Good morning. I dropped by ServiceNow’s World Forum in NYC yesterday to see what’s on the minds of enterprise tech leaders vis-à-vis AI—the good and the bad.
What did I overhear? As always, a haiku:
Fix the governance
Garbage in and garbage out
Process kills star teams
Today’s tech news follows. Have a wonderful weekend. —Andrew Nusca
Want to send thoughts or suggestions to Fortune Tech? Drop a line here.
Anthropic: We disrupted apparent efforts to develop biological weapons
Anthropic says it has disrupted several possible plans this year by scientists who used its AI models to research the development of biological weapons.
The San Francisco company wasn’t able to determine whether the scientists’ activity was a legitimate inquiry or a malicious attempt. It blocked it anyway in an abundance of caution.
The use of advanced generative AI to facilitate the development of advanced weapons has been a worry and risk almost since modern AI arrived on the scene. As well-meaning scientists use AI for breakthrough scientific discoveries like new drugs, there are concerns that malevolent researchers could use the very same processes to develop life-threatening pathogens.
Add the state-sponsored efforts to win the global AI race and there’s even more reason to be concerned.
Bioweapon research isn’t the only observed misuse of frontier AI, Anthropic said. Russian state media are trying to use it to generate propaganda that looks like real journalism, for example. And Chinese actors are trying to use it to design and develop conventional weaponry.
Anthropic didn’t name names, in part because it can’t determine intent. But it can make an educated guess. Some scientists purposefully dodged Anthropic’s blocks on certain regions. Others tried to hide their aims as they worked their way around Anthropic’s safeguards. (Such as consistent efforts to make a hypothetical virus more harmful.)
“As models become increasingly capable, their risks will increase,” Anthropic wrote, “unless AI developers and society’s defenders act to make them safer.” —AN
Intel-backed Altera may file for an IPO in the coming weeks
Altera, the programmable chip business that Intel spun off in January 2025, is reportedly preparing to file for an initial public offering in the next few weeks.
An IPO “could raise over $2 billion as early as 2026,” a Reuters report notes.
Intel still owns 49% of Altera. The remaining majority is controlled by the private equity firm Silver Lake, best known for taking a variety of household tech names private, including Dell, Electronic Arts, Endeavor, Qualtrics, and Skype.
If it happens, Altera would be one of the largest semiconductor listings in recent memory. The last big one was SoftBank-controlled Arm in 2023, which raised almost $5 billion.
It’s no secret that the IPO gates have swung open again, at least in the U.S. SpaceX broke records with its June 12 debut; peers Anthropic and OpenAI are imminent. Many more companies hope to draft behind them before market sentiment shifts again.
Altera makes a special kind of computer chip (most notably field-programmable gate arrays, or FPGAs) found in everything from financial trading systems and telecom networks to medical equipment and aerospace and defense systems.
Some computer chips are hardwired with function; Altera’s programmable chips can be repeatedly rewired with function. Standard CPUs execute software sequentially; FPGAs execute it in parallel, which is useful for certain tasks where reactive speed is important: routing data on a 5G network, guiding avionics, executing high speed stock trades, etc. —AN
OpenAI courts Wall Street with ChatGPT for financial services
OpenAI unveiled a new version of ChatGPT geared toward big banks and other financial firms, powered by its new GPT-6 Astra model.
The leading AI companies have increasingly sought to roll out products that package their AI models in ways that are tailored for specific industry verticals. They are hoping these more targeted products will help accelerate the growth of their enterprise revenue.
Nick Turley, OpenAI’s vice president and head of ChatGPT, said in a briefing for reporters that finance is one of the industry verticals OpenAI has chosen to build specific products for, along with cybersecurity and software engineering.
The company’s rival Anthropic already offers a version of Claude for financial analysis, which it debuted in July 2025. But OpenAI hopes its offering will be “the one product” large banks, or those with “tens of thousands of employees,” will ever need.
“This is the canonical product we are hoping the industry adopts,” Turley said.
Turley has been spending a lot of time in New York City, the center of the nation’s financial industry, to develop the product. His team worked directly with Morgan Stanley and boutique investment advisory firm Evercore to decide on the key features the product would include.
“There’s a difference between what looks good in a demo and what is actually a usable output, [and] you kind of rely on the experts” to achieve that, Turley said. —Emily Forlini
More tech
—Electronic Arts may merge with Savvy Games, but a $6 billion Moonton acquisition stands in the way.
—Nasdaq invests $100 million in Kraken parent company Payward, valuing the crypto exchange at $21 billion.
—Ant, Mastercard, and Visa collab on a new agentic payments standard. Agentic commerce, so hot right now.
—Bending Spoons will acquire Miro. The deal values the workplace collaboration platform at $1.36 billion.
—Amazon names Mandiant founder Kevin Mandia to its board of directors.
—Moonshot AI may list in both Hong Kong and Shanghai in the wake of HK’s recent weak debuts.
—Grab may buy a majority stake in Atome Financial at a $2 billion valuation. The buy now, pay later platform is owned by SoftBank-backed Advance Intelligence Group.