For two years, Jim Covello has been asking the same uncomfortable question: when does Wall Street\u2019s AI bet actually pay off?<\/p>\n
The head of global equity research at Goldman Sachs\u00a0raised the alarm in a widely read 2024<\/a>\u00a0report, questioning whether the torrent of capital flooding into artificial intelligence would ever generate returns commensurate with the spending. At the time, he estimated it would take 18 months to 2 years. Two years later, speaking on Goldman\u2019s\u00a0Exchanges\u00a0podcast<\/a>, he noted that he\u2019s been having this debate for close to four years now \u2014\u00a0and his answer has only grown more pointed.<\/p>\n \u201cAt some point, you\u2019ve got to make money,\u201d Covello said. \u201cYou make investments in a business so that you can generate returns and make money. And we\u2019ve gotten further away from that over the last couple years instead of closer to it.\u201d<\/p>\n Speaking with Allison Nathan, Goldman\u2019s senior macro strategist, and George Lee, co-head of the Goldman Sachs Global Institute \u2014 who has long taken a more optimistic view of AI\u2019s potential \u2014 Covello laid out three reasons why that patience may be running thin and asked another question that runs counter to the market narrative.<\/p>\n \u201cIf we\u2019re having the same debate two years from now and we\u2019re still saying, \u2018Well, it\u2019s early,\u2019 then we might have a challenge,\u201d Covello said, \u201cbecause at some point, when does the short-term become the long-term?\u201d<\/p>\n Anthropic and OpenAI weren\u2019t mentioned specifically on this podcast, but they are both nearing mega-IPOs, they are both valued at close to $1 trillion, and neither is profitable.<\/p>\n The hill keeps getting steeper<\/p>\n Covello acknowledged being wrong on some counts. Consumer adoption of AI has been, in his words, \u201cmagnificent\u201d \u2014 far exceeding his expectations. And the technology itself has advanced rapidly. But on the question that matters most to investors, his conviction has hardened rather than softened.<\/p>\n \u201cIn a lot of ways, companies are losing more money today implementing this technology than they were two years ago,\u201d he said. \u201cThe hill that has to get climbed is even steeper today than it was before, because we\u2019ve spent more money.\u201d<\/p>\n Goldman Sachs Global Institute<\/a>, which takes a more optimistic long-term view of AI, nonetheless agrees the math is daunting. Co-head George Lee estimated that $7 trillion to $8 trillion will ultimately be spent on AI infrastructure \u2014 and that simply disrupting existing profit pools won\u2019t generate sufficient payback. Net new economic activity, Lee argued, is the only way the numbers eventually work.<\/p>\n The critical variable for both men is the same: enterprise ROI. And that, Covello said, remains the central unanswered question. \u201cI really think it all boils down to one thing: do the enterprises make or save money implementing AI? If they do, this technology is going to fulfill its promise.\u201d But if, two years from now, people are still saying it\u2019s \u201cearly\u201d in the adoption wave, \u201cthen we might have a challenge.\u201d<\/p>\n So far, the evidence is thin. Widely cited MIT research<\/a> found 95% of organizations reporting zero return on AI pilots. A 2025 EY survey<\/a> found 99% of companies reported financial losses tied to AI-related risks, averaging $4.4 million per company.<\/p>\n The data tracks with what Fortune heard directly from executives at its COO Summit earlier this week. Cognizant, whose research team presented new findings at the event<\/a>, reported that 93% of jobs are already being disrupted by AI \u2014 six years ahead of their own 2023 projections \u2014 and yet the productivity gains that were supposed to follow haven\u2019t materialized. Their researchers called it an \u201cactivation gap.\u201d That dynamic surfaced in the executive debates, too: Francine Katsoudas, EVP and Chief People Officer at Cisco, noted that on teams using AI most intensively,\u00a0trust within those teams began to drop<\/a>\u00a0about 9 months in \u2014 a warning sign that the human-organizational dimension of AI deployment remains deeply unsolved. \u201cWe just have to invest so much more,\u201d she said.<\/p>\n The capex paradox<\/p>\n Part of what makes Covello\u2019s frustration notable is the sheer scale of spending without proof of return. Hyperscalers \u2014 Amazon<\/a>, Microsoft<\/a>, Google<\/a>, Meta<\/a> \u2014 have not only maintained their AI capital expenditure despite stock underperformance, but they\u2019ve also increased it. Covello predicted two years ago that sustained underperformance would trigger spending discipline. It didn\u2019t.<\/p>\n \u201cThere\u2019s a tremendous amount of FOMO at every level of the supply chain,\u201d he said, describing a dynamic where every company, from enterprise to model layer to hyperscaler, fears being left behind if a competitor cracks the economic code first. (The FOMO theme has been on Covello\u2019s mind<\/a> in recent months.)<\/p>\n At Fortune\u2018s Most Powerful Women Summit last October, Brilliant Earth chief brand officer Pam Catlett put it plainly<\/a>:\u00a0\u201cFor me, there\u2019s the pervasive sense of FOMO that\u2019s happening \u2014 fear of missing out. And the first word, fear, is not a good state to be in when you\u2019re thinking about how to better serve your customer.