Headlines this week - Dec 1, 2024
A look at how capital is being deployed across future opportunities
This week in the future:
AI valuations remain high, amid concerns about a financial bubble, and rumors that progress through scale has plateaued (which could affect Nvidia)
Global tech tensions rise with countries enacting tariffs against Chinese tech products. Huawei launching its own smartphone OS suggests a “technology split”
Rocket Lab, a startup following SpaceX’s approach (reusable rockets as a way to reduce launch costs) sees its stock fly (more than +100% in just one month)
Chinese electric car vendors are becoming even more efficient. A price war is expected in China, and this is making Chinese cars (like BYD) more competitive
Google has been working on Nuclear Fusion since 2014. Fusion has already become a hot topic for researchers, that many startups are trying to solve
Europe struggling to become self-sufficient in technology. The top European AI startup will open an office in California, looking for talent. Meanwhile, the fall of Northvolt highlights its dependence on China for EV batteries
Amazon has a plan to catch up with Nvidia in AI chips, but it will not be easy… (running AI chips is a complex task, and Nvidia is years ahead)
Intel finally gets a $7.9bn government subsidy to build chip manufacturing plants in the US. Qualcomm’s interest in buying Intel might be cooling down
Biotech continues to attract funding, with pharma companies acquiring startups and new IPOs on the horizon. Uncertainties remain about Trump’s health policy
1. Population & Natural Resources
Population growth may be sustainable, but it could also make us happier. Yet one more article by Tomás Pueyo this week, about the “fertility gap”: people are having fewer children than they want. The Moral Case for More People on Earth
Biotech
Funding for biotech startups keeps flowing (even if concerns are growing about how the new Trump administration will impact the industry):
Incumbent pharma companies want to buy them. Roche just announced a $1.5bn deal to buy Poseida, a specialist in personalized immunotherapies. Roche to Acquire Poseida Therapeutics in a Deal Worth Up to $1.5 Billion
A few of them are preparing IPOs: Odyssey Therapeutics (autoimmune therapies) and Aktis Oncology (cancer drugs) are meeting potential investors. Biotechs test IPO market despite concerns over Robert Kennedy’s health role
Engineering meets biology, to accelerate progress in medicine:
AI is speeding up biotech innovations: Jennifer Doudna just wrote an article for Wired about how AI can accelerate the creation of CRISPR-based therapies Combining AI and Crispr Will Be Transformational
‘Miniaturized Electrophoresis” (an emerging technology) is helping to rapidly identify antibiotic-resistant bacteria (and assess antibiotics’ efficiency). Tiny laboratories that fit in your hand can rapidly identify pathogens using electricity
Space
SpaceX wins a $257m contract for NASA. They will launch Dragonfly, a mission to explore Titan (a Saturn moon), which will be powered with nuclear energy. NASA awards SpaceX a contract for one of the few things it hasn’t done yet
Rocket Lab, a startup competing with SpaceX, is seeing its stock fly. Like SpaceX, they see re-usable rockets as the way to drastically reduce launch costs. And the market is happy with this. Rocket Lab Shows SpaceX Isn’t the Only Game in Orbit
Tech & Geopolitics
Trump is not alone in raising tariffs against China’s tech products. Countries like India, Vietnam or Turkey are already using this policy. All this could lead to a massive change in global trade. Breaking down the world’s tariffs against China’s tech industry
A different question is if the impact will be positive. The announced launch by Huawei of its own smartphone OS indicates an increasing fragmentation (not good…) Huawei to launch phone with own software in sign of China-US splintering
Meanwhile, under pressure from sanctions, China wants to import Western talent to strengthen its local tech ecosystem. Exclusive | China Is Bombarding Tech Talent With Job Offers. The West Is Freaking Out.
2. Efficiency & Productivity
Energy
Nuclear
Nature explains the first nuclear fusion experiment with a net energy gain. This has been one of the catalysts of the increasing interest of investors and researchers in this technology. How the world’s biggest laser smashed a nuclear-fusion record
Google has already been doing research on nuclear energy for 10 years. Their R&D nuclear energy lab was created in 2014, and they’ve worked in topics like nuclear fusion (both hot and cold versions…) The Inside Story of Google’s Quiet Nuclear Quest
Renewables
Huge challenges for Europe to become self-sufficient in EV battery technologies. Things are more complex than pure political will, as the recent Northvolt fiasco shows. The Northvolt dilemma: can European EVs avoid relying on Asian batteries?
