How will declining immigration impact the US economy?

    Immigration to the US is expected to fall from the elevated levels of the past three years, declining to a pace slightly below the pre-pandemic average, according to Goldman Sachs Research. If that occurs, the impact on the economy is likely to be limited, though more significant restrictions on immigration by the Trump administration could have larger repercussions.  Net immigration is expected to slow to 750,000 per year, well below the pace of the last three years but only moderately below the normal pre-pandemic pace, Goldman Sachs Research economists Elsie Peng, David Mericle, and Alec Phillips write in the team’s report. In their baseline estimate, the GDP impact from changes in immigration is likely to be limited: The slower pace of immigration would contribute 30-40 basis points less to potential US GDP growth than the 2023-2024 pace, but it would be just 5 basis points less than the pre-pandemic pace. The team’s baseline outlook for reduced immigration is based on an expected increase in border security and other immigration control measures as well as a moderate increase in deportations. More significant measures by the Trump administration could reduce net immigration further and increase the impact on labor and the economy.  What does reduced US immigration mean for the job market? The impact from reduced immigration on wage growth and inflation should be modest now that the US labor market is back in balance, according to Goldman Sachs Research. At its peak, the boost to labor force growth from immigration was 100,000 per month above the normal pre-pandemic pace. It has since fallen to 40,000 above the typical level and is predicted to return to normal by early 2026. Goldman Sachs Research has argued that the US unemployment rate would stop rising and start falling because labor demand has been healthy all along: The layoff rate remains historically low and job openings are high, and the pace of labor force growth will be more manageable now that immigration is slowing. While our economists think the natural path for the unemployment rate is a little lower — the unemployment rate has fallen slightly over the last two months to 4% — they note that the crackdown on unauthorized immigrant workers could cause more of them to end up unemployed. These dynamics might not show up in official statistics, as immigrants who are concerned about going to work might also be unwilling to respond to employment surveys.  Reduced immigration will have the largest impact on agriculture and construction The US government’s changes in immigration enforcement target asylum seekers, parolees, people receiving Temporary Protected Status, and those crossing the border illegally. Reductions in numbers for this group, rather than visa recipients or green card holders, made up the sharp decline in net immigration that was evident by the end of last year. Immigrants other than visa and green card holders account for 4%-5% of the total US workforce, and they make up 15%-20% in some industries, such as crop production, food processing, and construction. “Abruptly losing a significant share of these workers could be very disruptive for many of these industries,” the team writes. There could be temporary production bottlenecks, shortages, and price increases. In the team’s baseline forecast, the 750,000 of net immigration per year represents mostly visa and green card holders. Some 500,000 deportations are expected to largely offset roughly 500,000 people entering the country as asylum seekers and people entering illegally, which is the low end of the pre-pandemic range. The administration’s immigration policies may run up against constraints, including the number of enforcement agents available and space in detention facilities. Congress is expected to allocate at least $100 billion in additional resources for law enforcement later this year, much of which will likely be used to hire more agents. 

How will declining immigration impact the US economy? Read More »

How global stock market rankings are forecast to change

    The US has had the world’s largest economy for more than a century — a title it’s expected to relinquish in the coming decades. But even as US GDP is forecast to be surpassed in the years ahead, the country is projected to remain world leading when it comes to wealth and the size of its stock market, according to Goldman Sachs Research’s report “The Path to 2075.” Demographic projections and long-term drivers of productivity can help us glimpse how the global economy may look 50 years in the future. In fact, these longer-term forecasts are, in some ways, easier than shorter estimates over a year or two, which can be derailed by booms, recessions, and other surprises. That’s because some of the key variables underpinning long-term GDP growth are slower to change, while the shorter-term volatility of the business cycle tends to average out over time,  say Kevin Daly, co-head of Central & Eastern Europe, Middle East, and Africa Economics in Global Macro Research, and economist Tadas Gedminas. “Over the very long term, the things that tend to drive the size of economies are things like population growth and long-term productivity growth, which tend to be slower-moving and less variable,” Daly says.   The relative importance of capital markets in emerging economies is nevertheless projected to grow, from around one-quarter of total global market cap today to more than half by 2075. The research demonstrates that while rich, developed countries will remain critical in the decades that come, it won’t be possible to capture long-term, worldwide growth trends without exposure to emerging economies and their financial markets.

