In this week’s NFA Podcast, Dr. Nisheta Sachdev and Michiel Frackers unpack the wild week in tech and crypto. Nish covers China’s impactful stimulus, clarifies misconceptions around a Brazilian fintech firm’s Bitcoin adoption, and highlights Vietnam’s new crypto regulatory plans.
Frackers dives into President Trump’s strategic Bitcoin reserve and digital asset “stockpile,” analyzing implications for crypto markets. He also maintains a bullish outlook on Nvidia despite AI sector challenges, while Nish introduces BitTensor (TAO), an innovative decentralized AI network.
Together, they discuss navigating volatility in crypto investments, weighing altcoins versus Bitcoin in an uncertain global landscape.
(NFA podcast is educational and entertaining, not financial advice.)
They also explore current market dynamics, discussing the risks and opportunities in altcoins versus Bitcoin, and share insights on navigating market volatility amid global uncertainties.
The NFA podcast is for educational and entertainment purposes only and is not financial advice.
This week some long-standing trends in the technology sector surfaced more emphatically. First, the growing influence of large technology companies on politics and governance, and second, the question of whether AI models are economically sustainable, with even "life-threatening" market conditions for AI startups such as OpenAI. In the crypto market, it was complete chaos.
Big Tech and Political Influence
Leaders of technology companies are increasingly trying to gain influence over social and political processes. It is no longer just Elon Musk, who last month, as a born South African with Canadian passport and naturalized to American, tried to emerge as an expert of German society and openly called for support for the AfD, now Germany's second party with 21%.
Amazon founder Jeff Bezos is suddenly also a political activist. The surely not typically leftist Wall Street Journal is merciless to Bezos, delicately pointing out that Amazon recently paid $40 million for an authorized documentary on Melania Trump entirely by accident, "three times as much as any other bid." The WSJ continues:
"There were the flattering tweets with which Bezos applauded Trump's victory and his prominent presence alongside other tech leaders at the inauguration-Zuckerberg, Musk, Tim Cook of Apple and Sundar Pichai of Google, among others. Their seats on stage, directly behind Trump and in front of the cabinet, could be interpreted in two ways: as a historic gathering of new economic and tech powerhouses showing their support for the new administration, or as a hostage video of billionaires held captive by a menacing strongman.
Another shock followed this week when Bezos announced that the opinion pages of the Washington Post (bought by Bezos in 2013, MF) would henceforth be devoted to defending the principles of "personal liberty and free markets." This shift to the right led to the resignation of David Shipley, the section's editor-in-chief. Critics condemned the move as an attempt to suppress liberal opposition and criticism of Trump, while others noted that such views are widely represented in other publications."
Now the formerly media-shy Palantir CEO Alex Karp is surfacing, openly advocating a system in which democratically elected leaders are replaced by an AI-driven bureaucracy in a new book. His argument is that AI can make decisions more efficiently than human administrators. Bloomberg, also rarely accused of leftist leanings, judged Karp's book harshly:
"It’s a major complaint of the authors of The Technological Republic (Crown Currency, Feb. 18) that people today shrink from saying what they think. Too many of us, they insist, give mealy-mouthed, wishy-washy answers when asked. We have become uncomfortable with making moral and aesthetic judgments, they say.
I agree, and I'm going to break the taboos. The Technological Republic is a terrible book: badly written, boring and-when the ideas can be picked up between the jargon, clichés and repetitions-full of bad ideas ranging from questionable to reprehensible and disturbing. This book is abysmal in both form and content. It sketches a dark and depressing future."
Not a review that will make the back cover of the book, which should be seen primarily as a brochure for Palantir.
Big Tech cries out for government intervention
The founders of Big Tech companies Meta, Amazon and Tesla position themselves as indispensable to economic and social progress and argue that their success is the result of technological and market superiority. At the same time, Meta, with click-happy algorithms, is largely responsible for sharing disinformation and sowing social discord, Amazon is the pinnacle of consumerism with a history of sad working conditions, and Tesla enjoyed as much as $38 billion in government subsidies. Thus the Washington Post in an article that still dared to spout some criticism of Tesla's success based ons o-called free market forces.
With the imminent introduction of artificial general intelligence (AGI), technology is playing an increasing factor in society. Whereas at the end of last century politicians were often scornful of what was mostly described as nothing more than automation, it is only now being understood that the ongoing digitization of the world is spilling over into AGI systems that, without democratic control, run the risk of a small number of technology company leaders amassing disproportionate power.
Big Tech's interests do not parallel societal values such as privacy, democracy and public participation. The tech-bros think first and foremost about quarterly earnings.
The focus at OpenAI is not yet on making revenue. Just try updating your credit card.
The missing business model of AI
Speaking of quarterly earnings pressure; within the technology sector, there is an ongoing debate about the effectiveness of different business models. A major issue is always whether companies should bundle technologies or offer them as separate products. An example of this is Microsoft's acquisition of Skype for $8.5 billion, a sum that was probably never recouped due to ambiguity about the pricing model and lack of integration into MS Office.
Om Malik delightfully cynically concludes at the announcement that Microsoft is shutting down Skype: "Skype's demise is a good lesson in how ineffective middle management can destroy good acquisitions. I have never met a Skype manager on Microsoft's side who had any imagination. Most were such "drones" that next to them even a red clay brick would come across as a genius work of art.
Microsoft Teams is a terrible product-and I hate using it. In the simplest terms, Teams is the perfect summary of a bureaucratic, outdated and archaic 50-year-old company trying to reinvent itself as a leader in AI."
If the relatively straight forward product Skype, which already had millions of users worldwide, is so complex to be profitably exploited by a giant like Microsoft, it will be especially interesting to see if opaque billion-dollar investments in AI are ever going to deliver the intended returns. The quarterly earnings reports of Microsoft, Meta and Amazon are being increasingly scrutinized by analysts for their spending on AI. Although these companies are investing tens of billions in AI, Nvidia is the only one consistently benefiting from this trend, and it remains unclear whether the Big Tech companies will ever turn a profit on their AI investments.
