotal and as such do not represent any form of empirical proof, but they add to what is already being seen with existing models and how they are performing in reality.
  • First, an article from The Information: (see here) that claims that Open AI’s new models are not improving as quickly as expected and so the company is looking at new strategies to keep improving the performance of its new models.
  • Since its inception, Open AIs belief has been that with enough data and enough compute, artificial superintelligence would magically pop out at the end.
  • I have often referred to this as the “Infinite Monkey Theorem” (see here) and have held the opinion since 2020 (when I first wrote about LLMs (see here)) that this would not hold.
  • The radical underperformance of Open AI’s o1 model relative to what we were told is yet another sign that LLMs are beginning to experience the law of diminishing returns.
  • Second, commentary from an industry insider: who is the CEO of Deep Trading (algorithmic trading) who claims he was told that another one of the leading creators of LLMs has also hit a big wall of diminishing returns (see here).
  • This is even more tenuous than the article from The Information, but it adds weight and a second unrelated “data point” implying that LLMs are beginning to reach the limits of what they are capable of.
• Diminishing returns is a huge problem because it means that one has to use increasing amounts of resources in order to achieve smaller and smaller improvements.
• This very quickly becomes uneconomical, and any system that is funded by private money soon finds that willingness to pour more money in quickly dries up.
• This could easily trigger a correction of expectations which in turn would cause valuations of the most outlandish companies to fall meaningfully.
• It is my opinion that diminishing returns have been evident for quite some time supported by the fact that there is not much difference between the big models today in stark contrast to 2 or 3 years ago.
• It is the slowing improvements that allows the laggards to catch up which is why when one looks at the benchmarks these days, the differences are minor.
• LLMs still have substantial use cases that will deliver great economic benefits and spawn a new industry, but superintelligence is as far away today as it was 10 years ago.
• The LLM superpowers of using natural language as a man-machine interface and the ability to ingest categorise, cross reference and regurgitate unstructured data remain very much intact and when properly used will be extremely valuable.
• These two abilities alone open up many possibilities for new businesses as well as the replacement or improvement of businesses that already exist.
• I am not bearish on the outlook for LLMs but merely cautious on valuation as expectations have run far ahead of what is realistically possible meaning that a correction is needed to bring expectations back to reality.
• The robots are not coming to kill us anytime soon, but they will be making an appearance in the economy in ways that will make digital life for users better and more productive as well as allow companies to make far better use of the data that they already have but have forgotten about.
• There is a correction coming and the time to invest will be when everyone has given up on AI and moved on to the next bright and shiny theme.

• The loss of 60% of its founders just months after being founded and raising $220m despite having no products, looks like a sign that in the rush to invest in AI, red flags are being ignored and due diligence is being skimmed which always ends in problems.
• The H Company (formerly Holistic) was founded earlier this year by 5 founders from Google DeepMind and despite having no revenues or even products, was able to raise $220m from Accel Partners, Amazon, Samsung and other well-known faces from the billionaire class.
• Now, just 2 months after the raise, 3 of the 5 founders have left the company with the company citing “operational differences” after one of the leavers spoke to The Information and said the same thing.
• This is a strange turn of events and a cynic would immediately begin to wonder if this was an attempt to cash in on the DeepMind name and AI while it remains a super-hot commodity.
• This early break-up signals to me that the investors were so desperate to invest more money in AI that they ignored the warning signs that led to this which has put their investment at risk.
• Typically, this sort of break-up occurs at a point of stress when a company is forced to make a change of direction that had not been previously anticipated.
• This point of stress is usually triggered by either a realisation that the product is not fit for the market (H has no product) or the company runs out of money and has to make sales very quickly (H has just raised $220m).
• Hence, I can only conclude that whatever the cause of the break-up was, it was always there and did not suddenly appear meaning that anyone doing their due diligence properly would have found it.
• Instead, because there is so much money flowing into the sector, projects that would normally fall foul of a rigorous investment process get funded which in turn leads to failures, popped bubbles and so on.
• This is Accel Partner’s first foray into this space making me wonder whether the company was so determined to catch up with its peers that it failed to take a hard look at the H Company.