How global stock market rankings are forecast to change Read More »

How much will rising defense spending boost Europe’s economy?

March 6, 2025Shareshare   Defense spending by European Union member states is set to increase significantly in the next two years. The shift will have a positive — but limited — impact on GDP growth, Goldman Sachs Research economists Niklas Garnadt and Filippo Taddei write in a report. The team’s baseline assumption is that the EU will gradually increase its annual defense spending by around €80 billion ($84 billion) by 2027 — equivalent to roughly 0.5% of GDP, according to the report dated February 27. Defense expenditures in the euro area accounted for 1.8% of GDP in 2024 and Goldman Sachs Research expects them to rise to 2.4% by 2027.  The incoming German government recently said it intends to exempt defense spending from budget control measures and to allot €500 billion to an infrastructure fund. If implemented, the policies could result in faster-than-expected GDP growth from Europe’s largest economy. The economic impact of defense spending depends on the type of expenditure and whether it is imported or produced locally. Goldman Sachs Research estimates that additional spending on defense will have a fiscal multiplier of 0.5 over two years. That means every €100 spent on defense would boost GDP by around €50. The forecast is based on the assumption that imports of military supplies gradually decrease (and are substituted with domestic products) and that the higher spending initially focuses on equipment and infrastructure. What is the outlook for European defense spending? Spending on equipment has recently increased more than other areas of defense, reaching 33% of spending by European members of NATO last year, up from 15% in 2014. Europe bought a substantial amount of military equipment from non-EU suppliers immediately after Ukraine was invaded by Russia. However, a large portion of European defense supplies has historically been purchased from domestic companies, particularly in larger EU member states. The average domestic share of sourcing was around 90% in France, 80% in Germany, and 70% in Italy between 2005 and 2022. Europe’s share of global arms production declined between 2008-2016, although it has since started to pick up again. EU manufacturers have joined the global surge in arms production and are now poised to expand at a faster rate than their US counterparts, according to market pricing. As defense spending increases, there will be growing opportunity for equipment to be harmonized (made interoperable across the continent), for research and development to scale up, and for efficiency to improve. Such changes would increase the economic impact of military spending, and it would probably result in a higher fiscal multiplier after three years. How Europe could fund higher defense spending To meet a defense-spending target of 2.5% of GDP, the euro area needs to increase expenditures by an additional 0.6% of GDP annually, Taddei writes in a separate research report dated March 2. European leaders are discussing a common strategy for increasing defense spending, which could involve issuing more debt at the national or EU level, or setting up new lending facilities from European institutions. Issuing more national debt could be challenging given the new European fiscal framework, which requires countries to contain their ratio of debt to GDP. European rules allow a temporary exception in the case of “major shocks to the EU,” Taddei writes, known as the “escape clause.” EU President Ursula Von der Leyen proposed this option at the Munich Security Conference in February. Making this exception permanent for future defense spending needs (known as a “golden rule”) would require the approval of the EU Council and the EU Parliament. Taddei writes that the EU president’s proposal has the advantage of being relatively quick. But he adds that “introducing a ‘golden rule’ would leave national defense spending exposed to sovereign market stress and reduce the likelihood of coordinated and harmonized military spending within the EU.” How Europe could leverage supranational debt Alternatively, the EU could turn to existing lending programs that are available for European governments — either the European Stability Mechanism (ESM) or the European Investment Bank (EIB). “The EIB has struggled to identify projects worth funding in line with the European priorities, and the industrial reconversion needed to scale up defence spending in Europe would likely provide an ideal target,” Taddei writes. These options have limitations however. Only euro area members would be eligible for ESM lending, for example, and the ESM would only temporarily shift issuance from domestic to supranational debt. EU debt, meanwhile, would provide stable funding. This could come in the form of repurposing an existing Covid pandemic borrowing program (called NGEU), or as a separate program that is dedicated to defense borrowing. The latter is the only option to secure low rates for long-run funding. “However, it is also the option with the most cumbersome approval process,” Taddei writes. The team expects that setting up a new funding facility would take about a year from design to implementation. “We continue to expect the EU to use national debt, NGEU, and a new funding facility, but in that sequence,” Taddei writes. He adds that national debt, combined with the repurposing of spare NGEU financial capacity, could fund military spending until 2026.