Big problems for AI startups
For leading AI startups, it's all hands on deck this year. In recent weeks, Grok 3 (from x.ai, owned by Elon Musk), Claude 3.7 Sonnet (Anthropic) and ChatGPT 4.5 (OpenAI) have been launched. Analyses show doubts about the quality and efficiency of this latest generation of AI applications. "It's a lemon," headlines Ars Technica about ChatGPT-4.5.
Ethan Mollick, on the other hand, remains positive about advances in AI: "The intelligence of AI models is increasing, and costs are falling." But people like Mollick are now a minority among investors.
The problem for AI startups such as Elon Musk's X.AI, Sam Altman's OpenAI and competitors such as Anthropic (with Claude) and Mistral, is the increasing doubt about real technological advances relative to rising development and operating costs. Investors are increasingly questioning the long-term profitability of AI companies.
The discussion is no longer about the beliefs of AI proponents, who regard AI development almost as a religion, or the objections of the non-believers, let me call them "AI-theists," but about economic reality. The key question is not whether AI can continue to grow and fundamentally improve, but whether it can do so profitably. In the investment world, there are serious doubts about two things:
Whether the costs of AI development will ever outweigh the benefits and whether a sustainable business model is possible in which companies like OpenAI become profitable,
Whether AI models actually deliver the expected improvements that allow the end users of AI applications, the customers of OpenAI, Microsoft, Google etc, to operate more cost-effectively.
For OpenAI, this is an urgent problem. Companies like Microsoft, Google, Meta, Oracle and Salesforce invest tens of billions in AI every year, but can absorb losses with profits from other activities. OpenAI, on the other hand, is completely dependent on AI and remains heavily loss-making.
Legendary investor Vinod Khosla, among other early backers of OpenAI, openly says that he expects most investments in AI to be loss-making. Of course, that does not apply to his own investment in OpenAI, because he was in it so early that any sale of OpenAI will be a hit for Khosla.
The "disaster month" for Nvidia, compared to the Dow, S&P and Nasdaq Composite...
For now, Nvidia is benefiting from the confidence within Big Tech that increasingly powerful and expensive chips are the solution. The company again achieved record results, although profit margins are declining. Gross margin nevertheless remained at an impressive 72%. Not surprisingly, Nvidia rose another 4% on Friday and still ended February with a 7% gain, while the major stock market indices recorded losses. Barron's therefore half-jokingly called Nvidia a value stock.
DeepSeek with bad news for OpenAI
DeepSeek, OpenAI's Chinese nemesis, claimed yesterday on X to have a much more efficient cost structure: "Our cost-benefit ratio is 545 percent." A more detailed explanation later followed on GitHub, to which Techcrunch sharply concluded:
"The company (DeepSeek, MF) wrote that if it looked at the usage of its V3 and R1 models over a 24-hour period, and if all that usage had been billed at the R1 prices, DeepSeek would have already generated $562,027 in daily revenue. At the same time, the cost of leasing the required GPUs (graphics processing units) would have been only $87,072.
The company admitted that actual revenue is significantly lower for several reasons, including discounts during nighttime hours, lower prices for V3 and the fact that only a portion of services are monetized, while access via web and app remains free.
If the app and website were not free and other discounts did not exist, usage would presumably be much lower. Therefore, these calculations seem largely speculative-more an indication of potential future profit margins than a realistic representation of DeepSeek's current financial situation.
But the company is sharing these numbers amidst broader debates about AI’s cost and potential profitability."
Surely every investor now has the idea that DeepSeek has a much better chance of becoming profitable than OpenAI, which no longer has a substantial technological edge, dubbed in Silicon Valley as a "moat," nor does it have the financial capabilities needed to eliminate the competition. Consider how, for example, Mark Zuckerberg once bought fast-growing competitor Instagram with Meta. Sam Altman does not have that option.
The next few months will be decisive for OpenAI. The company desperately needs capital. It is now at the mercy of Softbank's Masayoshi Son, who is already raising $16 billion in loans with his cap in hand, indicating that the industry's biggest financiers are cautious. Even if all the money Softbank is now raising in loans were to go into OpenAI, which is doubted, the question is how far OpenAI will get with that money.
Savior from Abu Dhabi or a mirage?
Another possible investor is Tahnoon bin Zayed al Nahyan, an influential Abu Dhabi financier, irreverently dubbed the "Spy Sheikh" by the Wall Street Journal . As manager of several Abu Dhabi sovereign wealth funds, including MGX, he could turn out to be the financial decision maker on OpenAI's fate.
The question, meanwhile, is whether OpenAI can survive another year without rapid funding. If there is no urgent injection of billions, a takeover lurks. Microsoft already owns 49% of the shares, is the main provider of cloud infrastructure and could cough up as much as $100 billion to buy out the existing shareholders in OpenAI. That's a very different reality for OpenAI CEO Sam Altman than it was a few months ago, when he still thought he could raise $30 billion against just 10 percent of the shares.
Dr. Sachdev lost the bet on price predictions, but did buy crypto.
Blood bath in the crypto market
In episode 5 of the NFA Podcast (for Nish, Frackers and Anyone Else, and, of course, for Not Financial Advice), Nish eats an Indian green chili because she had lost the bet on a rising or falling market. She was dressed in red to symbolize the carnage in the crypto market.
In more relevant news: we discussed the drop in new token launches on Pumpfun, BlackRock's Bitcoin sales and the SEC's ruling that meme coins are not securities. That would normally be positive news for the speculative crypto market, but it didn't matter last week: almost everything plummeted.
The Bybit hack, which was linked to North Korean attackers, was also discussed in detail and showed vulnerabilities in multi-signature transactions. Finally, we discussed whether Bitcoin will still reach an all-time high this year and how its correlation with traditional markets is developing. We agreed, which is not the intent of the format.