• The problem here is that 3 of the 5 founders leaving the company will greatly reduce the market’s confidence in this business as a going concern making it harder to sell products (when there are some) and hire staff on the promise of a big equity-based payout.
• The investors are publicly rallying around the H Company stating that they “are confident that they can deliver on this mission and continue to support them” but I suspect that they are deeply concerned about the current outlook.
• To me, this is another symptom of the magic money tree at work and just like 1999 and 2000, many projects are being funded that should not be as a result of the oversupply of cheap money.
• At some point relatively soon, there is going to be a correction where the supply of money dries up, valuations become cheaper, and expectations are set that are based on reality rather than the blue-sky prospect of super-intelligent machines.
• This is why I remain very nervous about the valuations that are being paid for these companies and am staying well away from them.
• If I were forced to invest in this area, it would be Nvidia which actually has revenues and profits now or Qualcomm which is in a very good position to benefit as generative AI starts to be implemented at the edge.
• I already own Qualcomm which remains reasonably valued and has further upside from its position in automotive as well as the potential to take share from Intel and AMD in laptop processors.

• SoftBank, Nvidia and Microsoft are pouring $1bn into a system that relies entirely on machine learning to drive vehicles in the hope that if the model is big enough and trained with enough data magically, the answer will pop out at the end.
• I suspect that when it comes to the problem of driving cars, they will be disappointed.
• SoftBank is leading a $1.05bn round into UK-based autonomous driving start-up Wayve and I suspect that it is providing almost all of the funding.
• Microsoft is an existing investor and together with Nvidia, provides credibility to the investment even if they are risking very little in the transaction.
• This is because most of the money that Microsoft and Nvidia are investing will come back to them in revenue for Azure or sales of chips to run the models in the vehicles.
• While I think that there is nothing wrong with Wayve as a company, I think that trying to crack an infinite problem with a finite solution is not going to work.
• This is because all systems based on deep learning and neural networks have no causal understanding of what it is that they are doing and as such, are unable to deal with any situation that they have not been explicitly taught.
• Hence, for tasks where the dataset is finite and stable, deep learning can perform tasks to a superhuman level of ability.
• However, the road is neither of these things which is why I continue to think that deep learning on its own is insufficient to provide the causal understanding that is required to drive the road as safely as humans.
• I have long argued that the best solution for the road will be one that uses a combination of rules-based software and machine learning working together.
• This is because software can reason but it can’t learn while deep learning can learn but it can’t reason meaning that if they can be properly integrated, then they should be able to complement each other.
• This is how I have long thought that the autonomous driving problem will eventually be solved, but both Wayve and Tesla, with the latest version of its Full Service Driving (FSD 12.1), are going in the opposite direction.
• Consequently, without a system that has some understanding of causality and an ability to reason, I think that neither Wayve nor Tesla will arrive at a solution that can drive vehicles as well as humans.
• This issue is not limited to autonomous driving systems but is also rampant in large language models (LLMs) which consistently make things up, get things wrong and are generally unreliable as a result of their lack of understanding of causality.
• Wayve’s business model is to license its software to vehicle manufacturers rather than go direct to market with a fleet of robotaxis which is the right choice.
• However, I think that this will be problematic as I am not convinced that its solution will be better than Waymo, Mobileye, Cruise or any of the other very good Chinese offerings.
• Consequently, competition will be tough and prices low meaning that only the biggest and best solutions survive.
• This is why I think that the market for both robotaxis and autonomous driving software will not be nearly as big as SoftBank’s valuation of Wayze seems to indicate, leaving me very cautious about the sector from an investment perspective.
• The only sub-sector I like in this space is lidar where Ouster is my top pick as it is not dependent on the automotive sector to thrive and is in better financial condition than its peers.
• I have a position in Ouster which is discussed in more detail here.

• Capital continues to pour into the AI sector with very little attention being paid to company fundamentals in a sure sign that when the music stops there will not be many chairs available.
• The latest sign of this reckless abandonment is the fact that Cohere is able to raise money at a valuation that is 2.4x the valuation of 9 months ago despite the fact that its business development has been disappointing.
• Cohere will now be worth $5bn even though the annual run rate of its revenue in 2023 was just $13m (see here).
