We're In A Bubble When...

You can play a fun game with friends while waiting in line at the Employment Office. Take turns completing this sentence: you know we’re in a bubble when…

You know we’re in a bubble when Satya pays $10 billion for a chatbot.

You know we’re in a bubble when C3.ai trades at $177, or infinity times earnings.

You know we’re in a bubble when Dan Wright raises $55 million for a pre-revenue startup. That’s the same Dan Wright who presided over DataRobot’s slide from $30 a share to $3. A 10X disruption, but in the wrong direction.

Dan’s new company is Armada.ai. What does Armada dot ai offer? It’s the “world’s first full-stack mobile edge computing platform delivering internet connectivity, computing, and AI anywhere on the planet.”

Say that three times fast.?

Armada got a puff piece by Harvard grad Alex Konrad in Forbes to tout the launch. Some nuggets:

On oil rigs and at remote mines — not to mention advanced military bases, fire prevention outposts or even at elite surfing competitions — one basic problem slows the easy use of exciting new artificial intelligence models: data. By remote sensor, camera and drone, they’re generating it by the terabyte. Then it usually just sits there.

Remarkable. Palantir’s been doing this for 20 years, and they left all that data just sitting there. Dan Wright's in the saddle and riding to the rescue.

?“Nothing is being done with this data, and that just seemed crazy to me,” Dan Wright, CEO of startup Armada, told Forbes. “Once I got into the problem of bridging that digital divide, I couldn’t help it; I just couldn’t stop.”

Dan Wright “getting into” something – is that an endorsement?

When Dan talks tech, he sounds like George H.W. Bush seeing a supermarket scanner for the first time. “Gee whiz, will ya look at that!”??

While startups like OpenAI and Anthropic have raised billions of dollars in a race to build bigger and more powerful AI models, Armada is one of the most promising of another wave of startups looking to unlock their capabilities for business uses far from Silicon Valley or an Amazon Web Services data center.?

Another difference: OpenAI and Anthropic have customers.

Armada has no customers beyond a proof-of-concept trial, meaning its revenue remains at zero so far.

Yes, that’s generally the case for companies with no customers.

Armada’s valuation is now already approaching $250 million after the funding, a source with knowledge told Forbes.

$250 million post-money valuation minus the $55 million invested equals a $195 million pre-money valuation, which is only infinity times sales.

Startup valuations are bullshit.

Armada represents a high-difficulty, high-upside new act for Wright, who was last seen in the public eye resigning from his previous role as CEO of AI software unicorn DataRobot in July 2022 amid layoffs and an insider trading scandal.?

Insider trading is a felony; if Wright had done that, he’d be in jail. The scandal was about Wright’s hypocrisy – telling investors and employees about DataRobot’s “high upside” while selling his own shares. That’s not a crime, it’s bad leadership.

Wright declined to comment on that situation beyond a written statement.

Shocker.

A lawyer by training — he worked with cofounder Runyan at a Bay Area law firm serving startups more than a decade ago — Wright was general counsel at app performance manager AppDynamics when, two days before it was scheduled to go public in January 2017, it was acquired in a shock move by Cisco for $3.7 billion. Wright had been promoted to chief operating office before leaving for DataRobot in 2020, where he became CEO the following March.?

Note the career trajectory. Law firm. General Counsel. COO of a subsidiary. CEO of a unicorn that went splat.

Great leadership story. Our next Jack Welch.

Runyan, meanwhile, was coming off a long stint as general counsel of access manager Okta, including during its 2017 IPO.

Runyan “retired” from Okta in March of this year. Bloomberg Law reported: “Okta Inc., an authentication software company coping with layoffs and a shareholder lawsuit, is parting ways with general counsel Jonathan Runyan.”

A shareholder lawsuit filed in April alleges insider trading and includes Runyan among the defendants. In its most recent 10-Q Okta says it will vigorously defend these lawsuits. No doubt. But without Mr. Runyan. Doesn’t seem like a vote of confidence.

Runyan earned $15.7 million in his last year at Okta, or $52,333 for each layoff.

For what could prove to be a capital intensive, complicated business with multiple products and large-sized target customers that aren’t used to working with startups, Wright and Runyan seemed ideal founders, investors said, their networks more than making up for a lack of deep AI research chops, or direct hardware experience.

Networking helps when you don’t know shit about the tech you’re peddling.?

And investors were unfazed by Wright’s exit at DataRobot. “People that were involved came out saying they had a desire to work with Dan again, so that ultimately was a very positive signal for me,” said Lux general partner Shahin Farshchi.?

Presumably, he spoke with Damon, George, Parm, and Tom.