Episode 5 of the NFA Podcast. "Crypto bloodbath, the Bybit hack fall out and will Bitcoin 'go rogue' to hit an ATH?" is available to listen to now or watch here on YouTube and also here on Spotify.
You can subscribe to the special NFA Podcast newsletter, which will keep you informed of each new episode, here on LinkedIn. Thanks for your interest and see you next week!
Microsoft has announced it has reached a significant milestone in quantum computing. The tech company says it has created "topological qubits," a technological innovation that could be crucial to realizing stable and scalable quantum computing.
Quantum computing has long been considered the holy grail for computational power and data analysis, but the technology faces stability problems. Once quantum computing works at scale, it would make current encryption obsolete, including cryptographic keys for blockchain. That means everyone needs to think about alternative storage methods for their cryptoassets now, before quantum computing becomes a reality. In North Korea, they can't wait.
Former OpenAI leaders with competing AI companies
The AI sector remains in flux as former OpenAI executives raise billions for competing projects. It was previously known that ex-OpenAI Chief Scientist Ilya Sutskever founded Safe Superintelligence(SSI), which is in negotiations to obtain funding at a valuation of $30 billion.
Now it appears that former OpenAI CTO Mira Murati is also raising money for her AI startup Thinking Machines Lab, with a team largely made up of employees recruited from OpenAI and Anthropic. It is just not known what valuation Murati has in mind, but she will certainly be looking with an oblique eye at what former colleague Sutskever is getting done. Like SSI, Thinking Machines Lab has a one-page website with text on it in very small print. The most high technology companies make Web sites that look like they did in 1993.
President Xi Jinping brings together Chinese tech leaders
For the first time in six years, Chinese President Xi Jinpinghas met with top members of China's technology sector, including Jack Ma, the co-founder of Alibaba. The meeting marks a striking change of direction after years of the Chinese government cracking down on tech companies with strict regulatory measures and high fines.
The meeting is a sign that Chinese leaders recognize how crucial technology is to the economy. Jack Ma's return to the spotlight indicates that the government wants to give companies like Alibaba more freedom again, presumably to spur innovation and growth in a flare-up in international competition.
Thursday, Feb. 20, was a special day of trading for Alibaba, Palantir and Unity
Alibaba and Unity rise, Palantir falls sharply
The meeting was met with cheers from investors and Alibaba shares rose sharply. On the same day, Thursday, Palantir shares took a huge hit because President Trump is expected to cut defense spending, which could negatively impact Palantir.
Another notable stock was software maker Unity, which, although still loss-making, is being praised by investors for its change in direction, supplying not only software for computer games but also car companies such as Toyota. Thus, Thursday, Feb. 20, became a very special day for these three companies.
Interestingly, the share price of U.S. largest crypto exchange Coinbase, barely reacted to the news that the SEC, under the reign of President Trump, has halted the years-long lawsuit against Coinbase.
Formula 1 full of crypto sponsors
The Formula One season is about to start and was announced big time with an event at London's O2 Arena. It was noticed that many teams are sponsored big by crypto companies:
Aston Martin is sponsored by Coinbase, with quite a fuss being made about the fact that payment is made in the stable token USDC. That's a digital dollar, boy.
Red Bull Racing has had crypto exchange Gate.io on the back wing since this season, replacing Bybit. But that one made plenty of headlines later in the week.
McLaren has had OKX as a sponsor since 2022. OKX appears large on the side of the car, among other things.
Williams: American crypto exchange Kraken is a sponsor of Williams for the third year.
Stake F1 Team: Sauber, which will become Audi next year, has even sold the name to the crypto gambling site Stake.
This and more Nish and I discussed in episode 4 of the NFA Podcast (Not Financial Advice, or Nish, Frackers and Others), noting especially her extensive analysis of the Solana platform, which goes beyond just discussing the price of the day - as I like to do when her token favorites drop.
Despite the rise of competitors such as DeepSeek, OpenAI has not lost any traffic, unexpectedly growing faster than ever and leaving all competitors behind. The question is why OpenAI nevertheless needs $40 billion. We discuss this and much more in the new weekly podcast we are launching this weekend: The NFA Podcast, with co-presenter Dr. Nisheta Sachdev. The first two episodes have been online since this morning.
Also in this newsletter, we look at the final breakthrough of Palantir and the mysterious company SSI, which does not appear to have a product out, yet is valued at $20 billion.
Nish discusses the tokens PAIN and It Will Go UP, I talk about OpenAI's funding and the success of Palantir
New weekly podcast: NFA
In episode 1(viewable on YouTube here or on Spotify here), Nisheta and I introduce ourselves, talk about how we got into the tech and crypto world, and discuss the latest trends in web3 and tech, with a special focus on AI. According to Nisheta, the link between AI and crypto represents the definitive breakthrough of blockchain in the consumer market.
In episode 2(viewable on YouTube here or on Spotify here), we dive deeper into specific crypto projects such as PAIN, the token that last week raised $40 million in an instant and returned money to investors. We also discuss Palantir's extraordinary quarterly results and show footage of CEO Alex Karp, who in a jolly mood declares war on opponents of America and the West....
The name NFA was coined by Nisheta and stands for "Nish, Frackers & Anon" or "Anyone Else," because we hope to have interesting guests, but it also stands for "Not Financial Advice": we absolutely do not want to give financial advice.
Weekly, Nish and I discuss the most notable developments in crypto and tech, without trying to be pedantic or give investment advice. We share our thoughts, opinions and things we find interesting, such as interesting tokens and stocks that have come on our radar.
Whether you are a seasoned tech and crypto expert or an enthusiastic beginner, our goal with the NFA Podcast is to create a place to explore the market together and share experiences. All suggestions and feedback are welcome.
Who is Dr. Nisheta Sachdev?
Dr. Nisheta Sachdev, also known as "The Crypto Dentist," is a prominent figure in the blockchain world. Her inclusion in the "40 under 40" list underscored her breakthrough in this fast-changing world, where Nisheta stands out for her lucid analyses.