Fossil Fuels
The project to increase US oil production is under question. Trump’s Treasury Secretary pick has proposed increasing production by 3m barrels / day, but this might be unprofitable / not feasible. Trump Treasury Nominee’s Oil Math Doesn’t Add Up
New Transport Technologies
Electric Vehicles
Tesla under pressure for environmental violations in its manufacturing processes. This is an old issue about electric cars, now reactivated with Musk’s new roles in politics. As Musk Assumes Deregulation Role, Tesla Racks Up Pollution Violations
Concerns about Trump potentially eliminating EV subsidies:
This might hurt the industry in the US. Today, EV buyers in the US get a $7,500 tax credit. Eliminating that will probably decelerate sales (at least of the cheaper models). How Trump Could Upend Electric Car Sales
The timing looks particularly bad, because demand seems to be weakening now. Donald Trump’s plans threaten the EV transition
A plan by Chinese EV giant BYD to build a plant in Mexico is now under question. Authorities are worried about a potential retaliation by Trump if the plant is built. Mexico Gets Cold Feet Over New Chinese EV Plant After Trump Win
In China, BYD is preparing for a price war. They are very aggressively trying to reduce costs, and have asked suppliers to reduce prices by -10%. China’s EV price war set to intensify next year as BYD squeezes suppliers
Autonomous Cars
Singapore is now the most advanced market for autonomous car deployments. This is driven by actual needs / problems that self-driving vehicles solve (e.g.: scarcities in land and labor). How Singapore became the ideal testing ground for self-driving cars
Artificial Intelligence
The debate about a potential “AI bubble” goes on. The ECB applied the “b-word” to the industry in a recent report, concerned about the concentration of value. But (of course) not everyone agrees. AI stocks: what if this time it really is different?
AI: Agents
AI-agents already drive massive valuations. /dev/agents, the new startup focused on building an “agents operating system” has just reached a valuation of $500m. Why AI agent startup /dev/agents commanded a massive $56M seed round at a $500M valuation
AI: Apps
B2C / Consumer / General
Apple under regulatory pressure to find “AI partners” in China. Authorization to launch “Apple Intelligence” is being conditioned to work with Chinese companies (and this has obvious risks / implications) Apple faces ‘difficult’ process to launch its own AI in China
Google’s core business is under threat from AI apps. Perplexity was targeting this market since the beginning, and now OpenAI and Meta AI are also planning to add search to their AI apps. Googling Is for Old People. That’s a Problem for Google.
Perplexity’s CEO dreams about a $50 “voice-to-voice” AI-powered device. OpenAI once discussed something that, and other startups have already tried. It doesn’t look that easy… Perplexity mulls getting into hardware
Luma, a startup, launches a mobile app to democratize video creation. It is based on the company’s own AI video model (“Dream Machine”) Luma expands Dream Machine AI video model into full creative platform, mobile app
Anthropic claims its Claude model can now replicate a person’s writing style. At least if you have enough digital production to properly train it. Sounds interesting :-) Anthropic says Claude AI can match your unique writing style
B2B
Businesses across all industries are expected to embrace AI to accelerate R&D, under pressure to deliver new products much faster (a third of sales in the next 5 years depending on innovations, according to McKinsey) AI and the R&D revolution
Databricks, a “unicorn” AI startup selling tools for companies to build their own AI models, has reached a $55bn valuation in its latest funding round. Databricks closes in on multibillion funding round at $55 billion valuation to help employees cash out
Anthropic is addressing the challenge to integrate GenAI models into companies’ (often complex) processes, with a new tool that helps models connect to a diversity of datasets Anthropic launches tool to connect AI systems directly to datasets
AI: Robots
There is a race under way to build more powerful autonomous weapons, with obvious risks (misidentification of targets, lack of accountability / control, and potential escalation of conflicts, among others) The Technology for Autonomous Weapons Exists. What Now?
