TheChek News Flow

Quants have tried for decades with limited success at their biggest challenge

Where is Wall Street’s AI revolution?

Almost every industry, from architecture to entertainment, is testing generative AI, hoping to profit from a technology that can produce writing, images and art much like humans.

Wall Street has long used automated algorithms for tasks such as placing trades and managing risk. But investors haven’t made much progress relying on AI to tackle their biggest challenge: beating the market. While some see ChatGPT as a way to boost sales and research efforts, the investing results using AI haven’t been especially impressive.

“Progress in applying AI to investing has been limited, though innovations in language modeling could change that in the years ahead,” says Jonathan Larkin, a managing director with Columbia Investment Management Co., which manages the $13 billion endowment for Columbia University and invests in various funds.

Wall Street had a head start in AI. Four decades ago, mathematicians-turned-quants including Jim Simons, founder of Renaissance Technologies, developed algorithms to turn investing decisions over to their computers.

He and other quants have spent years using machine learning, a type of AI. They have built trading models that can extrapolate from past data to identify patterns and develop profitable trades, with limited human intervention.

But few firms have found success turning all of their operations over to machines, quants say. And they haven’t enjoyed dramatic advances with self-learning or reinforcement learning, which entail training computers to learn and develop strategies on their own. Indeed, Renaissance and others rely on advanced statistics rather than cutting-edge AI methods, say people at the firms.

“Most quants still take a “theory-first” approach where they first establish a hypothesis of why a certain anomaly might exist, and they form a model around that,” says Mr. Larkin.

One big problem: Investors rely on more limited data sets than those used to develop the ChatGPT chatbot and similar language-based AI efforts. ChatGPT, for example, is a model with 175 billion parameters that uses decades—and sometimes centuries—of text and other data from books, journals, the internet and more. By contrast, hedge funds and other investors generally train their own trading systems using pricing and other market data, which is limited by nature.

“The circumstances are different” when it comes to investing, says Jon McAuliffe, a former D.E. Shaw executive who now is the co-founder of Voleon Capital Management LP, a hedge fund that relies on machine learning. “We don’t have unlimited amounts of data to help us run models of unlimited size.”

Just as important, market data is “noisier” than language and other data, making it harder to use it to explain or predict market moves. In other words, earnings, share momentum, investor sentiment and other financial data only partly explain stock moves, and the rest is unaccountable “noise.” As a result, machine-learning models can identify correlations in various market data but prove incapable of predicting future stock moves.

Unlike languages, markets can change quickly—companies alter strategies, new leaders make radical decisions and economic and political environments shift abruptly—making it harder to make trades using models reliant on historic, long-term data trends.

And while ChatGPT has proven impressive, it regularly makes the kinds of obvious errors that would cost investors’ money and jeopardize their reputations.

Richard Dewey, chief executive of financial-technology company Proven, also notes that investing is “adversarial,” or it entails competing with rivals eager to take advantage of any mistake. That makes it harder to invest using AI than it is to use these approaches for natural language, image classification or self-driving cars.

“There’s a reason firms like Renaissance and D.E. Shaw still employ so many Ph.D.s,” says Mr. Dewey, who says humans are still essential in markets that are noisy and subject to human-behavior feedback loops. “When it comes to investing, it’s still hard to turn everything over to the machines.”

Still, there are signs that investors are becoming more comfortable relying on AI. Voleon is among a group of hedge funds launched in the past few years built around machine learning and other AI approaches.

San Francisco-based quant hedge fund Numerai used machine-learning techniques to score gains of 20% last year, the firm says. Also last year, three senior staffers at DeepMind Technologies, the artificial-intelligence subsidiary of Google parent Alphabet Inc., caused a buzz by leaving to start a machine-learning fund called EquiLibre Technologies, based in Prague.