Nisheta Sachdev, blockchain by day, dentist by night - or vice versa
Nisheta came into contact with crypto in 2018 and worked full-time in the sector from 2020, when she and her sister Nikita started the soon-to-be leading crypto marketing company Luna PR in Dubai. Within the blockchain sector, Nisheta became a passionate advocate for technologies such as cryptocurrencies, NFTs, the metaverse and AI. Nish has a unique ability to articulate complex concepts simply, making her a popular and respected speaker at international conferences.
Despite her keen interest in the crypto and technology sectors, Nish recently answered a higher calling and resumed her studies in medicine, specializing in reconstructive surgery of patients who have lost parts of their jaw and face, mostly due to forms of oral cancer. For those interested in learning more about Nisheta, I recommend this recent interview with her on YouTube.
ChatGPT resurrected from never being gone
After a period of stagnation, ChatGPT has returned to strong growth, leaving competitors such as Bing, Gemini, Claude and Perplexity far behind. They are "Boom Times For ChatGPT."
By January 2025, ChatGPT reached 3.8 billion visits on desktop and mobile, more than double that of Bing and well above the rest. This marks a major turnaround after more than a year of stagnation. The upturn comes at a crucial time, when OpenAI is being challenged by DeepSeek.
DeepSeek, which rapidly gained popularity at the end of January, achieved 49 million visits in just one day, a third of ChatGPT's audience, despite the company being barely known outside China. This shows that new competitors can gain ground at lightning speed.
OpenAI benefits from ChatGPT's strong brand name. To reinforce this position, it is launching its first Super Bowl commercial today, while Google is doing the same for Gemini. If OpenAI succeeds in positioning ChatGPT as the standard for AI, it can maintain its lead even as chatbots become an everyday product worldwide.
Which competitors?
ChatGPT is way ahead of the competition. In January 2025, ChatGPT had 3.8 billion visits, compared with 1.8 billion for Microsoft Bing, 267 million for Gemini, 99.5 million for Perplexity and 76.8 million for Claude. Although these figures reflect Web traffic only, they show that OpenAI has built a huge lead. It's especially painful for Google, which just can't seem to "redirect" traffic from its search engine to Gemini.
OpenAI's quest for tens of billions
Journalist and investor M.G. Siegler, who is particularly prolific as a blogger again these days, explored why OpenAI needs $40 billion and how the relationship between OpenAI, Microsoft and Nvidia fits together in the excellent piece "The IPOpenAI."
As discussed last week, OpenAI plans to raise $40 billion, primarily to fund the massive computing power needed to train and run its AI models. Although there is a battle going on for talent in the AI sector which means staffing is also not cheap, by far the largest portion of this amount is going toward computing infrastructure such as power-hungry servers and data centers.
OpenAI in battle with Big Tech
The Big Tech companies, with Microsoft, Meta and Amazon leading the way, can invest tens of billions of dollars annually in their existing cloud infrastructures, funded from their huge profits. To compete, OpenAI must also raise tens of billions and has now ended up with SoftBank and sovereign wealth funds, mostly from the Gulf region.
To put it in perspective, according to Reuters, OpenAI expects an annual revenue of nearly $12 billion in 2025 at a loss of possibly $15 billion, at least if margins do not improve. It's going up against Meta with $200 billion in annual revenue, Microsoft with $250 billion, Alphabet with $300 billion and Amazon with even more than $500 billion in revenue this year. Even Oracle ($58 billion) and Salesforce ($38 billion) are profitable giants compared to the heavily loss-making OpenAi.
OpenAI raises more than historical IPOs
If OpenAI manages to raise $40 billion, it would be the largest amount ever raised by a company, even more than oil giant Saudi Aramco's $30 billion in 2019. By comparison, Visa raised nearly $20 billion in 2008 and Facebook $16 billion in 2012. But Facebook (now called Meta) did so at a valuation of under $100 billion, or less than a third of what OpenAI now considers itself worth.
The trend in the IPO market is that the biggest promising companies are choosing to remain private, thanks to the abundance of available capital. OpenAI's current funding round is an example of this shift, with companies raising huge sums of money without going public.
It is unlikely that OpenAI will succeed in becoming profitable any time soon, and then it will be a hell of a job next year to find parties willing to step in at a valuation of say $500 billion. It won't be possible to go much lower if SoftBank comes in now at $300 billion. How many investors will dare to explain to their shareholders next year that they are investing roughly $50 billion in OpenAI and not even getting ten percent of the shares in return?
Nvidia always wins
If SoftBank invests the $40 billion OpenAI wants, OpenAI will pay most of that money to Microsoft, on whose cloud services (Azure) it largely depends. In turn, Microsoft, with hat in hand and checkbook in its inside pocket, will be knocking on the door of Nvidia, which supplies the necessary chips and data center infrastructure.
In the AI world, it's simple: when it rains money at the top, it eventually pours down at Nvidia. SoftBank's pennies will end up, via OpenAI and Microsoft, at Nvidia..
This is precisely why I wrote last week that it makes no sense to expect Nvidia's revenue growth and profit margin to decline in the coming years; all doomsday scenarios from Wall Street notwithstanding. Microsoft is investing $85 billion in AI this year and Meta $65 billion. There are reportedly countries, not companies, that will also place such hefty orders with Nvida over several years.
The Big Tech companies have no choice but to order what Nvidia can deliver, because missing the boat in AI could mean the end of the supremacy of Microsoft, Google, Oracle, Salesforce and Amazon Web Services. But note that this is No Financial Advice!
Palantir's share price already rose 47% this year, especially striking compared to mere share price declines at Nvidia, Apple, Microsoft and Google
Winner of 2025 so far: Palantir
An exception in the software sector struggling with the rise of AI is Palantir, which I have written about many times. The data analytics company remains secretive, but it seems increasingly likely that Palantir is one of the few software companies in the world that, through the use of AI, manages to both improve its software and reduce operating costs at the same time.
Despite selling relatively little Palantir software in Europe due to privacy concerns, the stock has risen 350% in a year. This week Palantir announced impressive financial results, again leading to a hefty rise in its share price: 22% in one day.