“There wasn’t anything from that experience that would give me less confidence in Dan’s ability to pull this off, his ethics and his moral compass in doing so,” added Founders Fund partner Trae Stephens.

We all understand there is a low bar for ethics in Silicon Valley. The world is full of sharks and pirates. But as an investor, would you trust a guy who painted a rosy picture about his company while selling his own stake?

When you do business with a person who thinks everyone else is a sucker, you are the next sucker.

Wright declined to answer specific Forbes questions about his DataRobot tenure. “My track record of helping build successful technology companies speaks for itself,” he said in a written statement.?

Lawyer, General Counsel, COO of a sub, CEO, splat.

“I look forward to continuing to support DataRobot’s current leadership [in] any way I can and applying the lessons I learned there and at AppDynamics toward building a generational company at Armada with our incredible team.”

Wright should focus his attention on finding a customer.

At this point in his essay, Ace reporter Konrad describes Armada’s product, the unfortunately named “Galleon”:

Made out of multiple layers of interior lining and and a reinforced outer casing of military-grade heavy gauge steel, a Galleon can fit on the back of a truck or rail car for its journey to a far flung location.

There are quite a few offshore oil rigs served by rail.

There on a trailer or level ground, it can be up-and-running within 48 hours.?

Fifteen years ago, you could unload a Netezza box from the truck and run queries in two. I should point out that “level ground” tends to be in short supply at oil rigs, remote mines, advanced military bases, fire prevention outposts, and elite surfing competitions.

The real value, however, is what’s inside: in the standard 40 ft. Galleon, six racks of computer processing units, or CPUs, or GPUs; in a 20 ft. version, the housing for three.

It’s not an edge device it’s a fucking mainframe.

Armada has a Galleon up and running at its Seattle offices, where it runs a number of AI applications in demos, but for now, it’s unclear how many are active in the wild.?

Well, sir, with no customers I’m guessing the answer is “none.”

Asked to connect Forbes to customers, Armada introduced Nexa Resources, a publicly-traded mining company currently conducting a proof-of-concept trial, CIO Marcelo Alves Santos said by email. Setting up remote data centers at mines could otherwise take as long as eight months, according to Santos. “Collaborating with Armada offers the prospect of more efficient, quicker and adaptable operations, a significant step up from our current capabilities,” he said. But Armada later clarified that Nexa isn’t currently operating any Galleons in the field.

Remarkable that Mr. Santos learned so much from an office demo.

A further-along trial is happening at a well-known media company, whose executive in charge asked to remain anonymous because he wasn’t authorized to speak in an official capacity.?

If it’s Vice Media, get the money upfront.

Armada’s Commander software suite, and specifically its Connect software for managing Starlink satellite usage, is proving the difference between the SpaceX satellites serving as a primary source of internet connectivity for streaming live events. By bonding up to seven Starlink terminals (on-ground receivers working with the satellites) together as one, the media company can ensure a high bandwidth baseline it couldn’t by working with SpaceX alone.

OK, I have to ask -- what is the point of shipping data by satellite from a device the size of a boxcar?

SpaceX did not respond to a comment request. Armada called its relationship with SpaceX a “close collaboration” but declined to answer specific questions.

Someone in the SpaceX press office wonders who Armada is and why Forbes is asking about it.

Other apps from inside Armada and through partners can have more industrial value, like filling out information forms required by regulators, or making maintenance suggestions before there is an emergency, the company said.

Using AI for regulatory compliance and predictive maintenance. “Gee whiz, will ya look at that!”

At least one railroad company, which Armada declined to name, is testing the use of its apps to scan and sort containers at a transit hub.

At a transit hub. Like what, the Port of Newark? Use a modem and a landline.

David Dunaway, a retired vice admiral…is actively advising Armada on its approach to the Department of Defense, where the startup is working to bid for initial contracts. Of course, such contracts could take years to procure.

But you can lose them in months. Like DataRobot did under Wright's leadership.

And another past Founders Fund investment, publicly-traded Palantir, already works with customers within the U.S. government and heavy industries to provide AI software.?

For twenty years. No worries, Wright and Runyan will muscle them aside with their magic networks.

At least one SpaceX investor who asked to remain anonymous for fear of retaliation said they had declined to invest in Armada over concerns its technology would be too easily matched by rivals over time.

Armada’s "strategic partner" thinks it's a garbage investment.

All the more reason for Armada and its founders to announce themselves — and tout their big-business credentials — ahead of a competitive land grab. “You can’t come in and just spin out of Stanford and run this kind of company,” claimed Runyan. “Dan and I have been building networks in Silicon Valley for 20 years.”

There are plenty of successful startups run by people fresh out of Stanford. I can’t think of any successful startups run by lawyers.