Wall Street isn’t immune from the plot line that has generative AI resulting in wholesale knowledge worker replacement. A new tool from AlphaSense, called Deep Research, won’t provide any comfort.
The generative AI agent functions like a team of analysts operating at what AlphaSense calls “superhuman speed,” generating research and market insights, and building investment-grade briefings.
But Jack Kokko, AlphaSense CEO and a former Morgan Stanley analyst and Wharton School MBA, isn’t worried about the job outlook for Wall Street professionals.
“It’s a popular narrative,” Kokko told CNBC of the job replacement fears during an interview on “The Exchange” on Tuesday after AlphaSense ranked No. 8 on the 2025 CNBC Disruptor 50 list. “But I would not be so sure,” he said.
More coverage of the 2025 CNBC Disruptor 50
What Deep Research does is tap into the AlphaSense universe of more than 500 million business and financial documents, which includes filings, press releases, content about public and private companies, and expert insights based on call transcripts. Last year, the company spent nearly $1 billion to buy Tegus and its library of a quarter-million business-focused interviews.
“There are a hundred on a single company, and no human can read it all, but Deep Research will read it all and ask questions,” Kokko said.
It will answer questions too, ones that Wall Street analysts are often paid to field, within minutes.
The company, which dates back to 2011 and has had Goldman Sachs Growth Asset Management as an investor since its origin, already offers rapid summaries of equity research and real-time customizable reports. And it already has a tool called Generative Search designed to think like an analyst, ask natural language questions and receive precise insights sourced from AlphaSense’s content, which covers 37 languages.
Any of of its enterprise intelligence customers in equities research, corporate development and finance, on or off Wall Street, will be able to plug their internal document libraries into Deep Research, which will then be able to take both pro and con positions, and offer internal and external perspectives, in a report generated in record time.
“It would have taken a human analyst days or weeks,” Kokko said. “I was an analyst,” he added.
Timothy A. Clary | Afp | Getty Images
The company says it counts majority of the S&P 100 as clients. That client base grew by about 25% in 2024, to more than 5,000, including Amazon, Nvidia, Microsoft, Pfizer and JPMorgan.
For companies making investments that run into the millions or billions of dollars, being able to make these decisions on the back of all of this information is a revelation, Kokko said, citing the experience of a private equity firm that told AlphaSense that Deep Research did the same or even better on a report the AI ran than its in-house analysts could do in weeks.
There are plenty of reasons to believe that this is all bad news for knowledge professionals like finance bros. And more CEOs are starting to talk that way, from Shopify’s CEO who recently said no job hire requisitions will be approved unless a manager can prove that the job can’t be done by AI, to fellow Disruptor Anthropic‘s CEO Dario Amodei, who recently said AI would wipe out up to half of entry-level office jobs and whose latest Claude model can work 7 hours straight without a break or burnout.
Wall Street’s long embrace of AI has only accelerated in the wake of OpenAI’s arrival in 2022. Last August, JPMorgan Chase rolled out a generative artificial intelligence assistant to employees that can help them with tasks like writing emails and reports, while Morgan Stanley has already released a pair of OpenAI-powered tools for its financial advisors. In January, Goldman Sachs gave its bankers, traders and asset managers access to a generative AI assistant, the first stage in the evolution of a program that will eventually take on the traits of a seasoned Goldman employee, Goldman Sachs Chief Information Officer Marco Argenti told CNBC.
But Kokko says the Wall Street bonus — for now, and as he sees it, into the future — is safe. He is still of the belief that the latest AI will enhance the jobs of Wall Street analysts rather than threaten them. “What it does is make human analysts and business people so much more productive,” he said. “That person will be operating with a higher ROI [return on investment] and companies don’t cut high ROI people,” he added.
What AI job doomsday soothsayers are dismissing too easily today is “the top line expansion that comes from being able to do things in a much more agile way,” Kokko said.
“It’s 10x prior productivity when it is you and the machine,” he added.
Sign up for our weekly, original newsletter that goes beyond the annual Disruptor 50 list, offering a closer look at list-making companies and their innovative founders.
Roblox on Friday announced new short-video and AI features that come amid increasing lawmaker scrutiny into how the company protects children on its platform.
With Roblox Moments, users 13 and older will be able to create and share video clips of their gameplay with others on a feed within the platform. The artificial intelligence additions, meanwhile, will allow users to generate advanced 3D objects for the games they create on the platform.
Tune in at 4:15 p.m. ET: Roblox CEO Dave Baszucki joins CNBC TV to discuss the company’s latest announcements coming out of its developer conference. Watch in real time on CNBC+ or the CNBC Pro stream.
Although users can share video clips from mature games on Moments, users who do not meet the age requirements of those experiences will be unable to view them, the company said. Roblox will moderate each video shared on Moments and will allow users to “flag content that they find is inappropriate,” said Matt Kaufman, Roblox safety chief. Moments is launching in a limited release on Friday.