For the fourth quarter of 2024, the company reported revenue growth of 36% over the previous year, with total sales of $827.5 million. The U.S. market contributed significantly to this, growing 52% year over year. Since 2009, the company has secured more than $2.7 billion in U.S. government contracts, and it is likely that under President Trump, the U.S. government is going to take an even much larger cut from Palantir.
Morgan Stanley raised the rating for Palantir from "underweight" to "equal weight"and adjusted the price target from $60 to $95, referring to the strong quarterly results and positive outlook for 2025. Nonetheless, concerns remain about the high valuation of the stock, which sits at a stratospheric price-to-earnings ratio nearing 600. Palantir has been part of the S&P 500 and the Nasdaq 100 since last year, so it will also have a stronger impact on the prices of index funds and ETFs.
Palantir has been led for years by quirky co-founder Alex Karp, who said confidently at the presentation of the quarterly figures that Palantir likes to help governments and goes far in doing so: 'when it's necessary to scare enemies and on occasion, kill them..'
Andreessen Horowitz invests (de)centrally
Due to the success of CEOs like Karp and before him people like Bill Gates, Steve Jobs and Elon Musk, the perception has emerged in the tech world that centrally managed companies are the most successful and best at innovating. In this excellent piece, Miles Jennings of investment firm Andreessen Horowitz argues the opposite. Yet even Andreessen Horowitz still invests the majority of its billion-dollar budget in traditional centrally organized companies, especially in the AI world.
No product, 1-page website: worth $20 billion
Safe Superintelligence, an unusual AI startup with a one-page website co-founded last year by former chief scientist and co-founder of OpenAI Ilya Sutskever, is fund raising at a valuation of at least $20 billion. Small detail: the company is not yet making revenue and apparently does not yet have a product.
Sutskever's track record and SSI's unique approach are enough reason for great interest among investors. I previously wrote about the extraordinary scientist Sutskever, who became famous in the tech world mainly because Elon Musk and Google founder Larry Page broke off their friendship over a dispute over whose company Sutskever would join.
Musk can get into an argument walking into an empty house, but his fight with Page showed that Sutskever must have special qualities. Musk said of this in Lex Fridman's podcast, "It was mainly Demis Hassabis (ex-founder DeepMind, now head of AI at Microsoft, MF) on one side and I on the other, both trying to recruit Ilya, and Ilya hesitated. In the end, he agreed to join OpenAI. That was one of the hardest recruiting battles I've ever experienced, but that was really the key to OpenAI's success."
The new funding round would quadruple SSI's valuation from the previous round less than six months ago (!), when the company was still valued at $5 billion and raised $1 billion from five investors, including Sequoia Capital, DST Global and, of course, Andreessen Horowitz. There are apparently enough investors in Silicon Valley who would rather buy a reasonable stake in SSI at a $20 billion valuation, than be fobbed off with a smurf sized share in OpenAI at a $300 billion valuation.
How DeepSeek would like the world to think about the youthful team. Image created with Midjourney.
It was the week of DeepSeek's CEO Liang Wenfeng, who seemed to appear out of nowhere to scare the hell out of everyone from Silicon Valley to Washington to Wall Street.
Apparently, not everyone has noticed that China is making the leap from an agricultural to a post-industrial society in record time. What chuckles there must have been in Beijing and Shanghai when Chinese New Year was celebrated last week.
Last week I wrote that Silicon Valley was rudely awakened by DeepSeek, and on Tuesday I added that Wall Street had overreacted. Today an attempt to chart the winners and losers, short- and long-term, of the rise of DeepSeek.
Who is Liang Wenfeng?
But first: who is Liang Wenfeng, the founder and CEO of DeepSeek? What is special about Wenfeng, as a startup founder, is his background as the founder of a hedge fund: High Flyer.
"When we first met him, he was this very nerdy guy with a terrible hairstyle talking about building a 10,000-chip cluster to train his own models. We didn’t take him seriously" one of Liang's business partners told the Financial Times.
During his time at High Flyer, Liang began buying Nvidia equipment and learned the various ways to develop algorithms for AI applications, lessons he now applies at DeepSeek. More remarkably, DeepSeek's sudden success is driven by Gen Z newcomers from diverse backgrounds. Liang likes originality and creativity from young smart people and values experience a lot less.
Liang also talked about hiring literature buffs on the engineering teams to refine DeepSeek's AI models. "Everyone has their own unique path and brings their own ideas, so there's no need to direct them." This is especially interesting to read in the week that Mark Zuckerberg boasts that he is getting rid of all diversity programs at Meta, in an effort to appease the Trump administration.
OpenAI worth $300 billion after all?
According to the Wall Street Journal, Japan's SoftBank would lead a $40 billion investment round in the ChatGPT maker, part of which is to be spent on its Stargate AI infrastructure project. With a valuation of $300 billion, OpenAI would become the second most valuable startup in the world, behind Elon Musk's SpaceX, the major rival of OpenAI CEO Sam Altman.
It would be downright amazing if Altman manages to raise money for his money losing company at that stratospheric valuation, in the week when its vision and technological architecture are being doubted worldwide. But let us not overestimate SoftBank: it is the same club and the same man, Masayoshi Son, who burned tens of billions in WeWork; all the way to bankruptcy. The question is: why won't anyone but SoftBank step in at this valuation?
Is Stargate science fiction?
Both OpenAI and SoftBank have declared they will invest tens of billions in Stargate, the $500 billion budgeted AI infrastructure project that is supposed to seal American hegemony in technology. The crazy thing is that OpenAI doesn't have that money at all, and neither does SoftBank. So when SoftBank invests in OpenAI, which thereby invests in Stargate, it's basically filling one hole with another one.
The Verge published a lucid analysis of the Stargate project. If Stargate fails, it would not simply be the end of a startup. It would be an expensive reality check for an entire industry that claims to transform the world through pure computing power.