\u201d<\/p>\n The result, Covello said, is an unusual inversion in the supply chain. Semiconductor companies \u2014 led by Nvidia<\/a> \u2014 are capturing nearly all the economic value being created, while the companies above them are bleeding cash. Covello said this is historically unprecedented \u2014\u00a0and he covered semiconductor stocks for 16 years. \u201cIn every cycle, the semiconductor stocks thrive when their customers thrive. Here in this cycle, the semiconductor companies are thriving at the economic expense of everybody above them in the chain.\u201d<\/p>\n He now favors hyperscaler stocks over semiconductors \u2014 a notable reversal \u2014 arguing that in two of three likely scenarios, hyperscalers win. Only a pure status quo, where semis alone keep profiting indefinitely, keeps the chip trade alive. \u201cThat can\u2019t go on forever,\u201d he said.<\/p>\n Workers aren\u2019t feeling it<\/p>\n A second pressure front is building inside companies themselves. Third-party surveys consistently document a widening gap between C-suite enthusiasm for AI and what workers on the ground are actually experiencing. Covello pointed to the pattern bluntly: \u201cThe line workers aren\u2019t getting as much benefit from it as the C suite expected.\u201d<\/p>\n Part of the explanation, he said, is a data readiness problem that rarely makes headlines. \u201cThere are agents today that are terrific. There are models today that are terrific. But in many cases, the data isn\u2019t ready to be agented yet. So we\u2019re putting agents on top of data that isn\u2019t ready to be agented. And that\u2019s creating another economic challenge for companies.\u201d<\/p>\n Large incumbent companies also face a structural drag that AI-native startups don\u2019t, Goldman\u2019s Lee noted \u2014 legacy systems, entrenched workflows, and resistance to change create a \u201cdrag coefficient\u201d that slows the realization of productivity gains at scale. The productivity leap is real, he argued, but primarily visible in companies built from scratch for this technology. Getting there for everyone else will take longer.<\/p>\n These are problems that Yale\u2019s Jeffrey Sonnenfeld examined in depth in Fortune last month<\/a>. In a piece titled\u00a0\u201cWhy your data infrastructure \u2014 not your AI model \u2014 will determine whether agentic AI scales,\u201d\u00a0the\u00a0Lester Crown Professor in the Practice of Management argued that the bottleneck in enterprise AI deployment isn\u2019t model quality but the organizational plumbing beneath it \u2014 siloed systems, inconsistent data governance and infrastructure built for a pre-AI world that simply wasn\u2019t designed to support autonomous agents making real-time decisions. Part of a series on agentic AI\u2019s enterprise implications, Sonnenfeld and his colleague Stephen Henriques found that data infrastructure readiness varies dramatically across sectors, with most large incumbents badly underestimating what it would take to make their data \u201cagentable\u201d \u2014 a word Covello, independently, used almost verbatim.<\/p>\n Politics enters the equation<\/p>\n Perhaps the most underappreciated pressure is the one building outside corporate boardrooms entirely. Lee raised it unprompted: a \u201csomewhat sudden turn towards populist resentment of AI\u201d that he described as \u201ca deeply unpopular technology \u2014 interestingly, almost uniquely in the US versus other parts of the world.\u201d<\/p>\n The signs are obvious. Mentions of AI are booed at commencement addresses by Gen Z sensing their futures being erased. Data center deployment has triggered threats and protests in local communities facing rising electricity costs. <\/p>\n President Trump convened seven major hyperscalers at the White House in March to sign a voluntary \u201cRatepayer Protection Pledge<\/a>\u201d \u2014 a political signal of just how quickly the issue has escalated. Data centers now consume over 4% of U.S. electricity<\/a>, with some government projections suggesting that could triple to 12% by 2028.<\/p>\n Covello connected the politics directly to the economics. \u201cI would put all of it in the broader economics bucket,\u201d he said. \u201cHow much of it is going to flow back to the individual? Can we make a case that individuals are benefiting economically from using the technology?\u201d Clearly, the implication is that the case is not being made.<\/p>\n The market, for now, is still granting a long leash \u2014 buoyed by a bull market that Covello himself acknowledged is partly\u00a0caused\u00a0by AI optimism, creating a circular dynamic that\u2019s difficult to unwind. But the patience is not unconditional, and three clocks \u2014 economic, operational, and political \u2014 are now running simultaneously.<\/p>\n \u201cThat doesn\u2019t mean it\u2019s never going to happen,\u201d Covello said of AI\u2019s eventual payoff. \u201cIt just means the stakes are higher.\u201d<\/p>\n","protected":false},"excerpt":{"rendered":"For two years, Jim Covello has been asking the same uncomfortable question: when does Wall Street\u2019s AI bet…\n","protected":false},"author":2,"featured_media":63620,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8],"tags":[53,15992,303],"class_list":["post-63619","post","type-post","status-publish","format-standard","has-post-thumbnail","category-anthropic","tag-anthropic","tag-bubbles","tag-goldman-sachs-group"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/63619","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=63619"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/63619\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/63620"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=63619"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=63619"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=63619"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}