• Another red flag was Microsoft’s ability to hire the CEO and 70 staff from the AI start-up Inflection AI.
• Things were not going well at Inflection AI because if the company had been doing very well, Microsoft’s advances would have been swiftly rebuffed.
• Cohere’s valuation equates to a historic price/sales ratio of 384x which indicates that investors have another bad case of FOMO (fear of missing out) and are rushing into anything that can be remotely associated with AI.
• This is precisely what happened with the Internet in 1999, autonomous driving in 2017 and now generative AI in 2024.
• Foxconn, Arm, TSMC, and Astera Labs are all companies that have benefitted greatly from investor appetite for AI indicating that the phenomenon is spread across both public and private markets, but it is private markets where valuations are the most outlandish.
• No one is immune from the FOMO effect and Amazon has thrown another $2.75bn of its total $4bn commitment at Anthropic, and I am pretty certain that Amazon will end up acquiring the company.
• This is a tacit admission of the fact that Amazon is not very good at AI which has been clearly on display for years to anyone who has either used Alexa or received an advertisement for a product that they have already purchased.
• Somewhat unusually, the majority of the cash being pumped into the generative AI sector is ending up in one place which is Nvidia which holds 85% market share of cloud-based AI processors and whose lock on its market appears stronger than ever.
• This is why Nvidia’s valuation is at a relatively reasonable 36x 2024 PER despite the shares increasing 9x relative to their low at the beginning of 2023.
• Nvidia is really the only company that is making tangible profits from the current boom in interest in investment in generative AI but when there is a correction, there will be nowhere for Nvidia to escape, although I suspect that it will be hurt much less than many others.
• The ones that are likely to bear the brunt of the correction are the providers of generative AI services who are raising money on the promise of selling their services for $20/user/month.
• The problem is that there are many of these all of whom are demonstrably similar, which combined with the many models that are available from open source for free, is going to put pressure on pricing.
• It is at this point that the flow of money is likely to slow down and then stop as falling prices will mean that targets are missed and start-ups go back to VCs cap in hand.
• This is unlikely to be nearly as brutal as what has happened to autonomous driving as generative AI has products available, use cases and can generate revenues, just not as much as the market currently expects.
• Investors will begin to be more selective in terms of where they invest and will ask harder questions about revenues and profits meaning that valuations will be lower.
• I think that this is a net positive for the industry as it means that only the viable projects with the most capable management will get funded resulting in a much better allocation of capital resources and better products.
• However, it does mean a shake-out and here I expect that we will see failing start-ups being acquired by the large companies who do not have an in-house foundation model off which to base their offerings.
• In the meantime, the frenzy continues but it is one I am perfectly comfortable staying well away from.
• If I was forced to invest in this area, it would be Nvidia which actually has revenues and profits now or Qualcomm which is in a very good position to benefit as generative AI starts to be implemented at the edge.
• I already own Qualcomm which remains reasonably valued and has further upside from its position in automotive as well as the potential to take share from Intel and AMD in laptop processors.

• Google held its Made by Google 2023 event where it once again demonstrated some great innovation which would work much better if it was in a device made by Samsung device rather than Google.
• However, both companies continue to be afflicted with chronic engineering disease meaning that Apple will remain pretty much unchallenged.
• Google launched the much-leaked Google Pixel 8, 8 Pro, Watch 2 and new colours for the Pixel Buds Pro that largely underwhelmed on hardware although the new screens on the Pixels are especially bright.
• There is also a new Tensor G3 chip although Google omitted to mention how much of an improvement the new hardware represents underlining yet again that Google is not a hardware company.
• Instead, Google allowed the software to speak for itself and here a series of exclusive features both make these devices shine and underline what a wasted investment they represent by limiting them to Pixel devices.
• These include 7 years of software support and updates, substantial enhancements to the digital assistant using Google’s generative AI models, and as always, image processing.
• The new AI assistant will be able to take accurate dictation running on device as well as integrate with Gmail and Docs and act very much like a smaller version of Microsoft’s CoPilot with the benefit of it not costing $30 per month.
• However, where the device really shines is in image processing with some really impressive new features.
• One of these is called Magic Editor which allows objects to be moved and resized within a photo which also enables the background to be changed such as the colour of the sky.