AI may someday help democratize trading, giving individuals and others programs as powerful as those used by big hedge funds, some AI specialists say. For now, though, there are too few firms focusing on machine learning and other AI methods to determine whether big returns are possible, says Jens Foehrenbach, chief investment officer of Man FRM, which invests more than $20 billion in hedge funds. And the early returns are inconsistent.

“Their results vary considerably,” Mr. Foehrenbach says. “Such strategies can do highly unexpected things, and that makes it hard for an investor to decide” whether to cut or increase their investments.

AI-adherents believe their approach will eventually perform well. Machine-learning models could eventually sort the meaningful from the meaningless.

“Building machine-learning strategies is more difficult, and there are more false starts.” Mr. McAuliffe of Voleon says. “But once you get them working, these strategies will make more accurate predictions.”

Martin Schmid, EquiLibre’s co-founder and CEO, says “reinforcement learning,” a form of machine learning in which computers are “punished and rewarded” for various trading investment decisions, will work for stocks and bonds, much like it has chess, poker and other games.

EquiLibre is still working on its trading models and hasn’t begun investing, Mr. Schmid says.

In the near term, the recent advances could shake up areas such as research and sales, some say.

“Now, you can create automated, bespoke information for clients, and that’s a lot of what salespeople [at investment banks] do,” says Jens Nordvig, a former Goldman Sachs and Bridgewater Associates staffer who runs MarketReader, which uses AI to distill financial news.

American partners told retrenchment is part of strategy to simplify business

EY’s US business will embark on a $500mn cost-saving programme after its opposition torpedoed plans for a historic split of the Big Four firm.

US leaders outlined a new strategy in a memo to partners sent shortly after EY’s global executive committee said on Tuesday it was abandoning ambitions to spin off its consulting and tax advisory businesses into a new company.

The collapse of the plan, which would have marked the biggest shake-up to the accounting industry in more than two decades, has pitched the global firm into a new period of recrimination and uncertainty.

The memo to US partners — signed by Julie Boland, US managing partner, and the rest of the US executive committee and seen by the Financial Times — said there was a strategic rationale for splitting the business in some way in the future.

However, the decision to ditch the planned split, which was codenamed Project Everest and had been worked on for more than a year, would allow EY to focus on freeing up capital for investment and to pursue governance reforms that had been put on hold, it said.

In particular, the US firm would act to “accelerate decision-making, streamline accountability and reduce complexity”. Without giving details, the memo said the “US simplification agenda will start immediately . . . and we expect to drive savings in the US of $500mn over the next 12 months”.

EY’s US firm accounts for 40 per cent of its global revenues, which were $45bn in the fiscal year ended June 2022. EY operates as a global network of member firms, and any split would have needed approval on a country-by-country basis.

In the memo, Boland and the executives said they would seek to beat the $500mn target “substantially” by “also streamlining global infrastructure and eliminating duplication in our global operating model”, though that would require co-operation with the global leadership.

Other Big Four firms have been trying to cut costs in the US after growth in their consulting businesses slowed sharply over the past year.

KPMG has announced it is laying off close to 2 per cent of its US workforce, and Deloitte’s consulting staff have been told to expect tougher performance reviews that will lead to more people leaving than in recent years, according to people familiar with internal communications.

The EY US leadership memo promised new investments in the audit business and tax practice that was at the heart of the disputes over Project Everest. US audit leaders opposed letting a majority of tax partners go to the new consulting business, saying it would damage the quality of EY’s audit work and threatened the financial strength of the audit-focused business.

EY’s global chief executive Carmine Di Sibio had championed the plan to split as a “road map” for the rest of the profession, saying it would free both sides of the business from conflict-of-interest rules that prevent consultants from selling many services to audit clients.

On Tuesday, after weeks of last-ditch negotiations, he and other global leaders said they would abandon the plan on hearing that the US would not take part.

In a separate email to retired partners later on Tuesday, the US executive committee gave further details of why it had vetoed Project Everest.