Altman likes to present himself as the protagonist in a classic science fiction story: the visionary who promises to transform society through technological power.
In say a year, we will know whether Stargate was the beginning of America's AI revolution, or just a techno-optimistic fantasy that could not survive in the real world.
DeepSeek's actual costs
Then to a much-discussed topic: the costs allegedly incurred by DeepSeek to develop the acclaimed R1 model. The wildest stories are circulating about this, while DeepSeek itself has been fairly transparent about it:
"Finally, we again highlight the economic training cost of DeepSeek-V3, as summarized in Table 1, achieved by our optimized co-designs of algorithms, frameworks and hardware.
During the pre-training phase, training DeepSeek-V3 on every trillion tokens requires only 180K H800 GPU hours, or 3.7 days on our cluster with 2048 H800 GPUs. This completes our pre-training phase in less than two months and takes a total of 2.664M GPU hours. Combined with 119K GPU hours for context length extension and 5K GPU hours for post-training, DeepSeek-V3 costs a total of only 2.788M GPU hours for full training.
If we assume that the rental cost of an H800 GPU is $2 per GPU hour, our total training cost is only $5.576M. Please note that the above costs include only the official training of DeepSeek-V3 and not the costs associated with previous research and tear-down tests of architectures, algorithms or data."
I highlighted the crucial part: all previous costs are not included in the cost calculation. It's like calculating the cost of a bodybuilder's meals on competition day without including how many meals it took to get to the competition.
Cheaper AI: who benefits?
Even more interesting than the cost aspect, DeepSeek offers the ability to install the model locally and develop on it. Microsoft CEO Satya Nadella pointed directly to Jevon's Paradox.
In short, precisely because of the reduced cost, the use of an innovationwill increase. It looks like Nadella is going to be right about that. In the long run, the "commoditization" of AI models and cheaper inference as demonstrated by DeepSeek will benefit Big Tech. Microsoft, for example, needs to spend less on data centers and GPUs, while benefiting from increased AI utilization through lower inference costs.
Amazon is also a big winner: AWS has not developed its own high-quality AI model, but that doesn't matter when there are high-quality open-source models available that it can offer at much lower cost.
Apple also benefits
Drastically reduced memory requirements for inference make AI on iPhones much more feasible. Apple Silicon uses a unified memory architecture, with the CPU, GPU and NPU (neural processing unit) accessing a shared memory pool, argues Stratechery in an excellent piece. This effectively gives Apple's hardware the best consumer chip for inference. Nvidia's gaming GPUs, for example, reach a maximum of 32GB of VRAM, while Apple's chips support up to 192GB of RAM.
Meta the biggest winner
AI is central to Meta's long-term strategy, and one of the biggest obstacles to date has been the high cost of inference. If inference and training become much cheaper, Meta can accelerate and expand its AI-driven business model more efficiently.
Sensibly, Zuckerberg has reportedly set up several war rooms to determine how Meta will react to the introduction of DeepSeek. Whereas in the short term DeepSeek is thought to be a threat to Meta's AI strategy with its Llama LLM, a structural reduction in AI development costs will actually lead to a huge advantage for Meta, which is on track to invest $65 billion in AI development this year alone.
Most of that is spent on hardware and data centers. If that kind of investment can be minimized by imitating DeepSeek's approach, Meta will see its net profits increase substantially without weakening its competitive position.
Google the loser?
While Google also benefits from lower costs, any change from the current status quo is likely to be a net detriment to Google. Every search in OpenAI, DeepSeek or a Meta agent, comes at the expense of a search on Google's search engine.
Despite all its efforts and hundreds of acquisitions over the last few decades, Google still depends largely on the search engine for revenue and profits. It remains to be seen whether Google will succeed in "redirecting" that traffic from the AI agents and chatbots the world so eagerly uses, back to Google's AI tools.
Nvidia not defeated by DeepSeek
Despite DeepSeek's breakthrough, Nvidia has two moats, according to Stratechery:
CUDA is the preferred programming language for anyone developing these models, and CUDA works only on Nvidia chips.
Nvidia has a huge leadwhen it comes to the ability to combine multiple chips into one large virtual GPU.
These two lines of defense reinforce each other. As mentioned earlier, if DeepSeek had had access to H100s, they probably would have used a larger cluster to train their model simply because it was the easiest option. The fact that they did not and were limited by bandwidth dictated many of their decisions in terms of model architecture and training infrastructure.
DeepSeek has shown that there is an alternative: heavy optimization can achieve impressive results on weaker hardware and with lower memory bandwidth. So paying more to Nvidia is not the only way to develop better models.
However, there are three factors that still work in Nvidia's favor.
First, how powerful would DeepSeek's approach be if applied to H100s or the upcoming GB100s? Just because they have found a more efficient way to use computing power does not mean that more computing power would not be useful.
Second, lower inference costs are likely to lead to wider use of AI in the long run. Microsoft CEO Satya Nadella recently confirmed this in his late-night tweet about Jevon's paradox.
Third, reasoning models such as R1 and o1 derive their superior performance from using more computing power. As long as AI's strength and capabilities depend on more computing power, Nvidia will continue to benefit.
Also, with a larger market, Nvidia will benefit from revenue growth in cheaper chips, although it will be hampered in that market by competitors such as AMD.
My subjective "Spotlight on AI" basket took relatively few hits last month.
DeepSeek thought 28 seconds about a hot dog
Joanna Stern of the Wall Street Journal did a funny test of DeepSeek and discovered how it differs from OpenAI's ChatGPT and Anthropic's Claude. Unlike OpenAI's reasoning models, DeepSeek shows its full thought process. When asked if a hot dog is a sandwich, DeepSeek thought about it for 28 seconds and responded with: "First, I need to understand what the definition of a sandwich is." It illustrates that there is no specific form of AI that works best for all issues.
The advance of AI throughout society is irreversible and with DeepSeek's approach, which will be copied frequently, the market will only grow larger. Therefore, despite all the doom-and-gloom news last week on Wall Street, it is fascinating that over the entire month of January, the performance in what I consider to be AI stocks has been better than one would expect.