• Another is called Best Take which takes multiple photos of a group shot and then enables the faces to be changed from one shot to another to enable the best face from each shot to be composited into the group photo.
• Another feature is called Audio Eraser which removes unwanted background noise from videos as well as an off-device feature called Video Boost which will automatically improve videos uploaded to the cloud.
• It is important to note that the demonstrations all featured images and videos where the contrast between the subject to be edited and the rest of the photograph was very high.
• This makes it far easier for the AI to strip out the subject from the rest of the photograph or video and in my experience, when the contrast is lower, these sorts of features do not work nearly as well.
• Despite this drawback, the features that Google demonstrated are the best I have seen on any device and far outstrip anything that Apple has to offer.
• However, these features will not cause Apple to lose any sleep because they will only be present on devices that ship in such tiny volumes that they will have zero impact on the purchasing decisions of smartphone users.
• This is where Google has been getting it wrong for years and it continues to labour under the idea that it can challenge Apple.
• The reality is that it has no chance, but if these features were available on Samsung products, then it would have a much better chance of having some impact.
• The same is true for Samsung which also labours under the same idea and continues to produce software for its devices that is easily outstripped by Google.
• I have argued for years that the best chance of mounting a challenge to Apple would be created by Samsung and Google working much more closely together with Samsung building the hardware and Google the software.
• Instead, each company continues to do both, resulting in suboptimal devices in either hardware or software.
• Both of these companies continue to be afflicted with engineering disease (see here) and there is no cure anywhere in sight.
• Consequently, Google will continue to earn a negative return on these excellent new features as its hardware volumes are likely to remain so low that it will be losing significant amounts of money in hardware.
• Samsung is also forgoing the opportunity to grow market share against Apple making Apple the real winner from these afflictions.
• With no cure in sight, Apple looks set to continue chipping away at the high end of the Android market.
• Hardware is a rounding error in Google’s numbers but the more Android users defect to Apple, the less time they will spend in the Google Ecosystem which will cost Google in terms of opportunity.
• The valuation of both Apple and Google look pretty full, but if I were to delve into this area (which I am not at the moment), Samsung is the one I would be looking at.

• Arm China is holding up the relisting of Arm and the less-than-ideal solution that SoftBank is proposing highlights again my view that SoftBank may be better off holding onto Arm and relisting it another day.
• The situation between Arm and Arm China remains unresolved which has meant that Arm is unable to audit Arm China which in turn is holding up the IPO.
• SoftBank’s solution is to move the shares into one of its subsidiaries and create a series of contracts (very much like the Chinese variable interest entity) that will mimic ownership.
• That way, the IPO can go ahead without the audit of Arm China but I think this increases risks for investors.
• Around 20% of Arm’s revenues come from Arm China (in which it owns 47.3%) which with this solution will mean that these revenues come from a black box.
• The whole point of an audit is so that shareholders can accurately assess the prospects and risks of an investment and this solution will reduce that visibility.
• Hence, I think that this increases the risks of investing in the IPO meaning a higher discount rate and a lower present value.
• This combined with the fact that Arm’s IPO is unlikely to fetch the same valuation that SoftBank would have achieved from Nvidia raises the possibility that the shareholders of SoftBank might be better off if the IPO is delayed for a few years.
• This would give Arm more time to return to the levels of profitability that it enjoyed before SoftBank acquired the company and the large investment cycle began.
• This would allow a better valuation to be achieved and a better return to be earned by SoftBank’s shareholders.
• Warren Buffet famously once said “the stock market is a device that transfers money from the impatient to the patient”, which I suspect is relevant here.

Alibaba & nReal – There can be only two.

• RFM research has concluded that there can only be one Metaverse if it is to be successful, but it increasingly looks like there will be one for China and another one for everyone else.
• The big Chinese ecosystems Alibaba and Tencent have yet to make their move into the Metaverse but Alibaba’s investment into nReal is the first concrete sign of where it is going to go.
• Alibaba is investing $60m into nReal at a valuation that I suspect is comfortably above $1bn.
• nReal is a maker of augmented reality glasses that has routinely surprised me in terms of the user experience that it is able to produce and the price point that it charges for its products.