“The analysis on the proposed strategies and perimeter for both organisations identified gaps in our ability to deliver exceptional client service, particularly to our largest global clients,” Boland and the 15 other executives wrote.

“The amount of time it would take to improve business performance and achieve a viable transaction has become much longer than anticipated,” they added, and “the transaction economics have become challenged given the current economic conditions and capital markets environment”.

EPA announces its ‘most ambitious pollution standard ever’ for cars and trucks

US environmental regulators on Wednesday proposed tough new emissions limits that would force carmakers to make 67 per cent of their American models electric by 2032.

EPA administrator Michael Regan called it “the most ambitious pollution standards ever for cars and trucks”. It would significantly increase EVs share of the new vehicle market, which stood at about 7 per cent in 2022.

The proposed rule would limit tailpipe emissions across all the vehicles in a carmaker’s fleet, forcing companies to make more battery-powered vehicles to meet the new standard. It would affect models built starting in 2027 through to 2032.

The proposal would also curb air pollution, boosting the US’s chance of achieving its Paris Agreement pledge to lower emissions by 50-52 per cent below 2005 levels by 2030. The rules would also force carmakers to speed up their timetables for electrification, and in doing so tilt the US car market towards EVs even as they remain more expensive than vehicles with traditional engines.

General Motors, Ford and Stellantis have pledged to make between 40-50 per cent of their US sales electric by 2030, and they have invested billions of dollars in developing new models and building factories. But there are questions about whether the carmakers can build cheaper electric vehicles, which continue to cost more than gas-powered equivalents.

The regulation was an “ambitious goal in a pretty short time period”, said Cox Automotive analyst Michelle Krebs.

“The big question is, ‘What is consumer acceptance?’” she said. “You can mandate them, but people have to buy them. A lot of that will depend on can automakers produce affordable ones, and ones they can be profitable with?”

Transportation generates more greenhouses gases than any other sector in the US. “If you’re serious about dealing with climate change problems, then you need to address the transportation sector,” said Howard Learner, executive director of the Environmental Law & Policy Center.

President Joe Biden issued an executive order in August 2021 calling for half of all new US car and truck sales to be electric by the end of the decade. While that order was not legally enforceable, the EPA regulation setting a higher emissions threshold would be.

The federal government’s target follows regulation by California eight months earlier to ban sales of petrol-powered vehicles in the state by 2035. The state is a heavyweight in the US vehicle market, with 13 per cent of all car and truck sales last year, which has allowed it to influence environmental policy for decades.

John Bozzella, chief executive of the Alliance for Automotive Innovation, a trade group representing car and battery makers, called the new standard “aggressive by any measure”, pointing out it exceeded the goal the Biden administration set just 20 months ago. 

“The question isn’t can this be done, it’s how fast can it be done,” he said. 

In the short term, the rule benefits manufacturers such as Tesla that make only electric cars. But moving to green energy required a number of successful carmakers, Krebs said. “The pie is going to get bigger, so you need more pieces of pie.”

Liquidity crunch leads investors such as Andreessen Horowitz to sovereign wealth funds in Saudi Arabia, Qatar and UAE

Silicon Valley investors are touring the Middle East, seeking to build long-term ties with sovereign wealth funds during the worst funding crunch for venture capital firms in almost a decade.

Top technology VCs such as Andreessen Horowitz, Tiger Global and IVP have jetted teams of executives to Saudi Arabia, the United Arab Emirates and Qatar in recent weeks, according to people with knowledge of the trips.

These visits come after their traditional North American and European backers contend with an economic downturn that has forced them to rein in private investments.

VCs are, in turn, being encouraged to come to the region, as Gulf officials and young royals seek to diversify their economy away from oil with investments into hot tech sectors such as artificial intelligence.

That has also meant that some VCs have quietly reversed earlier decisions to refuse meetings with, or cash from, Saudi Arabia over concerns about its human rights record following the 2018 murder of journalist Jamal Khashoggi.