How much will rising defense spending boost Europe’s economy? Read More »

Why the US economy may grow more slowly than expected

   The US economy may expand more slowly than previously forecast as tariffs on imports rise and the Trump administration signals that it may tolerate slower growth in order to implement its trade policies, according to Goldman Sachs Research. Our economists reduced their prediction for US GDP expansion to 1.7% in the fourth quarter of 2025 (year over year) from their earlier estimate of 2.2%. Goldman Sachs Research’s forecast for the world’s largest economy is, for the first time in more than two years, lower than the consensus estimate of economists surveyed by Bloomberg.  “Our trade policy assumptions have become considerably more adverse and the administration is managing expectations towards tariff-induced near-term economic weakness,” Goldman Sachs Research Chief Economist Jan Hatzius writes in the team’s report. The average US tariff rate is expected to rise by 10 percentage points this year. That’s twice Goldman Sachs Research’s previous forecast and about five times the increase seen in the first Trump administration. While some import taxes have been softened, our economists expect levies in the coming months on critical goods, global autos, and a “reciprocal” tariff. Reciprocal tariffs and the administration’s view of Europe’s 20% value added tax (VAT) are particularly important because the US considers the tax a tariff (even though Europe imposes it equally on imported and domestically produced goods). If applied mechanically, a reciprocal tariff that includes the effect of VAT could raise the average US tariff rate by 10 percentage points or more. Tariff carveouts will probably lower this number, but if the exemptions are less widespread than Goldman Sachs Research expects, the average tariff rate could rise as much as 15 percentage points. What are the economic effects of tariffs? Tariffs are likely to weigh on US economic growth via three main channels, according to Goldman Sachs Research. They raise consumer prices — and thereby cut real income — by an estimated 0.1% per 1 percentage point increase in the average US tariff rate. (In theory, the drag could diminish if the tariff revenue is recycled into additional tax cuts, but this revenue will not be scored in the ongoing budget negotiations if it results from executive as opposed to congressional action.) Tariffs tend to tighten financial conditions, although the impact in this cycle looks smaller than in the 2018-2019 trade war when scaled by the size of the tariff hikes. Trade policy uncertainty leads businesses to delay investment. All told, the team’s new baseline implies that tariffs will subtract an estimated 0.8 percentage point from GDP growth over the next year, with only 0.1-0.2 percentage point of this drag offset by the (relatively slow-moving) boost from tax cuts and regulatory easing. Will tariffs lead to higher inflation? Goldman Sachs Research now expects core PCE inflation to reaccelerate to 3% later this year, up nearly half a percentage point from their prior forecast. In theory, a tariff hike raises the price level permanently but only raises the inflation rate temporarily. In practice, this hinges on the assumption that inflation expectations remain well-anchored, which looks a bit more tenuous following the pickup in inflation-expectations measures from the University of Michigan and the Conference Board. Given their downgrade to the forecast for US GDP growth, our economists still expect the Federal Reserve to make two 25-basis-point cuts to the fed funds rate this year (June and December). Goldman Sachs Research’s near-term view is that the Federal Open Market Committee will want to stay on the sidelines and make as little news as possible until the policy outlook has become clearer.

Why the US economy may grow more slowly than expected Read More »

Carbonomics: Tariffs, deglobalization and the cost of decarbonization

Goldman Sachs Research has updated its Carbonomics cost curve which considers over 100 different applications for decarbonization tech across key emitting sectors, reflecting technological innovation and a growing push for local supply chains and tariffs.   CarbonomicsTariffs, deglobalization and the cost of decarbonization  

Carbonomics: Tariffs, deglobalization and the cost of decarbonization Read More »