The AI features will roll out to users before the end of the year. Roblox users will be able to use the artificial intelligence tools to create objects, like futuristic monster trucks, that match the aesthetics of the games users build on the platform, said Anupam Singh, Roblox’s senior vice president of engineering. Those creations will also be moderated, Kaufman said.
Roblox faces a number of lawsuits alleging that its design enables online predators to exploit underage victims.
Louisiana Attorney General Liz Murrill sued Roblox in August, alleging the company fails to implement robust safety protocols to “protect child users from predators.” At the time, Roblox said, “any assertion that Roblox would intentionally put our users at risk of exploitation is simply untrue.”
The company on Wednesday announced it would expand an age estimation program that Roblox debuted in July.
Google was on Friday hit with a 2.95-billion-euro ($3.45 billion) antitrust fine from European Union regulators for anti-competitive practices in its lucrative advertising technology business.
The European Commission, which is the executive body of the EU, accused Google of distorting competition in the so-called adtech market by unfairly favoring its own display advertising technology services to the detriment of rival adtech providers, advertisers and online publishers.
It also ordered Google to “bring these self-preferencing practices to an end” and “implement measures to cease its inherent conflicts of interest along the adtech supply chain.” The company has 60 days to respond.
“Today’s decision shows that Google abused its dominant position in adtech harming publishers, advertisers, and consumers. This behaviour is illegal under EU antitrust rules,” EU competition chief Teresa Ribera said in a statement Friday.
“Google must now come forward with a serious remedy to address its conflicts of interest, and if it fails to do so, we will not hesitate to impose strong remedies.”
Google’s global head of regulatory affairs, Lee-Anne Mulholland, said the EU decision is “wrong” and the firm will appeal.
“It imposes an unjustified fine and requires changes that will hurt thousands of European businesses by making it harder for them to make money,” Mulholland said. “There’s nothing anticompetitive in providing services for ad buyers and sellers, and there are more alternatives to our services than ever before.”
The case dates back to 2021 when the EU first opened a probe into Google to assess whether the tech giant favors its own online display ad technology services.
The news comes after Reuters reported earlier this week that the Commission had delayed the fine as regulators were waiting for the U.S. to cut tariffs on European cars as part of a trade deal.
Broadcom shares soared 15% on Friday after the chipmaker said on its earnings call that it had secured a new $10 billion customer. Analysts quickly pointed to OpenAI.
Following a better-than-expected earnings report late Thursday, Broadcom CEO Hock Tan told analysts that a fourth large customer had put in orders for $10 billion in custom artificial intelligence chips, which the company calls XPUs.
“One of these prospects released production orders to Broadcom, and we have accordingly characterized them as a qualified customer for XPUs,” Tan said. He added that the order increased Broadcom’s forecast for AI revenue next year, when shipments will begin.
Analysts at Mizuho, Cantor Fitzgerald and KeyBanc all said they think AI startup OpenAI is the customer. The Financial Times reported on Thursday, citing people familiar with the partnership, that the two companies co-designed a chip that will hit the market next year.
OpenAI declined to comment on the report.
While Broadcom doesn’t name its large web-scale customers, analysts have said dating back to last year that its first three clients were Google, Meta and TikTok parent ByteDance.
“During the call, the company surprised us by noting that it had secured a $10B order from a fourth XPU customer (we believe this is OpenAI), adding significant upside to the company’s three current XPU customers (Google, Meta, and ByteDance),” analysts at Cantor wrote in a note late Thursday. “Shipments are expected to commence in 2026.”
Broadcom’s stock has been on a tear of late as the company has joined Nvidia at the front of the race to build the kinds of processors and infrastructures needed for massive AI workloads. The stock is up about 130% in the past year, lifting Broadcom’s market cap past $1.6 trillion.
For the fiscal third quarter, Broadcom reported earnings and revenue that topped estimates. The company said it expects $17.4 billion in fourth-quarter revenue, higher than the $17.02 billion expected by Wall Street analysts, with AI revenue reaching $6.2 billion.
But news of an incoming $10 billion customer is what got Wall Street excited.
Tan said on the call that “immediate and fairly substantial demand” boosts the outlook for next year, “and really changes our thinking of what 2026 would be starting to look like.”
The company didn’t provide specific guidance for next year, but Tan suggested that growth in its AI could be above the 50% to 60% range he’d offered in the prior call.
Analysts at Mizuho raised their AI revenue growth estimate for next year to 76% up from about 60%, which would bring the total to $35 billion. Total revenue for the year ending in October 2026 is expected to increase about 30% to $81.8 billion from $63.1 billion this fiscal year, according to analysts surveyed by LSEG.
In addition to hardware, Broadcom has a large software business, keyed by its $61 billion acquisition of server virtualization software vendor VMware in 2023. Revenue in the infrastructure software business, which includes VMWare, rose 43% to $6.79 billion.