“We came to San Francisco looking for them in 2017. Now . . . everyone is coming to [us],” said Ibrahim Ajami, head of ventures at Mubadala Capital, a $6bn arm of Abu Dhabi’s $284bn sovereign wealth fund. “The tech correction has humbled the industry.”

The Financial Times interviewed more than a dozen Silicon Valley VCs who control tens of billions of dollars between them, as well as a string of advisers and bankers. They describe a new love affair between US venture funds and Middle Eastern cash.

A group of Silicon Valley executives received a personal invite from the office of Yasir al-Rumayyan, the governor of PIF, the $620bn Saudi sovereign wealth fund, to be his guest at last month’s Formula One Saudi Arabian Grand Prix in Jeddah, according to a person with knowledge of the calls.

Among the attendees, according to the person, was Andreessen Horowitz co-founder Ben Horowitz — the veteran financier’s second trip to Saudi Arabia in less than six months. Andreessen Horowitz declined to comment.

PIF’s venture arm Sanabil recently disclosed its partnerships with nearly 40 US venture firms, including Andreessen Horowitz, Coatue Management, David Sacks’ Craft Ventures, Insight Partners and 9Yards Capital, where former UK chancellor George Osborne is a managing partner. The sums invested in the firms were not revealed.

Horowitz, whose San Francisco-based firm raised just over $14bn last year, in particular has become a vocal supporter of Saudi interest in tech innovation.

In October, he spoke at the “Davos in the desert” conference in Riyadh and had lunch with Princess Reema bint Bandar al-Saud, the Saudi ambassador to the US. At a conference organised by PIF last month in Miami, he praised the kingdom as a “start-up country” and likened its crown prince, Mohammed bin Salman, to a company founder.

A year ago, Horowitz’s trips to Saudi Arabia would have been an anomaly among VCs flush with cash and keen to avoid the moral predicament of dealing with states with poor human rights records. Now the oil-rich Gulf, which enjoyed a petrodollar windfall last year, is buzzing with US start-up investors, according to several people who have visited there this year.

“The Four Seasons in Riyadh is basically Palo Alto,” said a partner at one large Silicon Valley venture fund.

That willingness to do business in the region has led to some criticism. Founders Fund partner Keith Rabois, who said in 2018 that Silicon Valley had been hypocritical for accepting Saudi money, said, “I don’t change my values and principles because a funding environment is difficult.”

But Lead Edge founder Mitchell Green, who made venture investments in Alibaba and Uber, said he had spent the past few weeks “building long-term relationships” with people and companies in the Gulf. “We think it will become an increasingly important area of the world over the next decade. It reminds us of going to China in 2003.”

Tiger Global partner Scott Shleifer also spoke at the Riyadh conference in October, and the firm has been pursuing investment from the kingdom as part of its latest $6bn fundraising, according to people familiar with the matter. A team from IVP, one of Silicon Valley’s oldest venture firms, led by partner Somesh Dash, went on a tour of the region, according to one person with knowledge of the trip. IVP declined to comment.

US venture capital has exploded in size in recent years, in part fuelled by a boom in tech valuations during the coronavirus pandemic. Marquee funds that once prized exclusivity, such as Sequoia Capital and Andreessen Horowitz, have raised funds of as much as $5bn and sometimes as large as $9bn. That shift has been intensified by large entrants to the market, such as Japan’s SoftBank and Tiger Global, which have deployed tens of billions of dollars into start-ups.

“These guys have built their models on high-volume, high-velocity investors — now they are a prisoner to the capital cycle,” said a partner at a venture fund with more than $4bn under management.

Dealing with nations such as Saudi Arabia is the “Faustian bargain these firms have made by scaling up”, the partner added. “They went for ubiquity and market share and gave up on scarcity, and because of that they have to play the game of selling themselves. Venture capital went from being the Hermès Birkin bag of investing to Target.”