• nReal’s products are good enough that Magic Leap felt that it had to sue nReal for patent infringement, but the case ended up being thrown out.
• Hence, nReal is one of the leading contenders to create the Chinese metaverse which increasingly looks like it will be isolated from everywhere else.
• The strategic rivalry between the USA and China has become an ideological struggle meaning that both countries are continuing to decouple from each other.
• Nowhere is this more evident than in technology and this is likely to mean that interoperability efforts will be limited to a Chinese metaverse for Chinese users in China and another one for everyone else.
• This is bad news for everybody because a balkanisation of the Internet will mean that it will create much less value than it otherwise would have done had it remained whole.
• I think that the same is true for other technologies like AI, robotics, 6G and so on which are also likely to end up with one standard for China and another for everyone else.
• This means less value creation from the Metaverse and therefore a lower return from investing in the companies that will create it.
• The Metaverse remains pretty uninvestable in my view because its reality is so far away that investing in the companies that are in pole position to create it today is an investment in their current businesses, not the metaverse.
• For example, if one buys Meta, one is investing in social networking and not the Metaverse and Unity is an investment in the video games industry.
• Hence investors have to like the shorter-term businesses as well as their prospects in the Metaverse in order to buy the shares.
• Fortunately, this is the case with Alibaba which trades at a tiny fraction of Amazon and where the case for a recovery is steadily improving.

• Magic Leap is back in contention with a new device and a funding round, but the company has not said anything that has not already been said by everyone else, leaving me wondering where it plans on finding its competitive edge outside of hardware.
• After what looked like certain death 2020 (see here), Magic Leap has clawed its way back into contention with new management, a new device, and $500m in new capital.
• The new money has been raised from existing shareholders but at a valuation of $2bn which is 70% lower than the peak valuation of $6.7bn achieved in 2019.
• However, because the money is coming from existing shareholders, they will not suffer dilution of their holdings, but they will have already marked the investment down materially in their books.
• Alongside the new money comes a new device, the Magic Leap 2 which looks a lot better than the old one.
• Magic Leap has increased the field of view significantly but interestingly, it has opted to expand it vertically far more than horizontally.
• I suspect that this is in line with its recent pivot towards the enterprise where utility is far more important than look or feel.
• In these use case scenarios, having a more complete picture of what is directly in front of you is more important than the periphery explaining this unusual decision.
• Furthermore, the device can also dim the lenses that let natural light in which will make the virtual display easier to read in bright lighting conditions.
• These characteristics combined with its much sleeker and lighter form factor (which the company states enables all day, every day wearing), are where I can see some differentiation.
• However, on the software and use case side, Magic Leap is late to the game and has a lot of ground to make up.
• As far back as 2018 (see here), it was obvious that in its initial iteration, augmented reality had to be an enterprise-focused offering.
• In response to that everyone except Magic Leap made the pivot and have been focused there pretty much ever since.
• Magic Leap only made this pivot in 2020 meaning that it has a lot of ground to make up to catch up with its rivals.
• Its new device is lighter than the others but this is partly due to the fact that it still has its compute in a puck that is connected to the headset via a cable.
• Everyone else has gone down the all-in-one route which leads their devices to have a heaver head-mounted unit than Magic Leap. But with no annoying cable.
• Despite, its late entry, Magic Leap has made some noise in the sector and has signed partnerships with Google Cloud, NVIDIA, and VMWare implying that it does have something to offer.
• It has also some endorsements from the areas that it is targeting like health care and the public sector which imply that it is also seeing some traction.
• At the same time, Magic Leap is competing against the big ecosystem companies that have very deep pockets and are able to outspend Magic Leap by orders of magnitude.
• The net result is that Magic Leap has bought itself a seat at the table with its new management, device, and funds but it still has an awful lot to prove.
• The use cases that it now talks about have migrated from cute aliens and whales to real-world scenarios but everyone else has been working on these for a much longer period of time.
• If the company really has something special outside of hardware, it is still very well hidden, and I presume that the company is deciding to keep it that way for now.
• Hence, I think this is still a long shot but if Peggy Johnson can pull this off, it will be one of the great turnarounds seen in recent times.