Traversal has raised $120 million in funding to help companies like American Express and Pepsi troubleshoot site outages.
The internet is being rebuilt, says Traversal CEO Anish Agarwal. With AI writing the code, software companies are shipping more applications and more features than ever, and faster.
The catch is what happens when all that new software breaks. It will, and when it does it will be harder to diagnose and patch than the old kind built from the ground up by humans. “They break in fundamentally new ways,” Agarwal says. An AI system might look fine on the surface while the model underneath has “drifted” into behavior no one intended or expected. An update might pass every test it was written to perform and then fail in scenarios no one thought to write.
Traversal, founded in 2023, investigates those failures for companies like Pepsi, Blackstone and American Express: a site that goes down, a bug that trips alarms across different apps, an update that starts breaking features across a website. Its AI agents comb through billions of logs and metrics in minutes, work out what went wrong and suggest a fix.
The startup announced Tuesday it has raised $120 million in a funding round co-led by Kleiner Perkins, Redpoint Ventures, and Khosla Ventures. The fresh capital brings its valuation to $1 billion. The startup says its annual recurring revenue has grown nine times in the past six months. Agarwal declined to share exact revenue figures. Its AI agents helped companies run over 150,000 investigations last quarter. AI has “created acuteness around the problem of cyber reliability,” says Kleiner Perkins Partner Mamoon Hamid.
The old way of handling an outage was to throw engineers at it, by the hundreds. One bank, which now uses Traversal, grouped thousands of its engineers in a so-called “war room” to find out why customers couldn’t log into their accounts, while every minute offline cost the company money. “The issue was actually a specific database deep in the bowels of their system that was running slowly,” Agarwal says. When a similar problem occurred a few months later, Traversal’s agents troubleshooted it in five minutes.
Now let’s get into the headlines.
BIG PLAYS
Anthropic’s revenue grew to $17 billion in the third quarter this year, according to its amended S-1 filing. But the surging revenue comes at a cost— the AI titan posted $23.3 billion in net losses and $567 million in operating losses in the last quarter. The company revealed that it plans to spend about $895 billion in spending commitments, most on data centers and chips.
Anthropic isn’t the only company earmarking more funds for AI infrastructure. Oracle plans to sell $20 billion of debt and equity to bankroll its AI buildout. The company’s shares fell 8.5% after the announcement.
COMPUTE
Data centers have become a key issue in the midterm elections this year, especially in states like Ohio, Pennsylvania and Michigan. The industrial-scale buildout of power hungry AI data centers has placed more strain on the national power grid and led to huge spikes in electricity prices— with utility bills going up as much as $18 per month, Forbes reported. Several candidates have proposed moratoriums on new data centers.
Also notable: Elon Musk’s SpaceXAI is offering to buy over 400 homes in Mississippi if residents promise not to join a class action lawsuit over noise pollution caused by its data center’s gas turbines. One resident said the noise is so “torturous” she can’t sleep.
HUMANS OF AI
OpenAI cofounder and president Greg Brockman and his wife, Anna Brockman, are among the biggest mega donors this midterms cycle. The power couple, worth about $30 billion, donated $25 million to MAGA Inc and $25 million to pro-AI PAC Leading The Future last year. But very little is known about Anna, the wife of one of the most powerful people in AI, according to a recent Vanity Fair profile. A onetime registered Democrat, Anna is an animal lover (a source says Greg and Anna live with over 20 animals) and an idealist who believes AI could solve major social problems.
SHOW ME THE MONEY
Tech’s hottest new job is no longer an AI researcher. Forward deployed engineers— engineers who are embedded within customers’ organizations— are drawing between $586,000 to $1.45 million annual salaries, the Wall Street Journal reported. As experts in AI, these engineers can help slow moving enterprises identify what to automate and how to do it quickly.
AI DEAL OF THE WEEK
Serval helps companies like Ramp, Notion and Perplexity automate everyday employee tasks like servicing IT tickets, filing expense reports and requesting time off. The company has raised $150 million at a $3 billion valuation, Forbes reported. The startup plans to hire over 100 “automation engineers” straight out of college and place them inside companies.
DEEP DIVE
In 2025, when AI startup Spellbook signed a long-term deal with Anthropic to use its models to review legal contracts, CEO Scott Stevenson wasn’t worried about the AI giant eating his lunch. Anthropic staffers told him explicitly they didn’t see the firm directly competing with Spellbook in the legal space, he recalls.
A year later, he was caught off guard when Anthropic informed him that it was launching Claude for Legal. The tool, tailored for law firms, has dozens of capabilities ranging from contract review to citation checking to tracking regulatory changes — many of which overlap with Spellbook. That made things, “a little bit awkward,” Stenvenson says. He decided to be more cautious sharing data and product roadmaps with the AI giant.
“Anthropic has been a great partner but it made us realize that they are a massive company and two parts of the org can have conflicting mandates,” he says. “Ultimately this is inevitable.” Up until recently, Spellbook used Claude to carry out over 500,000 monthly contract reviews for its customers, but it’s now moved a lot of its spend to OpenAI.
Spellbook is far from the only company navigating a tricky relationship with Anthropic. A large swathe of startups relies on the AI behemoth for its top-tier models, widely considered among the best in the industry. But Anthropic’s dominance and aggressive expansion has led some of its customers to feel as if they are walking a tightrope. At any time, the AI company could train its models to directly compete with their own products.
Read the full story on Forbes.
MODEL BEHAVIOR
AI avatars are stealing jokes from young female comics, copying their punchlines word-for-word and attracting hundreds of thousands of followers, The New York Times reported. The realistic-looking AI avatars, dressed in crop tops and bodysuits, are for the most part making jokes about dating and sex. In early October, comedians sounded the alarm about bots stealing their jokes.
Want to see more Forbes articles on your feed? Tap here to make Forbes Australia a preferred source on Google.
Look back on the week that was with hand-picked articles from Australia and around the world. Sign up to the Forbes Australia newsletter here or become a member here.