Defense spending to boost German and European GDP growth

   The economic growth outlook is improving in Germany — and in Europe as a whole — amid a fiscal plan that emerged after Germany’s federal election and the prospect of higher military spending across the region, according to Goldman Sachs Research. German voters in late February put Friedrich Merz in line to become Chancellor and gave his Christian Democratic Union (CDU) and the Social Democratic Party (SPD) a slim legislative majority that should allow for a two-party coalition. This outcome makes higher government spending more likely. The coalition partners have announced a fiscal plan to exempt substantial defense outlays from Germany’s so-called debt brake and to create a €500 billion ($546 billion) off-budget infrastructure and climate protection fund, among other steps.  In light of these developments, Goldman Sachs Research Chief European Economist Sven Jari Stehn and his team increased their forecast for real GDP growth in Germany this year to 0.2% from flat. They also raised their 2026 forecast by 0.5 percentage point to 1.5% and increased the estimate for 2027 by 0.6 percentage point to 2%. “Growth could be higher with quicker implementation,” Stehn and his colleagues write in a report. “In practice, we think the implementation will be more gradual given capacity constraints and well-known challenges with stepping up public investment.”  Why the German economy is improving The researchers examine the potential impact of three key elements in the fiscal plan. Defense spending in excess of 1% of GDP would become exempt from the debt brake, Germany’s constitutional limit on structural deficits. The team sees military spending ramping up to 3% of GDP by 2027 and reaching 3.5% after that. The off-budget infrastructure and climate protection fund, designed to last 12 years, would boost spending gradually, raising expenditures by €40 billion above our economists’ pre-election baseline in 2027. A third feature of the fiscal plan increases the permissible structural deficit German states can run. This and the freed-up space in the federal budget may be partially used for tax cuts. The lower house of parliament (Bundestag) passed the package this week and our researchers expect the fiscal package to also pass the upper house (Bundesrat) later this week, before newly elected Bundestag members are seated in late March. Business leaders and investors have been pushing for a loosening of Germany’s debt rules and a boost in government spending, as the economy has been sluggish for several years, a growth laggard among the large European nations. The outlook for euro area GDP growth The researchers also raised their forecasts for the euro zone as a whole. They added 0.1 percentage point to this year’s growth estimate, bringing it to 0.8% for the region. They increased the 2026 forecast by 0.2 percentage point to 1.3%, and boosted the 2027 numbers by 0.3 percentage point to 1.6%. “One reason is that we expect stronger growth in Germany to spill over into neighboring countries,” Stehn writes of the forecast change. “Another reason is that we now expect the rest of the euro area to step up military spending somewhat more quickly in response to the German announcement.” The team sees France boosting defense spending to 2.9% of GDP by 2027, Italy reaching 2.8% of GDP, and Spain boosting outlays to 2.7% of GDP. This is a 0.3 percentage point increase from the researchers’ previous estimates. Some of the increases in defense outlays could be offset by spending cuts elsewhere or tax increases, the researchers note, as these countries bump up against their own fiscal limits, resulting in a smaller economic boost.      “We see risks in both directions around our new forecast” for the euro zone, Stehn writes. A steeper increase in public spending, especially in Germany, could create faster-than-forecast growth in 2026 and 2027. On the other hand, the researchers acknowledge the ongoing risk that tariffs and trade tensions with the US might have a greater-than-expected impact. The researchers have as a baseline a 0.5 percentage point drag on growth from targeted tariffs and trade policy uncertainty in 2025. “An across-the-board tariff could imply an additional hit to growth of 0.5% this year,” they write. The prospect of increased government spending across the euro zone decreases pressure on the European Central Bank to cut rates below the neutral policy rate, the researchers find. They now expect that the central bankers will be satisfied by cutting rates to a terminal rate of 2%, with 0.25% cuts expected in April and June (the policy rate is 2.5% now), rather than lowering it further in July. 

Defense spending to boost German and European GDP growth Read More »

Chinese measures to raise birth rates are boosting dairy stocks

    Recent policy announcements in China highlight new government efforts to raise birth rates. For investors, this suggests an improving outlook among dairy and infant formula companies that have sales in China, according to Goldman Sachs Research. It also creates a positive storyline for companies outside Asia that make ingredients for infant nutrition. The policy developments include a March 13 announcement by leaders in Hohhot, Inner Mongolia’s capital, of child-raising subsidies. The city will offer a one-time payment of RMB 10,000 ($1,383) to help support a family’s first child; provide RMB 10,000 per year up to age five for a second child, for a total of RMB 50,000; and will grant a subsidy of RMB 10,000 per year for 10 years for a third child or additional children. The announced subsidies in Hohhot also included a plan to provide milk to parents for one year after a child is born, through coupons for dairy products worth RMB 3,000. “Hohhot’s initiatives resonate with the government’s recent policy direction,” Goldman Sachs Research analyst Leaf Liu and her colleagues write in a report. How is China attempting to increase birth rates? A few days after Hohhot’s announcement, China’s government unveiled a special action plan that signaled the potential for more childcare subsidies nationwide. The plan reinforced policies to promote consumption that emerged from the annual plenary sessions of the National People’s Congress and the Chinese People’s Political Consultative Conference in Beijing earlier in the month. The subsidies for parents in Hohhot are high compared with similar programs announced in recent years in other Chinese cities, Andrew Tilton, chief Asia Pacific economist and head of Emerging Markets Economic Research, writes in a separate report. The macroeconomic impact will be limited if Hohhot is the only place offering subsidies at that level. Still, Goldman Sachs economists estimate that these types of supports, if implemented nationwide, could add between 0.1 and 0.3 percentage point to annual GDP. Shares of dairy companies that can benefit from these measures in China have risen: A basket of stocks that includes large makers of liquid milk, milk powder, and infant formula rallied more than 7% in just a few days. China’s fertility policy could boost stocks outside China Companies in Europe may also benefit from China’s efforts to boosts birth rates and provide greater support for families with young children, Georgina Fraser, head of the European Chemicals team, writes in a separate report. Policies to increase domestic consumption and enhance citizens’ quality of life could drive more demand for premium and higher-value dairy products. Investors may find opportunities in biotechnology companies that have engineered human milk oligosaccharides (HMOs), a type of carbohydrate that occurs naturally in human breast milk and promotes immune health and gut function. “The commercialization of HMOs is on the back of more favorable regulation,” Fraser writes. By 2030, there may be HMOs in 50% of the infant formula produced worldwide, up from just 5% today, she says in her team’s report. Some European companies make HMOs. Fraser writes that the market for these products could broaden across age groups. “HMOs are increasingly being recognized for supporting immune and gut health for a broader demographic,” Fraser writes. The outlook for demographics in China Births have been falling in China for years, but they rose in 2024. There’s further room for birth rates to rebound, Liu writes. Mothers aged 20 to 24 are estimated to be having children at half the pace they were before the pandemic, and mothers in the 30 to 44 age range have a birth rate notably below levels seen in Japan and South Korea for that age range. As a result, there’s scope for a recovery in birth rates. If policy support for having more children turns out to be significant nationwide, “our population model points to a potential uptick in new births” over the next decade, Liu writes.

Chinese measures to raise birth rates are boosting dairy stocks Read More »

Nvidia’s Jensen Huang dissects the AI revolution

Superfast chips are in high demand — not just by the artificial intelligence industry, but also by companies that work in computer graphics, robotics, autonomous vehicles, or drug discovery. “It’s fun to see all these amazing applications being created,” says Jensen Huang, the CEO of Nvidia. Speaking to David Solomon, the CEO of Goldman Sachs, at the Communacopia + Technology conference in San Francisco, Huang explained how computer graphics, for example, rely heavily on AI infrastructure. “We compute one pixel, and we infer the other 32,” he says in an edition of Goldman Sachs Talks. “Computing one pixel takes a lot of energy. Inferring the other 32 takes very little energy, and you can do it very fast. And the image quality is incredible.” Given this speed and flexibility, this infrastructure more than pays for itself, Huang says, responding to a question from Solomon about returns on investment for customers. By spending on such equipment, “the computing cost goes up a little bit — maybe it doubles,” Huang says. “But you reduce the computing time by a factor of about 20. You get 10x savings.” How Huang sees the data center market Chips that accelerate computing are everywhere, but there is no such thing as a universal accelerator, Huang says. Instead, every time a chip company enters a new market, it must learn new algorithms. They differ according to purpose; the algorithm for image processing would be different from the algorithm to model fluid dynamics. “Usually, some 5-10% of the code represents 99.999% of the run time,” Huang says. “So if you take that 5% of the code and offloaded it onto an accelerator, then technically you should be able to speed up the application a hundred times.” The promise of this kind of accelerated computing has led to keen investor interest in the data center market, Huang says. He thinks this infrastructure can yet be improved. For one thing, the average data center is “super-inefficient, because it’s filled with air, and air is a lousy conductor of electricity.” Making data centers denser — eliminating the air, in other words — will make them cheaper and more energy efficient. Another revolution lies in how data centers now understand not just how to process data but the meaning of the data itself, and how to translate one form of data to another, Huang says: “English to images, images to English, English to proteins, proteins to chemicals.” The chip supply chain needs to be resilient The ecosystem of manufacturers and suppliers to the chip industry is sprawling and complex, and particularly concentrated in Asia. As a result, Nvidia tries to design diversity and redundancy into every aspect of its supply chain. Companies need to have “enough intellectual property” to be able to shift their manufacturing from one “fab” — or chip-making facility — to another if they have to, Huang says. “Maybe the process technology won’t be as great, or you won’t get the same level of performance or cost, but you will still be able to provide the supply.”

Nvidia’s Jensen Huang dissects the AI revolution Read More »

Can generative AI overcome questions around scalability and cost?

With the world’s biggest companies racing to build the most sophisticated artificial intelligence, questions about how far and how fast the technology can be scaled are coming into focus, according to Goldman Sachs Research.  The advent of generative AI has created a surge of excitement about the future of the technology. Unlike other types of AI, gen AI can create its own outputs in natural language. Because it’s “multimodal,” it can also generate responses in other formats, including text, numbers, videos, and sound. Large tech companies are citing initial use cases, and some enterprise players see scope for major productivity gains in certain areas such as code writing, which could make the most valuable employees even more productive. However, the exact size of future economic benefits is the subject of debate. The cost of building gen AI at scale is extremely high, with big tech companies investing hundreds of billions of dollars — although the cost per query has come down considerably since the technology first launched. At the inaugural Goldman Sachs European Virtual AI and Semis Symposium, 20 speakers — from CEOs and technologists to macro economists — came together to assess the prospects for AI. In particular, they discussed key topics including use cases, the total addressable market, challenges for further development and adoption of the technology, and implications for European hardware and semiconductors. We talked to Alexander Duval, head of Europe Tech Hardware & Semiconductors in Goldman Sachs Research, about the main findings of the symposium. What role are the large tech companies known as “hyperscalers” playing in the development of AI? So far, US tech giants have been at the vanguard of generative AI use. They’ve been developing large language models, which they can use both for their existing business and also potentially in creating new business tools. The symposium heard how the technology is generating a quarter of one hyperscaler’s code and saving meaningful engineering time for another. Broader use cases in the real economy include its use to predict the structure of proteins, and even to “de-age” an actor’s appearance in a movie. Those are striking examples. But it’s worth bearing in mind that hyperscalers have been spending hundreds of billions on this. Together, they have spent around $200 billion on AI this year, and that will probably increase to $250 billion next year. Developing large language models can cost tens or hundreds of millions of dollars. And that’s why at this symposium, we really wanted to look at whether it’s feasible or desirable that the technology could scale to address many more use cases. These hyperscalers have a lot of free cash flow, and we are starting to see examples of use cases, but a number of industry observers believe that at some point we need to see a return on investment for a broader array of use cases and verticals. Have any key use cases emerged for artificial intelligence in the broader economy? Because generative AI is multimodal, it could theoretically apply to multiple fields: customer support, coding, medical analysis, marketing and many others. Given that there is a very significant level of investment in AI, the aggregate benefit of such use cases will need to be demonstrated in order to justify a solid return on investment. That being said, some participants at the symposium said that it might not be imperative for AI to scale in one particular area — in other words, a single key use case may not be necessary — as long as the economic benefits from all the different use cases are sufficient in aggregate. You could see efficiency gains across the board. Some speakers pointed out there are a number of examples of very large successful tech businesses where you could argue that there wasn’t a key use case at first. Take the example of ride hailing apps. There was already a perfectly good solution: Walking to the end of the street and hailing a taxi physically. But by leveraging software and network effects, you could create very large economic benefits, as well as benefits to consumers. Is there still room for smaller technology companies to compete? Some speakers at the symposium had interesting insights on small language models. At first, technology players were focused on building large language models — and those are still important. But there is also a trend of developing smaller and more efficient models. Small language models are easier to fine tune, they may have lower energy consumption, and they can be customized to meet an enterprise’s specific requirements in a given domain (such as legal, medicine, or finance). They’re also generally less expensive, because they’re smaller and use less power. Large language models will remain important, and tech behemoths have the resources, free cash flow, and balance sheets to drive the development of those. But speakers pointed out that there will be other, perhaps smaller players in the ecosystem who can innovate and develop small language models that will sit on top of those larger models. Some speakers thought this presented an opportunity for small companies to drive innovation at the top of the stack and highlighted the large number of companies being founded daily to do so. Could the high cost of generative AI hold back development? Training LLMs requires very high levels of capital investment. You need to build a data center, you need all the semiconductors — that includes both GPUs and memory chips — and you need hardware, power, and utilities. Speakers mentioned that the cost per query in some domains is multiple times higher than for a regular search algorithm.  That said, there has been steady progress on reducing costs. The cost of a generative AI query at some large tech companies has come down significantly since the launch of the technology, and a new gen AI company has said that revenue generated by the latest generation of LLMs exceeded the cost of training prior models. While some speakers stated there could be a risk that spending on AI could reduce if significant returns

Can generative AI overcome questions around scalability and cost? Read More »

What to expect from AI in 2025: hybrid workers, robotics, expert models

More than two years after ChatGPT’s debut, Goldman Sachs Chief Information Officer Marco Argenti says the potential for generative artificial intelligence is coming into focus. The development of increasingly powerful large language models (LLMs), supercharged by advances in robotics will, in Argenti’s view, begin to bring sweeping changes to everything from employment to regulation of the technology in 2025. Argenti, the former vice president of technology of Amazon Web Services, makes five predictions about how AI could evolve and interact with businesses and society in the near future: The new hybrid workforce: AI systems are becoming more like humans. So why not employ them like humans? The capabilities of AI models to plan and execute complex, long-running tasks on humans’ behalf will begin to mature. This will create the conditions for companies to eventually “employ” and train AI workers to be part of hybrid teams of humans and AIs working together. The question becomes: How best to get humans and AI to work together. Companies will reskill human managers to oversee a hybrid workforce. The role of human resources will evolve into a department for human and machine resources. The first AI “layoffs” could eventually emerge, in which AI models will be replaced by better AI tools or humans if they perform poorly compared to their peers. The emergence of expert AI: The AI version of PhDs will arrive. Companies will integrate AI with their proprietary data, either with retrieval-augmented generation (RAG) — an architecture that can connect LLMs to external, specialized datasets — or via a process known as fine-tuning, which involves enhanced training of an LLM with a smaller, specialized dataset. As a result, expert AI systems, or large expert models, will gradually emerge with advanced capabilities and industry-specific knowledge — for example, specialized models for medicine, robotics, finance, or material sciences. Robotic breakthroughs powered by AI: So far, AI models have been trained by reading essentially all the books in the world. What if they’re trained on the world itself? Children learn to walk before they learn to read. In the same way, the intersection of LLMs and robotics will increasingly bring AI into, and enable it to experience, the physical world, which will help enable reasoning capabilities for AI. At the same time, these models will transform commodity hardware into specialized components capable of performing far beyond their default capabilities. Advanced cameras using cheap sensors, studio-quality microphones using low-cost transducers, and off-the shelf mechanical joints capable of performing complex manipulation tasks will drive down costs for combining advanced AI with robotics and will speed up innovation. Regulation goes from global to local: As the world awaits regulatory clarity, responsible AI principles will take center stage for CEOs and boards.   In addition to (and somewhat separate from) national, state, or sectoral regulations, companies across sectors will continue to see the benefit of implementing proper controls, such as responsible AI principles (i.e., a form of self-regulation). Responsible AI will become an even bigger priority for CEOs and boards of major companies. Large model consolidation: The Formula One experience arrives for AI. Given the cost and complexity of engine development in the Formula One motorsport competition, there are many cars but only a few engine makers. Likewise, the investment required to train and maintain large frontier models (those that are the largest and most advanced) for AI will eventually result in only a handful of providers. Consolidation will mirror what has taken place in cloud infrastructure, databases, and operating systems, where the total number of companies developing large AI engines will be countable on one hand. Startups that are now “model-centric” will shift towards building solutions that are model-agnostic, focusing instead on other aspects of AI such as compliance, safety, data integration, orchestration, automation, and user experience.

What to expect from AI in 2025: hybrid workers, robotics, expert models Read More »