ARM's 29% rise is remarkable and is largely based on ARM's participation in Stargate. The remarkable thing is that SoftBank owns ARM and therefore there is a good chance that Masayoshi Son will use the shares in ARM as collateral when raising loans, which SoftBank can then use to pay for investments in OpenAI and in Stargate. Time will tell whether this approach leads to a skyscraper, or a house of cards.
This is how the main parties of the DeepSeek crash closed on Wall Street yesterday
What did America's tech billionaires buy from Trump?
President Trump has often expressed hostility toward major technology companies and their leaders, calling Facebook an "enemy of the people" and labeling Jeff Bezos as "Jeff Bozo," for example. Yet these gentlemen were in the front row at the inauguration, having lapped up significant sums of money. This was obviously no coincidence, and the technology sector wants something back from Trump soon. Bloomberg looked at each of them and mapped out what they each want to accomplish.
As we take stock of the performance of Big Tech stocks in the month of January at the end of the second week in Trump's second reign, it appears that the short-term results are not yet what Trump's new tech pals had hoped for. Despite all of Trump's presidential decrees and appointments, stock market results have been rather mixed, to say the least.
What is particularly striking is that investors are sharply divided over the tech sector as a whole. Meta rose mainly due to good quarterly earnings, but how could Microsoft fall while Google rose? Did Apple fall in January due to the possibility of a trade war with China? It is strange that the financial media was mostly focused on last week's results and ignored what happened in terms of price swings earlier in the month. Consider, for example, Palantir, up nearly 10% in January and already up 385% in the last year.
Huang at Trump, Liang at Li Qiang
President Trump and Nvidia CEO Jensen Huang discussed the impact of DeepSeek and possible restrictions on AI chip exports to China during a meeting at the White House on Friday. Huang will certainly have been thinking about the possible impact on Nvidia's stock price.
DeepSeek's Liang Wenfeng also met with an important politician this week: as the sole representative of the AI industry, he met with Premier Li Qiang, China's second most powerful man. Both meetings underscore the importance of technology to economic power in the new world order defined in part by AI.
Palantir CEO Alex Karp told CNBC that the rise of DeepSeek is a sign that the U.S. needs to work faster to develop advanced AI. "Technology is not necessarily good and can pose threats in the hands of adversaries. We need to recognize that, but that also means we need to run harder, go faster and make a national effort."
Boring: success begins with homework
Europe is no longer a consideration in the geopolitical shuffling between continents; how can it be, with so much talent among half a billion people?
Malaysian comedian Ronny Chieng summed up the West's problem perfectly: people are willing to die for their country, but they don't want to do homework for it. Chieng is talking about America, but it applies just as well to Europe.
Among all the news on AI, you would almost forget that in two weeks, Apple is expected to announce its first new device since the Apple Watch in 2015. It will be mixed reality glasses that could cost as much as $3,000 (three thousand!). All the omens are that the Apple Vision Pro will be a flop - a flop by Apple standards, that is. But that's not a bad thing at all. At least Apple is trying to develop something new again, and that's better than unimaginatively buying back its own shares for hundreds of billions, as it has in recent years. Apologies in advance for this long speech below, but I felt like it.
Beautiful render of the Apple headset by Marcus Kane, creator of beautiful renders and lucid analysis. Another stunning render was created by Denis Lukianenko.
What can Apple's ski goggles do?
The glasses, likely to be called Apple Reality, are expected to feature an internal screen for virtual reality, while outward-facing cameras will allow users to view the real world inside the headset with augmented reality overlays. This combination is known as mixed reality, or blended reality.
Angry tongues claim that the first pair of glasses Apple is expected to release late this year is a poor compromise between these two visions, with external video cameras thus capturing the environment and displaying it on the screen in front of the eyes when users switch the headset from VR mode to AR mode. In technical terms, video pass-through.
What does this mean in normal Dutch? Bassie (of Adriaan) would say that the glasses allow you to see the inside of your eyes. Because the first function is that of "regular" virtual reality glasses, giving you the experience of standing on stage next to Beyonce or at the top of Mount Everest. But the second function offers the user the chance to simply look through the glasses at the real world, in which mom asks if you've taken out the trash yet. Possibly supplemented by information projected over that sad reality (augmented reality), showing, for example, that the garbage truck won't arrive until the day after tomorrow and you can continue gaming in peace.
But for how many people are the potential applications actually relevant, informative or fun? The Wall Street Journal logically headlined: Can it be more than a nerd helmet? The expected primary applications of the Apple headset are FaceTime, Apple Fitness+ and gaming. Those are applications that already work well on computers, iPhones, gaming consoles and smart watches, such as the underrated Apple Watch.
Is the Apple headset becoming serious business?
Among all the rumors about the Apple headset, I enjoyed the garbled listing of all the materials purchased by Apple, called the Bill of Materials. If correct, Apple has already spent $1500 on component purchases. More details here.
These expensive components are one reason that the much made comparison between the Apple Headset and the Apple iPod is completely flawed. Optimists point out that when the iPod was introduced in 2003, only 3 million MP3 players had been sold worldwide, compared to over 8 million VR headsets today. In other words, Apple is now entering a much larger, more mature market than back then with the iPod.
This ignores the fact that the first iPods cost $399 and $499 respectively in 2001, which for a device you use every day was expensive, but not insurmountable. Compare that to the whispered introductory price for the Apple Headset of $3,000 for a device you just don't use every day. A market for that kind of device in that price range does not exist and is not going to exist.
Apple hopes that like the iPod and iPhone previously, people will use the Headset for hours, if not continuously. But the applications to do so are lacking. That leaves you with nerds and gamers, and that seems like a great market. But not at these prices and especially not when there is little spectacular new content available. Before MP3 players, there was music. In fact, more music than ever before. The device feasted on the huge supply of pirated music that flooded the Internet via Napster, Limewire and Kazaa.