PIF, in particular, has gradually permeated US tech through its investments over the past decade. It contributed $45bn to the $100bn SoftBank Vision Fund in 2016. It has made large direct investments in US tech companies, including a $3.5bn investment in Uber in 2016 and more than $1bn in electric-car maker Lucid Motors in 2018. That same year, Elon Musk said he was in talks with PIF to help him fund a $72bn deal to take Tesla private, although a deal did not materialise.

In the wake of the murder of journalist Jamal Khashoggi by Saudi operatives in late 2018, a number of high-profile western businesses, including many tech investors, stopped publicly working with the country.

That continued until the recent economic downturn, which has meant pools of capital available for venture funds at large western institutions have dried up. Fundraising by venture capital firms hit a nine-year low at the end of 2022, according to research firm Preqin.

VCs are sitting on a record $300bn of “dry powder” — money raised that has not yet been deployed. But many are struggling to find lucrative investments in start-ups and will be unable to raise a new venture fund.

Cash that VCs put into start-ups has plunged more than 50 per cent over the past 12 months, according to data provider Crunchbase.

As a result, many have been lured back to the Gulf, which “is the most liquid place on the planet right now”, according to the head of a $1bn venture fund.

“This is a unique opportunity for funds like Mubadala to really take a leading role in the development of technology over the next 20 years,” Ajami said. Mubadala has invested in or alongside a number of big Silicon Valley firms, including private equity group Silver Lake and Sequoia Capital. It recently invested in fintech group Klarna alongside Sequoia.

Meanwhile, the Qatar Investment Authority said in 2019 it would raise investments in the US from $30bn to $45bn, including in technology.

For some investors, there remains a tricky moral debate. “The US is buying oil from Saudi, we’re selling them drones, where do you draw the line?” said one venture capitalist, who admitted they had shifted from a stance of never accepting Saudi money to being more open to it as fundraising had dried up.

For others, particularly those who control smaller funds and therefore have still been able to tap up western pension and endowment funds, the issue is more black and white.

“I’ve been in the Valley for 20 years and I’m increasingly disappointed with the way we behave,” said one senior banker who handles deals for venture firms. “If you’re really good at what you do, go to Norway [to raise money].”

L'ancien président du FMI est sorti de son silence sur les réseaux sociaux et interpelle directement le chef de l'État sur le contexte social de ces dernières semaines. Il accuse l'exécutif de multiplier les «erreurs».

Sa parole est extrêmement rare depuis qu'il n'est plus sous le feu des projecteurs. Dans une longue publication postée sur Twitter ce lundi et intitulée «La cinquième erreur», l'ancien président du FMI Dominique Strauss-Kahn revient sur le contexte social tendu de ces dernières semaines en France, lié à la réforme des retraites. «Aujourd'hui, ce n'est pas l'âge de la retraite qu'il faut changer, c'est la conception même du système», indique l'ancien ministre de l'Économie. Qui interpelle directement le président de la République, à l'origine, selon lui, de quatre erreurs. L'occasion de louer, par contraste, la «réforme systémique» qu'Emmanuel Macron avait voulu entreprendre, lors de son premier quinquennat, en 2019. Alors que la nouvelle architecture des retraites avait été mise en pause face à la première vague épidémique, avant d'être abandonnée sous cette forme, «DSK» voit dans la réforme paramétrique de 2023 - le relèvement de l'âge légal de départ de 62 à 64 ans - une «erreur de méthode».

Pour Dominique Strauss-Kahn, la France, qui «est dans un triste état», doit «sortir d'un système qui repose sur l'âge de départ à la retraite pour construire un système fondé sur la durée de cotisation.» Cette dernière «s'ajusterait lentement et régulièrement avec l'évolution des besoins de financement.» Et l'ancien prétendant putatif à la présidentielle de 2012 d'apostropher Emmanuel Macron sur la temporalité de la réforme dans le contexte inflationniste et social ambiant. Une autre «erreur». «Le moment n'était-il pas particulièrement mal choisi qui crée une crainte nouvelle sur la retraite quand d'autres détresses assaillent les Français ?», se demande alors DSK à quelques jours de la décision du Conseil constitutionnel sur la réforme des retraites.