AI: Foundational Models
Debates on the potential “scaling crisis” in AI models and its implications:
Investors are concerned. We talked about this last week. If progress depends less on scale, that could clearly affect Nvidia (which benefits from a race to build larger data centers) Nvidia and the AI boom faces a scaling problem
OpenAI may be seeing an opportunity to create a moat. If “brute force” stops being the way to build better models, and more sophisticated tools emerge, patenting them could be a good idea. OpenAI moves to trademark its o1 'reasoning' models
But “reasoning” algorithms don’t seem exclusive to OpenAI. In China, Alibaba is releasing an open-source model that could have a similar performance as OpenAI’s o1. Alibaba releases an 'open' challenger to OpenAI's o1 reasoning model
Valuations remain high:
SoftBank is back, and they’ve joined the OpenAI party. The exuberant Japanese company / fund has invested $1.5bn in OpenAI, at a valuation of more than $150bn. OpenAI gets new $1.5 billion investment from SoftBank, allowing employees to sell shares in a tender offer
Only 6 “Big Tech” companies have created $8trn of shareholder value since Jan 2022, driven by GenAI. These are the traditional “Big 5” (Alphabet, Meta, Apple, Amazon, Microsoft) plus Nvidia. ChatGPT’s $8 Trillion Birthday Gift to Big Tech
With xAI, Elon Musk wants to catch up with OpenAI. This looks challenging (xAI is far behind in terms of scale and revenue generation, and they’re largely dependent on Elon Musk’s companies), but not impossible (Musk’s companies have access to massive amounts of unique data, and he’s already demonstrated an enormous engineering talent) Inside Elon Musk’s Quest to Beat OpenAI at Its Own Game
Talent seems a bottleneck for European AI leaders. The best known GenAI startup in Europe (Mistral) is opening an office in Palo Alto, looking for best-in-class talent, among other things. Europe’s Mistral expands in Silicon Valley in hunt for AI staff
Nvidia plays the forward-integration game: Nvidia just demonstrated a new AI model built by them, focused on audio, which targets producers of music, films and video games. Nvidia shows AI model that can modify voices, generate novel sounds
AI: Infrastructure
Energy needs remain a key concern for AI data center players. Prometheus Hyperscale, a startup specialized in building data centers for AI, has just hired an ex-BP CEO as the new chairman. Former BP Chief Joins A.I. Data Center Developer
Is distributed computing the future of AI? Quanta Magazine just published a nice recap about distributed computing, that might lead to the debate about this also impacting AI training (currently centralized in huge data centers owned by a few players) What Is Distributed Computing?
AI: Chips
Amazon has a plan to catch up with Nvidia in AI chips. A long Bloomberg article just made this (sort of) “official”. Amazon’s scale and market power with AWS could help with this. But Nvidia fully dominates the AI chip market and has a significant advantage in technology. Amazon’s Moonshot Plan to Rival Nvidia in AI Chips
Intel gets a (lower than expected) subsidy of $7.9bn (just before Trump / Musk enter the house with the big scissors…)
The discount has been justified by Intel’s delays investing in a new production facility in Ohio. The original figure was $8.5bn Washington Curtails Intel’s Chip Grant After Company Stumbles
The company is receiving separate funding through a defense contract. The $3bn contract is focused on build secure facilities producing microchips for U.S. military applications. Intel Gets Up to $7.9 Billion Award for U.S. Chip-Plant Construction
As a condition to receive the money, Intel will not be allowed to sell its manufacturing unit. This (potential sale) was a rumor some time ago, and it would be completely at odds with the rationale for the subsidy. Intel's $7.86 billion subsidy deal restricts sale of its manufacturing unit
Qualcomm seems to be renouncing to buying Intel. This was another rumor, from a few weeks ago, but the improvement of Intel’s financial health (with the subsidies) may have made Qualcomm cool down. Qualcomm’s Takeover Interest in Intel Is Said to Cool
Is Samsung losing competitiveness? Samsung, a world leader in memory chips, is reorganizing its chip business unit, amid concerns about losing technological competitiveness in the field. Samsung shakes up struggling chip business for second time this year
Intelligence Augmentation
Brain-Computer Interfaces
A new Neuralink test: controlling a robotic arm with a neural implant. The company remains focused on healthcare applications, as a first step. Neuralink Plans to Test Whether Its Brain Implant Can Control a Robotic Arm
3. Other things - Economy & Society in the future
Big Tech enters the political debate. The recent result of the US election, and the support for Trump from many Silicon Valley executives, is discussed in this FT opinion column, which claims that Big Tech firms should not be treated at the same level as states. The return of the techno-libertarians
The content-creation powers of AI models require new rules to protect people from “deep fakes”. As often in the past, there is a debate about who should be accountable: technology platforms or people using these platforms. The legal battle against explicit AI deepfakes
LinkedIn is full of AI-Generated content. Obviously it is not surprising that most people have not noticed :-) Yes, That Viral LinkedIn Post You Read Was Probably AI-Generated
People keep questioning the need to learn code, as models are increasingly good at creating code. However, coding probably should still be part of a “basic” curriculum producing versatile people. Do Coding Boot Camps Make Sense in an A.I. World?