Because what is often forgotten is that MP3 players, including the iPod, benefited from the ability to listen to music by artists whose entire CDs you would never otherwise have bought. I am man enough to confess that How Do I Live by LeAnn Rimes was high on my iPod playlist for years, but I had never bought an entire CD of hers.
Content development for the entire VR/AR/XR market, on the other hand, is complex, expensive and time-consuming. The Apple Headset will have to run on legal content (read: no porn) and that costs money. So where the iPod was relatively cheap and played free content, the Apple Headset will be an expensive device with expensive content. And that suddenly reminded me of my graduate thesis and the huge failure of the Apple Newton.
In Search of the Holy Grail
The Apple Newton did just fine. As a bookend and dumbbell. Source: Ars Technica
Exactly thirty years ago, in 1993, Apple launched with much fanfare the Apple Newton, a handheld computer whose main asset was the possibility of handwriting recognition, which would make a keyboard unnecessary. That same year, Frans Straver and I graduated together on a study of success and failure factors of interactive media in the consumer market, entitled In Search of the Holy Grail. The conclusion after a year-long analysis of more than 600 scientific articles and pieces from the international trade press, was not shocking: interactive media that want to be successful in the consumer market must be cheap, easy to operate and preferably provide a hefty amount of erotic coziness.
In 1993, university professors asked how we had gotten this photo on the cover of our thesis. Video camera set on an old, painted flower pot. Idea of the brilliant photographer Morad Bouchakour.
Philips misread our conclusion and was kind enough to let us present the results at a conference in Ahoy. After I showed the picture showing that Philips' interactive compact disc player, the cd-i, would be mercilessly crushed between the PC at the top of the market and the game console at the bottom, Frans and I were thrown out before lunch.
What we ourselves and the brains at Philips overlooked at the time was that the Apple Newton suffered from exactly the same shortcoming as the CD-i: high price, no necessary new applications and no supply of pink content, as it was so nicely called at the time.
Wired published a great article 10 years ago about the failure of the Newton. CNET seems inspired by this and recently came out with this video in which it takes 8 minutes and 30 seconds to draw the parallel between the Apple Newton and the Apple Headset.
The best analysis I've come across so far on the chances for the Apple Headset is this video from the Wall Street Journal. Within Apple, there also seems to be division over the potential of the glasses, and executives are now keeping their appropriate distance from the project. Hopefully the tech gods will be merciful to whoever presents the Apple Headset on June 5. My guess is that it won't be CEO Tim Cook.
Apple Headset will be a flop - by Apple standards
Reports are that Apple hopes to sell half a million to a million Headsets in the first year. Even if that market doubles every year for the next few years, which it won't, that's still change for Apple. Because that's the downside of a company heading for $500 billion in annual sales: it's unimaginably difficult at that scale to bring something new to market that has more impact on sales than, say, a marketing campaign for a new type of Airpods.
But I'd rather see a relative failure of a fundamentally new product, than more of those hopeless share repurchases that Apple has been peddling in recent years. Is there nothing better to invent, build or buy than spending $90 billion on share repurchases? That Headset isn't going to be it, but surely there will be products in development within Apple that can match the success of the... Apple Watch!
Apple had over 30% global market share in smart watches by 2022
that 30% market share led to as much as 60% revenue for Apple from every penny that went into the smart watch market (30% market share vs. 60% dollar share)
Apple will sell over 50 million Apple Watches this year
the annual growth rate of the smart watches market is nearly 20%
Apple has higher watch sales than the entire Swiss watch industry
Swatch Group, LVMH and Richemont collectively achieve lower watch sales than Apple
Rolex's annual sales are around $10 billion
Apple does not publish specific figures on Apple Watches sales, but with 50 million Apple Watches sold, Apple is at least twice the size of Rolex
Combined with the success of the Airpods, which, with the Apple Watches, fall under the Wearables division, with the emergence of the Services division, this leads to a dramatic change in the structure of revenue for Apple. See these excellent charts on Apple's revenue by product line and region over the years. Sales from Wearables now exceed those of the iPad and Mac. Apple is a computer company where revenue from traditional desktop and laptop computers is still only 8%.
It would make sense for Apple to organize revenue distribution differently, for example:
computer hardware (Mac, iPad, Apple TV)
services (Music, Movies, TV Shows, Apps, Books)
wearables (Airpods, Watches, Headset)
iPhone (you know)
I conclude about the Apple headset with the same consideration as when the Google Glass came out 10 years ago: over half the people on earth are women. Do you know one who will walk around with a device on her head all day, messing up her hairstyle?
Holland's favorite tech company apparently has servers made of papier mâché.
Spotlight 9: Google again winner of the week
Alphabet stock was again the riser of the week and I have no idea why. Google is throwing generative AI at every product right now and experts say this is a very risky strategy for the company.
I can't possibly characterize the bizarre market right now any better than this article on Crunchbase:
"Suppose you would have invested $100 in a Nasdaq Composite Index fund at the height of the boom. That would have been on Nov. 19, 2021, when interest rates were low and technology stocks were very popular.
Today, that investment would be worth about $76. It's a disappointing return that reflects how technology ratings have steadily declined in recent quarters. But it could have been even worse.
Now imagine if, instead of the index fund, you had chosen a basket of promising, venture capital-backed startups founded in the past 15 years. You know, companies like Airbnb, Coinbase, Rivian and Uber.
Let's say you had invested at the market high point in November 2021. And let's say you bought a share of the startups that launched the 19 biggest IPOs of the past 10 years. If you had held on until now, every $100 you had invested would be worth about $32. That's a much steeper drop than the drop in the technology-focused Nasdaq Composite Index and points to the public's deeper disappointment in mostly unprofitable new market entrants."
Two conclusions: profits are currently considered more important than growth. And investing in tech stocks is and remains highly risky. Because even Apple does not score on every try.
Opinions like this in your mailbox weekly? Sign up for the free newsletter and receive my views on the latest developments in technology, entrepreneurship and innovation every week.