«Relancer le cycle de violences»

«On pense d'abord à la baisse du pouvoir d'achat découlant de l'inflation qui frappe durement la partie la moins aisée de la population, s'inquiète l'ancien cacique du PS. Si comme nous l'apprend un sondage récent, près de la moitié des Français les plus modestes ont été amenés à supprimer un repas par jour, alors c'est là qu'est l'urgence.» Car «rien ne nous obligeait à agir tout de suite. Rien, sauf peut-être le sentiment qu'il fallait au plus vite tenir une promesse.»

L'erreur suivante est «stratégique». «La France n'est pas un pays dans lequel on peut mener une réforme sociale d'envergure en se référant uniquement à un rapport de force politique et en négligeant le rapport de force social et, au-delà, le consensus minimum au sein de la société. C'est pourtant ce qui a été fait», cingle l'ancien patron de Bercy. Le déclenchement du 49.3 ayant entraîné une série de violences dans le pays, «la crise traduit (...) le blocage profond de la société française». Brocardant cette nouvelle «erreur», l'ancien ministre de l'Industrie considère que «le pouvoir politique», «mesurant son absence de majorité politique», «s'est retrouvé dans une impasse.»

Si la réforme va au bout de son cheminement et a l'aval du Conseil constitutionnel vendredi prochain, «la loi risque, à sa publication, de relancer le cycle de violences», prévient-il. Raison pour laquelle DSK, qui souhaite qu'Emmanuel Macron «renoue les fils d'un dialogue social trop brutalement interrompu», redoute une cinquième erreur et soumet une proposition : «On peut l'éviter (la violence, NDLR) en repoussant cette publication jusqu'à ce qu'un point d'équilibre ait été trouvé avec les forces sociales. Faute de quoi le péril est grand de voir le pays s'embraser à nouveau.»

Startup AlphaSense Snags $100M As AI Stays Hot; Alphabet’s CapitalG Invests

At this point one needs an AI platform to keep track of all the big money AI startups have raised.

AlphaSense became the latest AI-enhanced startup to lock up big money, as it raised another $100 million from investors that included CapitalG — Alphabet’s independent growth fund. Other investors include the growth equity business within Goldman Sachs Asset Management and Viking Global Investors.

The new money values the company at $1.8 billion.

The cash is in addition to the $225 million Series D that valued the company at $1.7 billion last June.

The New York-based startup has created a market intelligence and search platform — powered by AI and natural language processing — to help clients form corporate and investment strategies.

Google and AI

Although CapitalG is an independent growth fund under Alphabet’s umbrella, Google and its parent have certainly shown significant interest in the AI space. In February, Google introduced its conversational AI tool Bard, which it plans to integrate into many of its products including its dominant search engine.

Earlier this year, Google also invested between $300 million and $400 million into San Francisco-based Anthropic, a ChatGPT rival.

AI frenzy

Of course, that is far from the only AI-powered company to raise big money, as VCs and large strategics including Microsoft, Google and Salesforce all continue to show unabated interest in generative AI startups.

Just last month, Character.ai closed a $150 million Series A at a $1 billion valuation led by Andreessen Horowitz. The Palo Alto, California-based AI startup allows people to create their own personalized AI chatbot using language models and deep-learning algorithms.

Also in March, San Francisco-based Adept AI raised $350 million in a Series B — at a reported post-money valuation of at least $1 billion.

Anthropic was reported to be raising another $300 million round at a pre-investment valuation of $4.1 billion. Spark Capital is reportedly leading the round.

Of course the craze started in January with news of Microsoft’s massive $10 billion investment into OpenAI — creator of ChatGPT.

DAX:

MDAX:

SDAX: